REGISTER  |  SIGN IN
United States Patent5745555
MarkApril 28, 1998
Title
System and method using personal identification numbers and associated prompts for controlling unauthorized use of a security device and unauthorized access to a resource
Abstract
A system and method using personal identification numbers and associated prompts for controlling unauthorized use of a security device and unauthorized access to a resource. The method includes requesting an authorized user of a security device to select a set of N PINs and N distinct phrases, each one of the N distinct phrases being associated with a corresponding one of the N PINs for acting as a prompt to remind the user of the corresponding one of the N PINs. A current user of the security device is prompted using one of the N distinct phrases and the user's response to the prompt is compared to the associated PIN to determine whether the current user of the security device is the authorized user. The current user is granted access to the resource or is granted use of the security device if it is determined that the current user of the security device is the authorized user.
Inventors:Mark; Andrew R. (New York, NY)
Assignee:Smart Tone Authentication, Inc. (New York, NY)
Appl. No.:657594
Filed:June 7, 1996
Current U.S. Class:379/93.03 379/93.17 235/380 379/355.04 379/361 
Field of Search:379/201,67,88,89,216,200,199,93,355,95,131,112,192,188,189,190,356,361 380/19,20,23,3,4 364/408 348/5.5
U.S. Patent Documents
3706855December 1972Pitroda et al.
3806874April 1974Ehrat
3896266July 1975Waterbury
3976840August 1976Cleveland et al.
4007355February 1977Moreno
4126768November 1978Grenzow
4130738December 1978Sandstedt
4201887May 1980Burns
4205203May 1980Mehta et al.
4223185September 1980Picou
4302810November 1981Bouricius et al.
4320387March 1982Powell
4342882August 1982Gravenhorst et al.
4355210October 1982Soulliard
4399535August 1983Southard
4451701May 1984Bendig
4471165September 1984DeFino et al.
4482802November 1984Aizawa et al.
4490579December 1984Godoshian
4494114January 1985Kaish
4563548January 1986Misherghi et al.
4570260February 1986Basehore
4571462February 1986Janssen
4571463February 1986Shefler
4585904April 1986Mincone et al.
4614861September 1986Pavlov et al.
4620064October 1986Kessler et al.
4638481January 1987Crane et al.
4667087May 1987Quintana
4677657June 1987Nagata et al.
4679233July 1987Richardson et al.
4689478August 1987Hale et al.
4689760August 1987Lee et al.
4697072September 1987Kawana
4710613December 1987Shigenaga
4719322January 1988Guzik et al.
4720860January 1988Weiss
4759056July 1988Akiyama
4761807August 1988Matthews et al.
4763355August 1988Cox
4776006October 1988Comerford et al.
4797910January 1989Daudelin
4799254January 1989Dayton et al.
4815128March 1989Malek
4817135March 1989Winebaum
4827512May 1989Hirokawa et al.
4856062August 1989Weiss
4866756September 1989Crane et al.
4866773September 1989Lubarsky
4868852September 1989Tsumura
4879743November 1989Burke et al.
4882750November 1989Henderson et al.
4885778December 1989Weiss
4910767March 1990Brugliera et al.
4922527May 1990Nonami
4930129May 1990Takahira
4930160May 1990Vogel
4935608June 1990Tanaka
4935961June 1990Gargiulo et al.
4939470July 1990Kang
4941172July 1990Winebaum et al.
4972469November 1990Saltwick et al.
4972479November 1990Tobias, Jr. et al.
4980910December 1990Oba et al.
4995077February 1991Malinowski
4998279March 1991Weiss
5022080June 1991Durst et al.
5023908June 1991Weiss
5027389June 1991Chiue et al.
5027401June 1991Soltesz
5036461July 1991Elliott et al.
5050207September 1991Hitchcock
5054051October 1991Hoff
5058150October 1991Kang
5060255October 1991Brown
5063590November 1991Koshiishi
5068894November 1991Hoppe
5070233December 1991Takizawa et al.
5073941December 1991Locke
5120939June 1992Claus et al.
5125078June 1992Matsuda et al.
5168519December 1992Scarinci et al.
5168520December 1992Weiss
5170046December 1992Kusakabe
5177781January 1993Ponton et al.
5177785January 1993Itani et al.
5181238January 1993Medamana et al.
5181744January 1993Betheil
5182767January 1993Bernard
5193114March 1993Moseley
5199061March 1993Kim
5199062March 1993Von Meister et al.
5199081March 1993Saito et al.
5208446May 1993Martinez
5212373May 1993Fujioka et al.
5220158June 1993Takahira et al.
5223699June 1993Flynn et al.
5227612July 1993Le Roux
5227613July 1993Takagi et al.
5239166August 1993Graves
5239583August 1993Parrillo
5243175September 1993Kato
5247578September 1993Pailles et al.
5251259October 1993Mosley
5257309October 1993Brandman et al.
5276735January 1994Boebert et al.
5297194March 1994Hunt et al.
5329578July 1994Brennan et al.
5343519August 1994Feldman
5353342October 1994Pietrowicz
5357566October 1994Dowling, Jr. et al.
5363443November 1994Petty
5365574November 1994Hunt et al.
5369685November 1994Kero
5369689November 1994Kawamura
5377263December 1994Bazemore et al.
5392342February 1995Rosenthal
5392447February 1995Schlack et al.
5406619April 1995Akhteruzzaman et al.
5408582April 1995Colier
5414755May 1995Bahler et al.
5420914May 1995Blumhardt
5425084June 1995Brinskele
5425085June 1995Weinberger et al.
5428678June 1995Fitzpatrick et al.
5430827July 1995Rissanen
5452352September 1995Talton
5454035September 1995Oba et al.
5465290November 1995Hampton et al.
5469499November 1995Lanning
5485370January 1996Moss et al.
5499288March 1996Hunt et al.
5506885April 1996Hamilton
5510777April 1996Pilc et al.
5550575August 1996West et al.
5608784March 1997Miller
5619564April 1997Canniff et al.
5623539April 1997Bassenyemuskasa et al.
Foreign Patent Documents
5-308302Nov., 1993JP
WO 88/03294May., 1988WO
Other References
Fitzgerald, Karen, "The quest for intruder-proof computer systems", IEEE SPECTRUM, vol. 26, No. 8, pp. 22-26, Aug. 1989. .
Davis et al., "Wallet Terminal Keyboard with Acoustic Coupler", IBM Technical Disclosure Bulletin, vol. 10, No. 3, pp. 188-189, Aug. 1967. .
Marven, Craig, "General purpose tone decoding and DTMF detection", Texas Instruments Ltd., Regional Technology Centre, Bedford, England (date unknown)..~
Primary Examiner: Zele; Krista M.
Assistant Examiner: Wolinsky; Scott
Attorney, Agent or Firm:Testa, Hurwitz & Thibeault, LLP
Parent Case Text


This is a divisional of application Ser. No. 08/286,825 filed on Aug. 5, 1994, and issued as U.S. Pat. No. 5,583,933 on Dec. 10, 1996.
Claims

What is claimed is:
1. A security method for controlling unauthorized access to a resource, comprising the steps of:
requesting an authorized user of a security device to select a set of N PINs and N distinct phrases, each one of the N distinct phrases including at least one word, each one of the N distinct phrases being associated with a corresponding one of the N PINs for identifying the corresponding one of the N PINs and for acting as a prompt to remind the authorized user of the corresponding one of the N PINs, wherein N is a positive integer;
storing in the security device the N PINs and the associated N distinct phrases;
transmitting an encoded signal from the security device to a verification service, the encoded signal representing the N PINs and the associated N distinct phrases;
receiving the encoded signal from the security device at the verification service;
decoding the encoded signal at the verification service to obtain the N PINs and the associated N distinct phrases;
selecting, by the verification service, a first one of the N PINs;
prompting, by the verification service, a current user of the security device for a first time using the one of the N distinct phrases associated with the first selected one of the N PINs as a first prompt;
receiving at the verification service a first response to the first prompt from the current user;
determining at the verification service, as a function of the first response, whether the current user of the security device is the authorized user, wherein the step of determining includes the step of comparing the first response to the first selected one of the N PINs to determine if there is a match; and
granting, to the current user by the verification service, access to a resource if it is determined that the current user of the security device is the authorized user.

2. The method of claim 1, wherein the step of selecting a first one of the N PINs further includes the steps of:
generating a number using a number generator, wherein the number generator is selected from the group consisting of a pseudo random number generator, a random number generator and an incrementing register; and
selecting, as a function of the generated number, the first one of the N PINs.

3. The method of claim 1, wherein the step of prompting the current user of the security device for a first time includes the step of providing an audio prompt to the current user, the audio prompt representing the one of the N distinct phrases associated with the first selected one of the N PINs.

4. The method of claim 1, wherein the step of receiving a first response includes the step of receiving a plurality of dual tone multi-frequency (DTMF) signals representing the first selected one of the N PINs.

5. The method of claim 1, wherein the step of receiving an encoded signal includes the step of receiving a set of tones, and wherein the step of decoding the encoded signal to obtain the N PINs and the associated N distinct phrases further comprises the steps of:
monitoring a frequency independent signal characteristic of the received set of tones to obtain a plurality of signal characteristic measurements; and
using a look-up database in conjunction with the plurality of signal characteristic measurements to decode the received set of tones to obtain the N PINs and the associated N distinct phrases encoded into the set of tones.

6. The method of claim 5, wherein the received set of tones includes a plurality of DTMF signals.

7. The method of claim 1, wherein the security device is an acoustically programmable auto-dialer, the resource is a long distance telephone service provider and wherein the step of receiving an encoded signal includes the step of receiving an encoded DTMF signal representing a telephone number, the N PINs and the associated N distinct phrases.

8. The method of claim 1, further comprising the step of:
generating, by the verification service, an acoustic signal, to which the security device is programmed to respond by deactivating itself, upon a determination that the current user of the security device is not the authorized user.

9. The method of claim 1, wherein the step of determining whether the current user of the security device is the authorized user further comprises the step of:
providing the current user a second opportunity to provide proof that the current user is the authorized user if it is determined that the first response does not match the first selected one of the N PINs, the step of providing the current user a second opportunity further comprising the steps of:
i) selecting, by the verification service, a second one of the N PINs;
ii) prompting, by the verification service, the current user of the security device for a second time using the one of the N distinct phrases associated with the first one of the N PINs and the one of the N distinct phrases associated with the second selected one of the N PINs;
iii) receiving at the verification service a second response; and
iv) comparing at the verification service the second response to the first selected one of the N PINs and the second selected one of the N PINs to determine if there is a match, wherein a match of both the first and second selected ones of the N PINs indicates that the current user of the security device is the authorized user.

10. The method of claim 1, wherein the step of storing in the security device the N PINs and the associated N distinct phrases further comprises the step of programming the security device by transmitting audio signals to the security device.

11. The method of claim 1 wherein the resource comprises the verification service.

12. A security method for controlling unauthorized use of a device, comprising the steps of:
requesting an authorized user of a device having memory, a display and input keys, to select a set of N PINs and N distinct phrases, each one of the N distinct phrases including at least one word, each one of the N distinct phrases being associated with a corresponding one of the N PINs for identifying the corresponding one of the N PINs and for acting as a prompt to remind the authorized user of the corresponding one of the N PINs, wherein N is a positive integer;
storing in the device the N PINs and the associated N distinct phrases;
selecting, by the device, a first one of the N PINs;
displaying, by the device, a first list of PINs from which a current user can select a PIN by using the input keys, the first list of PINs including the first selected one of the N PINs;
prompting, by the device, the current user of the security device for a first time using the one of the N distinct phrases associated with the first selected one of the N PINs as a first prompt, to select from the displayed first list of PINs the first selected one of the N PINs;
determining, by the device, whether the current user of the device is the authorized user, wherein the step of determining includes the step of detecting whether the current user selected the first selected one of the N PINs in response to the first prompt; and
enabling, by the device, the current user to operate the device upon determining that the current user is the authorized user.

13. The method of claim 12, wherein the step of determining whether the current user of the device is the authorized user further comprises the steps of:
selecting, by the device, a second one of the N PINs;
displaying, by the device, a second list of PINs from which the current user can select a PIN by using the input keys, the second list of PINs including the first selected one of the N PINs and the second selected one of the N PINs;
prompting, by the device, the current user of the security device for a second time using the one of the N distinct phrases associated with the second selected one of the N PINs, to first select from the displayed second list of PINs the first selected one of the N PINs and to then select from the displayed second list of PINs the second selected one of the N PINs; and
detecting, by the device, whether the current user of the device selected the first selected one of the N PINs and then the second selected one of the N PINs to determine if the current user is the authorized user.
Description
FIELD OF THE INVENTION

The present invention is directed to methods and apparatus for communicating data, telecommunications access methods, and, more particularly, to auto-dialers, security and information devices that can transmit and receive data.

BACKGROUND OF THE INVENTION

In the modern world, telephone transactions involving extensions of credit, payment of bills, fund transfers and the providing of other types of services are commonplace. Generally, such services are provided in response to a user dialing the telephone number of a service and by then entering identification information and/or credit card information using standard "touch tones", i.e., dual tone multi-frequency signals ("DTMF tones"), to represent the identification information being entered and transmitted. Normally, such touch tone signals are produced using a standard telephone keypad input device.

Touch tones are generated using a dual tone multi-frequency (DTMF) encoding technique, as opposed to a frequency shift key (FSK) encoding, which is frequently used for data transmission purposes. In accordance with the DTMF technique used to generate touch tone signals, tone signals are produced by generating two tones such that one tone is selected from a high frequency band group and the other tone is selected from a low frequency band group. In standard telephone systems, the high frequency band group includes four high frequencies (1209, 1336, 1477, and 1633 Hz) while the low band frequency group includes four low frequencies (697, 770, 852 and 941 Hz). Each of the high and low frequencies is referred to as a fundamental frequency.

Each one of the low frequencies corresponds to one of the four rows of keys on a standard extended telephone keypad while each one of the four high frequencies corresponds to one of the four columns of keys on standard extended telephone keypads. Accordingly, low frequency tones represent row tones and high frequency tones represent column tones. It should be noted that extended and non-extended keypads differ in that extended keypads include the additional fourth column of keys not found on non-extended standard keypads such as those commonly used with public telephones and household telephones, although these additional tones are found in most modem hardware/software systems.

Each different telephone key is represented by a signal including a unique combination of one tone from the high band and one from the low band. Sixteen different signal states may be represented by this encoding technique with one signal state corresponding to each one of the sixteen keys that can be found on a standard telephone keypad. Referring now to FIG. 8A, there is illustrated a chart which lists the 16 different numbers/symbols that are represented by the 16 different signal states and the Hi-tone and Lo-tone frequency associated with each of the 16 different signal states.

A received DTMF tone signal is determined, in telephone switching and DTMF tone signal detection devices, as having a valid signal state if five conditions are met. The first of these is that the tone signal contain exactly one valid tone, and only one valid tone, from each of the low and high band frequency groups, i.e., the signal must contain only one valid Hi-tone and one valid Lo-tone frequency. The second condition is that the low and high tones are present for a predetermined minimum time duration, i.e., at least 35-40 milliseconds. Third, the difference in power (level), commonly referred to as "twist", between the low and the high tone must fall within a predetermined range, i.e., the Hi-band tone signal power level can not be more than 4 dBm greater or 8 dBm less than the Lo-band tone signal power level, where dBm is a logarithmic measure of power with respect to a reference power of 1 milliwatt. Fourth, the amplitude level of each tone signal in the tone pair must be in the range of 0 to -25 dBm. Fifth, consecutive tone-pairs must be separated by a period of silence for at least 35-40 milliseconds.

Thus, if too many or too few tones are detected the detection criteria will not be met and a valid signal state will not be indicated. When a valid signal state does occur, the particular combination of low and high tones is decoded to produce an indication of the corresponding key or signal state that was responsible for the generation of the DTMF tone signal. It is this key or signal state information that represents, e.g., one of the numerical digits comprising a telephone or credit card number.

Placing a voice call, using a calling card number or other credit card number, is exemplary of one of the most common types of telephone credit transactions involving the use of DTMF tones.

Referring now to FIG. 1, there is illustrated a flow chart illustrating a standard telephone call transaction involving the use of a credit card for billing purposes. As illustrated, the standard telephone call transaction comprises the first step of making a decision to place a credit card call 1001 followed by the actual step of dialing 1002. As part of the dialing step 1002 a user enters, using, e.g., the telephone keypad, an access code identifying the desired long distance telephone carrier, and a destination number, i.e., the telephone number of the party being called. Both the access code and telephone number are represented as DTMF tones, i.e., touch tones, generated by the telephone in response to the keypad input.

The telephone system, e.g., the local switching office to which the telephone is linked, connects the caller to the appropriate long distance carrier represented by the access code input by the user. The long distance carrier then generates an audible signal/message indicating to the caller that it is ready to receive billing information as indicated in step 1003.

In response to the audible signal/message generated by the long distance carrier, the caller then inputs, e.g., using the telephone keypad, a credit card number as illustrated in step 1004. In response to receiving DTMF tones representing the credit card number, generated by the telephone, the long distance carrier checks the credit card number for validity as illustrated in step 1006. If the credit card number is determined to be valid, the call is placed as illustrated in step 1008. However, if the credit card number is determined to be invalid, the call is rejected and the telephone connection is disconnected as illustrated in step 1010.

The standard procedure for placing a call using a credit card has several drawbacks. For example, requiring a caller to manually input through the telephone keypad a carrier access number, a destination number, and a credit card number introduces into the calling procedure ample opportunity for a user to accidentally input an incorrect number for any one of the required values, which can exceed, e.g., 35 required inputs- more if the call is placed to or from a foreign country.

It is generally understood that the likelihood of entry errors increases in proportion to the number of digits to be entered. Such an error normally results in the call being rejected by the telephone carrier, requiring the caller to repeat the entire calling procedure from the point of connection to the long distance carrier. Generally, the long distance carrier pays fees to the company which owns the originating local switching office of the caller from the moment of connection, and is unable to start billing for the ultimate connection until the moment of connection. In such a case, each entry error is an increase in the unbillable time that a long distance carrier must absorb without any offsetting revenue.

Generally, the requirement that a caller manually input a large series of numbers to place a call leads to a higher error rate during call placement than results when the caller has to input fewer numbers, e.g., when calls are placed without the use of credit cards. In addition, requiring a caller to manually input a calling card number discourages some callers from using a credit card to place a call because of the additional time and frequent input errors associated with the initiation of a call as compared to calls placed without using credit cards.

In addition to error problems associated with the manual input of credit card number information, security problems are also associated with the manual input of credit card data into a telephone using a standard keypad. For example, a person viewing the initiation of a telephone transaction can record the credit card number input to the telephone keypad and then later use the calling card number to place unauthorized calls.

Portable electronic information cards and auto-dialers that are capable of being acoustically coupled to telephone systems to perform dialing functions are well known in the art.

Such known devices which generate a series of DTMF tones representing the numbers which must be input to initiate a call, have reduced or eliminated the need to input telephone number, carrier number, and credit card number information manually each time a call is placed.

Frequently, to enhance the versatility of such devices, they are made programmable with individual devices being programmed to store different telephone numbers, and/or calling card, credit card or personal identification numbers (PINs). While such programming is generally performed by electronically coupling such programmable devices to a programming unit, for example, as described in U.S. Pat. No. 4,882,750 to Henderson et al., it has been suggested that such devices should be designed to be capable of being programmed by acoustic signals received from a telephone. For example, PCT patent application Ser. No. 02837, now abandoned, suggests a portable electronic information card capable of being programmed in response to acoustic signals.

While known portable electronic information cards and auto-dialers that can be acoustically coupled to telephone systems facilitate telephone dialing and the supplying of billing information over the phone, the lack of a practical workable device which deals with past data transmission errors and security problems has inhibited widespread acceptance and use of such auto-dialer devices.

The introduction of errors into the data being sent to the telephone system, e.g., as the result of the use of an acoustic coupling, is one example of a data error problem that may prevent the telephone system from completing a call. As will be discussed in detail below, errors associated with the use of an acoustic coupling result from various factors affecting the acoustical transmission of DTMF tones. Such factors include variations between components used in present auto-dialers to generate the DTMF tones, temperature variations affecting battery voltages and the amplification levels applied to DTMF signals, speaker proximity to a telephone handset's microphone used to receive the acoustic DTMF tones, distortions introduced by the microphone receiving DTMF tones output by an auto-dialer, as well as, ambient noise levels. Errors may also be introduced from the lines that connect a caller to the ultimate telephone call destination as well as from other sources.

In addition to the error problems associated with the use of known auto-dialers, known devices for providing calling card and caller identification information acoustically to a telephone system present many security problems. For example, an unauthorized tape recording of a calling card number generated by the known systems, can easily be created by connecting an input cable to the telephone cable connected to a coin phone and then played back by an unauthorized user seeking to obtain access to the telephone system. Generally, the known calling card systems fail to provide a security method for preventing unauthorized users from gaining access to the telephone system via the use of such an unauthorized recording. Furthermore, known systems fail to prevent an unauthorized person from obtaining a calling card number and its related Personal Identification Number (PIN), both of which can be repeatedly used in their identical form for subsequent calls, through other methods, e.g., video-taping an authorized caller in a public place, going through the trash outside of large office facility, etc. Once known, calling card number and PIN data can be used to initiate multitudes of calls, often to foreign countries. The aggregate cost of the fraudulent use of unauthorized calling card calls in the United States is estimated to exceed $1 billion per year. While long distance carriers are now providing software analysis of many calls placed on their network to determine if there is a likelihood of unauthorized use, many such calls are paid by authorized customers who fail to notice the unauthorized calls on their bills.

Accordingly, there is a need to provide a secure device and method for storing and providing information regarding a user, and, more specifically, for providing such information locally or transmitting it remotely by, e.g., calling card, credit card, and user identification information either through or to a telephone system.

In addition to the reliability and security issues discussed above, there are also several convenience problems associated with known auto-dialer devices some of which are inherent in the standard credit card calling transaction illustrated in FIG. 1. These convenience problems include the need to generate and/or input separate telephone number and calling card identification information when attempting to place a credit card call. Furthermore, the procedures often used to initiate a calling card call often require that a user interlace carrier access codes (up to 20 digits) followed by the user's desired destination number (up to 17 digits) followed by an account/pin number array (usually at least 14 digits). These input sequences, which may expand in the future to accommodate greater call/user volumes are difficult sequences for a user to supply without input errors, and because of their complexity often result in users holding a publicly viewable card showing the account number and dialing sequences needed, creating security risks from unscrupulous users of calling card account and access data.

Additionally, because many countries have a distinct dialing sequence and numbering plan, carrier access codes are rarely identical from country to country, thereby increasing the difficulty that a traveler experiences when placing a calling card call.

Other convenience problems with the known devices relate to size and general ease of use issues. For example, known devices may be of such a size that it is inconvenient for a user to keep the device with them throughout the day. In addition, the size and shape of many of the known auto-dialers makes it difficult to properly align the auto-dialer with the microphone of a standard telephone handset making it somewhat difficult to use if accurate transmission of DTMF signals is to be achieved. Furthermore, the number of keys and the complexity of the controls frequently encountered on the known portable electronic information cards and auto-dialers have tended to inhibit the known devices from gaining widespread acceptance.

While the problems described above mainly refer to the problems and errors associated with the initiation of calling card calls, each of these problems may also affect access security and convenience in relation to any telephone based transaction (e.g. credit card purchases of over the phone, access to secure voice and data networks, etc.).

SUMMARY OF THE PRESENT INVENTION

The present invention provides a communication device capable of transmitting and receiving data. In accordance with the present invention, data may be encoded into a signal comprising a series of tones or tone pairs, e.g., a conventional DTMF signal, by controlling various identifiable components or aspects of the signal, e.g, the amplitude, frequency, lack of tone, or the period between tones, etc. When multiple signals are used, e.g., a combined Lo-frequency tone signal and a Hi-frequency tone signal, the twist, i.e., amplitude difference between the signals, may also be used to encode data.

In accordance with one embodiment of the present invention, the signal characteristics described above are used to encode data into a DTMF signal which is still detectable by a standard DTMF signal detector. In this manner, the present invention permits data to be embedded or encoded into a DTMF signal being used, e.g., to place a call. The embedded data may represent, e.g., a telephone calling card number, destination telephone number information or other data.

In one embodiment, the present invention is implemented as an auto-dialer also suitable for use as a smart card which is capable of transmitting and receiving information over conventional telephone lines, e.g., between a database and the auto-dialer, without the need for a specialized interface (other than a standard telephone). The auto-dialer of the present invention is capable of being acoustically coupled to the receiver of a telephone and being reprogrammed in response to acoustic signals. The programming and other features of the auto-dialer can be individually enabled or disabled in response to pre-selected signals, e.g., a string of DTMF tones received by the auto-dialer. In this manner, in one embodiment, the auto-dialer of the present invention requires an acoustic "key" to enable/disable some functions.

Using the encoding technique described above, an auto-dialer according to the present invention can encrypt calling card and other data into, e.g., destination telephone numbers by selectively altering pre-selected characteristics of a DTMF tone sequence, such as the duration of tones, the period of silence between tones and the twist between Lo-band and Hi-band tones of DTMF tone pairs in the conventional DTMF protocol which represent the desired destination telephone number. In accordance with the present invention, the encryption of data into the destination telephone number does not affect the ability of standard telephone switching equipment to recognize the destination number. However, information encrypted into the DTMF signals will be undetectable to standard telephone switching circuitry because it is encrypted using DTMF signal characteristics not normally used to represent data related to conventional call processing. In one embodiment, other tones or frequencies are also used to transmit data which cannot be detected by standard DTMF tone detectors.

In addition to the encryption capability described above, the auto-dialer of the present invention, in one embodiment, has a system clock that is used to drive a pseudo random number generator used in various data security schemes.

The auto-dialer of the present invention incorporates various calibration features which permit the calibration of the audio output and system clock. In one embodiment calibration adjustments are made by programming the auto-dialer with various calibration factors. This programming may be done by an acoustic coupling device incorporated into the auto-dialer or via another input device, e.g., wired to the auto-dialer either at the time of manufacture or over the phone at a later time. The calibration features permit the easy calibration of both the system clock and various characteristics of the tones generated by the auto dialer.

In addition to the various encoding and calibration features of the present invention, the present invention is also directed to a variety of security schemes and features which are designed to generally increase security when placing a telephone call and/or when providing other confidential or user-identification-related information.

Each of the above described features of the present invention, as well as numerous other features, will be described in greater detail below.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a flow chart illustrating the steps involved with placing a standard telephone call using a credit card.

FIG. 2 is a schematic block diagram of an exemplary embodiment of an auto-dialer in accordance with one embodiment of the present invention.

FIG. 3 is a more detailed diagram of the auto-dialer illustrated in FIG. 2.

FIG. 4 is a diagram illustrating DTMF decoder circuitry suitable for use in the auto-dialer of FIGS. 2 and 3.

FIG. 5 is a diagram illustrating DTMF encoder circuitry suitable for use in the auto-dialer of FIGS. 2 and 3.

FIG. 6 is a flow chart illustrating the steps associated with the placing of a credit card call using the auto-dialer of the present invention.

FIG. 7 is a block diagram of the auto-dialer of the present invention being coupled to a destination telephone via a carrier switching center.

FIG. 8A is a chart illustrating the fundamental difference between the tones of various DTMF tone pairs generated by standard DTMF generators.

FIGS. 8B-8E are charts illustrating the accept/reject and out-of-range frequencies, of a standard DTMF detector circuit, in relationship to each of the four fundamental Lo-band frequencies and four fundamental Hi-band frequencies used to generate standard DTMF signals.

FIG. 8F is a chart illustrating exemplary output levels of both carbon and electret microphones when receiving acoustic high-frequency tones associated with the DTMF tone pair representing the indicated digits.

FIG. 8G is a chart illustrating exemplary output levels of both carbon and electret microphones when receiving acoustic low-frequency tones associated with the DTMF tone pairs representing the indicated digits.

FIG. 9A is a chart illustrating the fundamental difference between the various tone pairs when the frequency of the Lo-band and Hi-band tones are selected, in accordance with the present invention, to reduce third tone errors.

FIG. 9B is a chart illustrating twist values of DTMF tone pairs generated in accordance with one embodiment of the present invention.

FIG. 9C illustrates accoustic output levels of Hi and Lo tones.

FIG. 10A is an illustration of a bottom view of one embodiment of the auto-dialer of the present invention.

FIG. 1OB is an illustration of a top view of the auto-dialer illustrated in FIG. 1OA.

FIG. 1OC is an illustration a side view of the auto-dialer illustrated in FIG. 10A.

FIG. 1OD is an illustration of a cut away side view of the auto-dialer illustrated in FIG. 10A.

FIG. 1OE is an illustration of the auto-dialer illustrated in FIG. 1OA mounted on a key ring.

FIG. 11 is a chart illustrating exemplary signal characteristic values, for a DTMF tone pair representing the digit 3, that can be used to represent data in accordance with the present invention.

FIG. 12A illustrates an exemplary Lo-band tone level to data conversion table in accordance with the present invention.

FIG. 12B illustrates an exemplary Hi-band tone level to data conversion table in accordance with the present invention.

FIG. 12C illustrates an exemplary tone duration to data conversion table in accordance with the present invention.

FIG. 12D illustrates an exemplary interdigit period duration to data conversion table in accordance with the present invention.

FIG. 12E illustrates an exemplary Lo-band tone frequency deviation to data conversion table in accordance with the present invention.

FIG. 12F illustrates an exemplary Hi-band tone frequency deviation to data conversion table in accordance with the present invention.

FIG. 12G illustrates an exemplary tone pair twist level to data conversion table in accordance with the present invention.

FIG. 13 illustrates a chart which displays the data conversion results obtained by using the signal characteristic values of chart 11 in conjunction with the conversion tables of FIGS. 12A through 12G to decode information encoded into a DTMF tone pair in accordance with the present invention.

FIG. 14 is a schematic block diagram of an access control device implemented in accordance with one embodiment of the present invention.

DETAILED DESCRIPTION

The present invention is directed to methods and apparatus for communicating data through the use of acoustic or electrical signals including, e.g., DTMF signals. Various embodiments of the present invention are directed to, e.g., portable acoustically coupled auto-dialers, calling cards, credit cards, paging devices, smart cards, and other information card type devices, debit cards, etc. In addition to such portable embodiments, several embodiments of the present invention are directed to telephone switching circuitry and DTMF and other tone recognition circuitry which may be incorporated into existing telephone systems, as well as other data systems.

Referring now to FIG. 2 there is illustrated an auto-dialer device, generally indicated by the reference numeral 100, in accordance with one exemplary embodiment of the present invention. As illustrated, the auto-dialer 100 comprises a microprocessor 104 coupled to a read only memory (ROM) 106, device control keys 105, and a random access memory (RAM) 108. The ROM 106 may be located within the microprocessor 104 or externally thereto.

Via device control keys 105, the microprocessor 104 receives input signals from a user, which input is stored in the RAM 108 or processed by the microprocessor 104 using other information stored in the ROM 106. In addition, in various embodiments, the RAM 108 is used to store data relating to voice signals and/or tone signals.

The auto-dialer 100 further comprises a DTMF encoder 110 and a DTMF decoder 112 which are coupled to the microprocessor 104 and to a speaker/microphone 114. In the illustrated embodiment, the speaker/microphone 114 serves as both an input device for receiving acoustic signals, such as DTMF tones, and as an output device for outputting signals such as DTMF tones and other signals generated by the encoder 110. In other embodiments, a separate microphone is used for receiving audio signals and the speaker/microphone 114 is used only for outputting signals.

As illustrated in FIG. 2, the auto-dialer 100 may be acoustically coupled to a standard telephone 122 such as a public pay phone. While the speaker/microphone 114 is illustrated in FIG. 2 as receiving a signal from and sending a signal to, a telephone handset 121, it is to be understood that the speaker/microphone 114, in this embodiment, can not be used to perform these operations simultaneously. Furthermore, it should be noted that when receiving signals from the handset 121, the speaker/microphone 114, which serves as a transducer, is positioned in close proximity to the handset's speaker 120 and while sending signals to the handset's microphone 118 the speaker/microphone 114 is positioned in close proximity to the microphone
118. Thus, to change between the send and receive functions, in the illustrated embodiment, a user must move the auto-dialer 100 from being in close proximity to the microphone 118 to a position where it is in close proximity to the speaker 120. However, in another embodiment, a separate microphone is included for the receipt of data in addition to the speaker/microphone 114. In accordance with such an embodiment, data may be received and transmitted simultaneously by the auto-dialer 100, without the requirement of moving the auto-dialer 100.

In another embodiment, the auto-dialer 100 is designed to acoustically monitor its output and perform an auto-calibration sequence prior to or at the beginning of each period of use of the auto-dialer 100 which follows a period of dormancy of a preselected time period, e.g., a selected number of hours or days or when the auto-dialer 100 senses a temperature outside of a preselected temperature range, e.g., representing the temperature the auto-dialer 100 is expected to work at.

As is well known, battery voltage output varies as a function of temperature. Variations in voltage output are particularly noticeable in cold weather. Additionally, other components of the auto-dialer 100 are subject to the effects of temperature. Accordingly, in one embodiment, the auto-dialer 100 incorporates an auto-calibration feature which causes the generation of a preselected group of tones used for calibration purposes. As these tones are generated, the microphone 109
receives the tones and converts them to electrical signals which are analyzed by the microprocessor 104. The microprocessor 104 analyzes the generated signal levels and compares them against stored desired signal level values. In the event that an adjustment is required in any tone output level, or other signal characteristic, the microprocessor 104 calculates the appropriate change, and alters a tone generation control parameter stored in the RAM 108 to correct the output signal level. The auto-dialer 100 may then re-test the generation of the tone which had a tone signal level problem to insure accurate generation of the tone to insure that the calculated parameter change produced the desired result. If the desired output level was not achieved, the microprocessor 104 repeats the calibration sequence. In one embodiment, when it is detected that a tone signal level fails to achieve the pre-determined level, e.g., desired signal level after one or more attempts to adjust the output level, the auto-dialer 100 indicates a "don't use" condition on a display device 202.

Referring now to FIG. 3, the auto-dialer 100 of FIG. 1 is illustrated in greater detail. In FIG. 3, components that are the same or similar to those of FIG. 2, will be referred to using the same reference numerals as used in FIG. 2. As illustrated in FIG. 3, the auto-dialer 100 may further comprise the display device 202 for displaying data and other information output by the microprocessor 104 and/or system components, a main battery 208 for powering the auto-dialer 100, a back-up battery 206 for supplying power to the microprocessor 104 as well as other system components when the main battery fails, and a voltage comparator 210 for detecting the condition of the main and backup batteries 206, 208.

As illustrated in FIG. 3, the auto-dialer further includes a micro-power amplifier 226 coupled to the output of the speaker/microphone 114. The amplifier 226 serves to provide a wake-up signal to the microprocessor 104 as described below. The amplifier 226 generates a signal in response to a signal generated by the speaker 114 in response to received acoustic signals. The signal output by the amplifier 226 causes the microprocessor 104 to become fully active from, e.g., a sleep mode which is automatically entered into after a long period of inactivity in order to conserve power. In an alternative embodiment, an input of the microprocessor 104 is coupled to a light sensor or other activation device such as a radio frequency sensor, which causes the auto-dialer 100 to become fully active in response to an outside stimulus which may be provided by, e.g., a light or sound source associated with an automatic teller machine (ATM) or telephone device. Thus, in accordance with such an embodiment the auto-dialer 100 can be made active by the excitation of a transducer or other sensor by, e.g., a light, radio frequency signal or the receipt of an acoustic signal having a pre-defined frequency and a minimum, pre-defined intensity level. These predefined levels or values are a matter of design choice and may be programmed into the ROM 106 or RAM 108 at, e.g., the time of manufacture.

With regard to "wake-up" features, in one embodiment, the auto-dialer 100 incorporates a reflective surface, e.g., a reflective ring 160 as illustrated in FIG. 10A or other component which can be detected by an interface of e.g., an ATM machine. For example, an ATM machine may detect the presence of the auto-dialer 100 because of its unique shape when it is place in contact with the ATM machine. As will be discussed below, with regard to FIG. 10A, the auto-dialer 100 may include a notch or cut out designed to mate with the shape of the input area of the interface with which the auto-dialer 100 is intended to communicate. Because of the auto-dialer's reflective ring 160 shape, or other physical characteristic, the interface of the machine with which the auto-dialer 100 is designed to communicate can detect the presence of the auto-dialer 100 and signal the auto-dialer through, e.g., a series of tones, to become active, e.g., "wake-up". Accordingly, the auto-dialer 100 permits interfacing hardware to recognize the presence of the auto-dialer 100 by, e.g., the presence of a highly reflective material, e.g., a mirrored film or other similar material, which reflects an emitted light from the interfacing hardware in such a fashion as to cause a light detecting sensor in the interfacing hardware to sense the reflection of such light, and upon the sensation of such light, cause the interfacing hardware to emit a pre-determined set of tones which will cause the auto-dialer 100 to become active and enter a mode of operation which may not otherwise be available to the user.

The ROM 106 includes a series of memory locations which contain information that serves as a set of permanent data tables, as well as a computer program instructions for controlling the operation of the microprocessor 104. The permanent data tables may contain e.g., long distance carrier information, area code information, data encoding/decoding information, and/or credit service related information as will be described further below.

The RAM 108, on the other hand, is used to store information that is device dependent, is likely to change, or for other reasons is more easily stored in an alterable memory device than in a ROM. As illustrated by the representative memory map
211, the RAM 108 may dedicate memory space to storing DTMF transfer and receive parameters 214 used for encoding/decoding signals, mode tone pairs information 216, e.g., frequency information relating to supported tone pairs, display memory 218, system control data 224, e.g., calibration parameters, user data 222, (e.g., destination phone numbers and billing information relating to the individual who is authorized to use the auto-dialer 100) and device data 220, (e.g., one or more numeric or alpha-numeric sequences which identify the particular auto-dialer 100, manufacturing date information, etc.).

A circuit suitable for use as the DTMF decoder circuit 112 illustrated in FIGS. 2 and 3 will now be described with reference to the schematic block diagram of FIG. 4. As illustrated in FIG. 4, the DTMF decoder circuit 112 comprises an amplifier and filter circuit 302 that has an input coupled to the output of the speaker/microphone 114 and a received signal output coupled to the input of a Hi-band passband filter 304 and a Lo-band passband filter 306. In this embodiment, the speaker 114 acts as an inductor converting acoustic signals received from, e.g., the speaker 120 of the telephone handset 121, into electrical signals which are amplified and filtered by the amplifier and filter circuit 302 and then further filtered by the passband filters 304, 306. The Hi-band passband filter 304, is designed to pass the corresponding Hi-band frequency DTMF signals while eliminating noise and other signals. Similarly, the Lo-band passband filter 306 is designed to pass the Lo-band frequency DTMF signals and to eliminate other signals. In this manner, the Lo-band and Hi-band signals are segregated from each other with noise, i.e., signals having frequencies outside the bands of the DTMF signals being removed to facilitate the later decoding of the signals.

An output of the Hi-band filter 304 is coupled to the input of a column frequency detector 308 for detecting which frequency of the set of Hi-band tone frequencies is being received. Similarly, the Lo-band filter 306 has an output coupled to an input of a row frequency detector 310 for detecting which frequency of the set of Lo-band frequencies is being received. In particular embodiments, the column and row frequency detectors 308, 310 as well as Hi and Lo band filters 304, 306 may be designed to recognize and pass additional Hi-band and Lo-band tones, respectively, which are outside the range of standard DTMF tones to thus increase the number of signals which can be used to transmit data to add additional security, increase data transmission rates, or provide additional features.

An output of the column frequency detector 308 and an output of the row frequency detector 310 are coupled to corresponding inputs of a DTMF signal detector 312. The DTMF signal detector 312 receives the Lo-band and Hi-band tones output by the column and row frequency detectors 308, 310 along with information signals indicating the frequency of the received tones. The DTMF detector 312 determines if the received tones constitute a valid tone pair or other signal which the DTMF signal detector
312 is programmed to recognize. If the DTMF signal detector 312 detects a valid tone pair or a signal it recognizes, it sends a signal to a DTMF tone to data converter circuit 316 of the microprocessor 104 to convert the detected DTMF tone or signal into the data it represents, e.g., a symbol or number.

Because the auto-dialer 100 is programmable, it may be programmed or reprogrammed to accept one or more signals as valid tones. These tones may include tones other than those used for standard DTMF signals. Furthermore, it can be programmed to reject or ignore input data which does not conform to predetermined signal characteristics which are stored in the RAM 108 of the auto-dialer 100. In one embodiment, these signal characteristics (e.g. maximum tone-length) may be remotely modified via, the acoustic reprogramming of the auto-dialer 100 in response to the auto-dialer 100 receiving a series of DTMF tones. Such tones act as a signal or key which is required to enable the reprogramming of the auto-dialer 100. In addition, because the auto-dialer 100 is designed to be both responsive to, and capable of, generating audio tones, e.g., both standard and encoded DTMF tones, as will be described below, the auto-dialer 100 is capable of receiving, storing and transmitting both standard and encoded DTMF tones for a variety of purposes including for use as passwords and "keys" to enable certain functions of the auto-dialer 100 or the device which the auto-dialer 100 is used to communicate with.

In yet another embodiment, the auto-dialer 100 is programmed to, upon the receipt of pre-selected series or group of tones, representing commands or instructions to the microprocessor 104, perform mathematical computations based on either data stored within the auto-dialer 100 and/or on data which is received by the auto-dialer 100 in response to the received commands. In such an embodiment, the performed computation(s) is in accordance with a received instruction and is performed in such a manner that a user can not effect the result of the computation by manipulating the keys on the auto-dialer 100. In this manner, the auto-dialer 100, because of its programming and ability to receive commands and data from an outside source, can perform, e.g., debit/credit calculations with the user being unable to manipulate the result from the device control keys 105 of the auto-dialer 100. Additional security features to prevent unauthorized manipulation of data stored in various locations within the memory 106, 108 of the auto-dialer 100 will be discussed below.

As discussed above, for security reasons, the reprogramming feature is, in some embodiments, enabled only upon the receipt of a pre-selected group of acoustic tones which serve as a key to indicate to the auto-dialer 100 that an authorized individual is in fact reprogramming the device. Different acoustic keys or tone sequences may be used to limit access to different memory locations. In this manner, one key, e.g., a series of tones, may be required to reprogram one memory location while another key may be required to reprogram another memory location. In this manner, the ability to alter the contents of various memory locations containing, e.g., personal identification telephone numbers, prefix information, etc. can be restricted so that the user cannot change the contents of certain memory locations and so that only services authorized to alter particular items in memory, e.g., dialing prefixes, country codes, etc. can alter such information. In such an embodiment, a first series of tones is required to alter the contents of a first memory location while a second series of tones is required to alter a second memory location. Additional tone sequences or "keys" may be associated with additional memory locations.

In one embodiment, the DTMF signal detector 312 of the present invention referred to as an enhanced DTMF signal detector is able to detect alterable characteristics of a DTMF signal, e.g., signal twist, Lo-band and Hi-band tone amplitude, Lo-band and Hi-band tone duration, tone frequency, etc. which may be used to encode information into a DTMF signal without affecting the ability of a standard DTMF signal detector to detect the symbol/number represented by a DTMF tone pair. If the DTMF signal detector 312 detects encoded information the encoded information is supplied to the DTMF tone to data converter 316 for processing. A particular signal or sequence of tones is used in some embodiments to provide an indicator signal to indicate to a receiver that encoded DTMF signals are being transmitted. In such embodiments, a DTMF signal detector detects the receipt of encoded DTMF signals by checking a received signal for such an indicator signal or indicator sequence of tones.

The DTMF signal detector 312 also has start and stop signal outputs coupled to corresponding inputs of a non-tone demodulation circuit 314 of the microprocessor 104. In this manner, the non-tone demodulation circuit 314 receives timing information concerning the starting and stopping of each received signal. As will be discussed below, this information can be used, in accordance with one embodiment of the present invention, for decoding information encoded into one or more DTMF signals and/or for distinguishing of a string of signals which represent meaningful data as opposed to nonsense signals added for security reasons as well as to enable the device to provide non-frequency dependent data that is encoded into the interdigit periods, i.e., the time between DTMF tone pairs.

Referring now to FIG. 5, DTMF encoder circuit 110, which may be used as illustrated in FIGS. 2 and 3, will now be described in detail. The DTMF encoder 110 comprises a high frequency register 424, a tone select register 426, and a low frequency register 428.

The high and low frequency registers 424, 428 have a first input coupled to a data output of the microprocessor 104, a second input coupled to a tone select output of the microprocessor 104 and a third input coupled to a tone select signal output of the tone select register 426.

The tone select register 426 receives tone signal information from a tone store output of the microprocessor 104 which is then processed to generate a control signal which is supplied to the low and high frequency registers 424, 428 through the third input of the registers 424, 428.

The high and low frequency registers 424, 428 are responsive to signals received from the microprocessor 104 and the tone select register 426 to produce a control signal indicating the fraction of the microprocessor's clock frequency the desired Hi and Lo tones correspond to.

The Hi-band DTMF tone of each DTMF tone signal pair is generated by a Hi-band frequency signal generation circuit 401. The Hi-band frequency signal generation circuit 401 comprises a programmable divider 430, which is coupled to a Johnson counter 434. The Johnson counter 434 is coupled to digital to analog converter 438 which has an output coupled to an amplifier 458 which is responsible for amplifying the Hi-band DTMF tone signals of each tone pair.

The programmable divider 430 receives as input signals the output of the high frequency register 424 and the microprocessor's oscillator. Using the control information provided by the high frequency register 424, the programmable divider generates a digital signal having the desired frequency of the Hi-band tone to be generated from the oscillator signal. This digital signal is then further processed by the Johnson counter 434 before being converted into an analog signal by the D/A converter 438.

It should be noted that to avoid the problems that may result from harmonics associated with squarewaves, the D/A converter 438 only generates pure sine waves. The analog Hi-tone output signal, output of the D/A converter 438, is amplified by the amplifier 458 which has a gain control input coupled to a Hi-band amplitude control signal output of the microprocessor 104.

As will be discussed below, the degree of amplification performed by the amplifier 458 is controlled by the microprocessor 104. In this manner, the microprocessor 104 can introduce intentional twist into the DTMF signal being generated and/or encode information into the DTMF signal by selectively varying signal strength and/or twist associated with the tone pairs comprising the DTMF signal being generated.

A Lo-band frequency signal generation circuit 403 comprising a programmable divider 432, a Johnson counter 436, a (D/A) digital to analog converter 440, and an amplifier 460 is responsive to the output of the low frequency register 428, the microprocessor's oscillator, and the microprocessor's Lo-band amplitude control signal, to generate the Lo-band DTMF tone in the same manner as described above with regard to the generation of Hi-band DTMF tones.

The output of each of the amplifiers 458 and 460 which comprise the Hi-band and Lo-band tones, respectively, of each DTMF tone pair being generated, are supplied to first and second inputs of a dual ported amplifier 462 for additional amplification. The amplifier 462 has a control input which is coupled to a timing control output of the microprocessor 104.

The timing control signal is used to control the amount or level of amplification the amplifier 462 provides In addition, by asserting the timing control signal the microprocessor 104 enables the amplifier 462 during periods of data transmission. On the other hand, when the speaker/microphone 114, which has an input coupled to the output of the amplifier 462, is being used as a receiving device or microphone, the microprocessor 104 de-asserts the timing control signal thereby deactivating the amplifier 462 and thus the output of the DTMF tone encoder 110. The timing control signal may also be used to inhibit signal output during the interdigit periods.

Various features of the present invention, directed to overcoming the data error, security and convenience problems associated with known auto-dialer devices will now be described with reference to the auto-dialer 100.

Each feature of the present invention will be discussed in detail below beginning with a discussion of the features of the present invention which are directed to reducing the error rate associated with the acoustic transmission of information represented by DTMF signal tones, e.g., telephone number and credit card number information, from the auto-dialer 100 of the present invention to a DTMF signal receiver such as a telephone handset 121. This particular feature of the present invention may be described as an error avoidance feature.

I. Error Avoidance Features

In accordance with the present invention, several methods are used to avoid or compensate for the occurrence of errors commonly associated with the acoustic transmission of a DTMF signal to a standard telephone system, e.g., a handset. These methods are directed to eliminating, or compensating for, common sources of errors that are associated with acoustic transmission of a DTMF signal. The methods of error avoidance of the present invention will generally be discussed according to the source of the error and the particular method of the present invention directed to eliminating or compensating for such error sources.

A. Third Tone Errors

One of the most common sources of errors associated with known acoustically coupled auto-dialers is generally referred to as the "third tone" problem. This problem is, as the name suggests, associated with the detection of a third, otherwise valid tone, at the detector stage of a receiver where a DTMF tone signal is being decoded. As described above, a DTMF tone signal is only considered valid if it includes a single pair of valid tones, i.e., one valid Hi-band tone and one valid Lo-band tone. Thus, when multiple valid Hi-band or Lo-band tones are received at the same time, the DTMF signal is considered invalid and can not be properly decoded. The relative amplitude of a received tone compared to the other received tones may, in some cases, be used to distinguish valid tones from erroneous invalid tones.

When used as collectors of DTMF tones, carbon based microphones, which are commonly used in standard telephone handsets because of their low cost and high degree of reliability, often generate and transmit erroneous tones, e.g., third tones, in addition to the tones actually received by the microphone. Such errant third tones can cause errors in some tone detection receivers, and particularly those systems which do not utilize digital signal processing equipment for tone detection.

The transmission of DTMF tone signals through carbon microphones causes the carbon granules within the carbon microphone to vibrate in relation to the driving frequencies. As a result of the harmonic effect of the varying vibrations of the granules, various residual tones are generated, with the third tone being the most powerful of these residual tones. This third tone can be relatively powerful, e.g. as much as one half the power level of the higher of the two received acoustic DTMF tones passing through the microphone. The frequency of this third tone will normally be the arithmetic difference between the frequencies of Hi-band and Lo-band tones being received by the carbon microphone.

Referring now to FIG. 8A, there is illustrated a chart with four columns. In the chart, the number or symbol in the first column 11 represents those numbers or symbols available on a standard telephone keypad. The second column 13 represents the Lo-band frequency associated with the corresponding number or symbol in the first column 11 while the third column 15 represents the corresponding Hi-band frequency. The fourth column 17 represents the fundamental difference between the Lo-band and Hi-band frequencies listed in columns two and three 13, 15. It is this fundamental difference in frequency that represents the frequency of the third tone that is generated by a carbon microphone when the corresponding Lo-band and Hi-band frequencies are received. As will be discussed further below, the fundamental difference frequency, or third tone frequency associated with the symbols/numbers 3, 6, a, b, c, and d will fall within the passband of the filters of many standard DTMF detectors.

The third tone error problem is particularly significant with regard to the tone pairs representing the numbers symbols 3, 6, a, b, c, and d because, in their case, the third tone created from the harmonic effect associated with the carbon microphone falls within the tolerance range of the low frequency band tones and the band pass filters corresponding to these frequencies. Thus, in the case of these numbers and symbols, the third tone signal will fall within the passband of the filters of many standard DTMF detectors and will therefore not be filtered out.

As discussed above, the presence of such an errant tone within the frequency range of valid tones, may cause a detector to detect two valid Lo-band tones, when only one should be present. While the power level of the deliberately generated tone will normally be much higher than the errant signal tone, without the use of digital signal processing which is able to select the tone with the higher power level, a detector will have difficulty in determining which of the two Lo-band tones received should be used. Normally, when the DTMF detector is unable to select the valid or deliberate tone from the two Lo-band tones, the DTMF tone detector will ignore the tone-pair signal which includes the third tone, causing the entire dialed string to be lacking the errant digit and, thus, preventing the connection of the call or the completion of a data string being used for other purposes.

While the third tone will not always be of such an intensity that it results in an error, and while digital signal processing in DTMF detectors is becoming more common, for an acoustically coupled auto-dialer 100 to offer maximum versatility it must be capable of transmitting telephone and credit card number information accurately to the vast majority of existing telephone systems including those that do not perform such digital signal processing. Accordingly, the third tone problem associated with carbon microphones needs to be reduced or compensated for to increase the reliability of acoustically coupled auto-dialers if such auto-dialers are to work reliably with the vast majority of existing telephone systems.

The present invention, addresses the third tone problem in two ways. The first approach is directed to avoiding using symbols/numbers which are likely to produce third tone errors. The second approach is directed to altering the nominal frequency of the tones which are likely to generate third tone errors in an attempt to avoid such errors.

i. Avoidance of the Use of Symbols/Numbers Likely to Produce Third Tone Error Problems

As discussed above, the first approach of the present invention for dealing with the third tone problem revolves around avoiding the use of numbers/symbols that are most likely to produce the problem in the first place, i.e., the numbers/symbols
3, 6, a, b, c, and d. Because most credit card, calling card, and telephone numbers use only the digits 0 through 9 found on standard, non-extended keypads such as those found on pay phones, most third tone problems can be eliminated by merely avoiding the use of the numbers 3 and 6.

Thus, in accordance with one embodiment of the present invention, when assigning numbers, e.g., telephone numbers which must be used to obtain connection, e.g., via a local telephone switching office which may not contain digital signal processing equipment, to a central office, only the digits 0, 1, 2, 4, 5, 7, 8, and 9 are used. In this manner, third tone problems associated with the numbers/symbols 3, 6, a, b, c, and d are avoided when sending information to telephone switching offices which may not have digital signal processing equipment capable of distinguishing the actual tone from the undesired third tone signal.

Once a connection has been made to a telephone switching network with digital signal processing equipment, such as the type currently found in most long distance carrier telephone switching offices, the risk of errors due to third tone problems will be greatly reduced. Thus, all digits may be used including 3, 6, a, b, c, and d once a connection has been established to a system known to include digital signal processing equipment. Accordingly, the primary time for avoiding the digits associated with third tone problems is when establishing connections to local offices or other telephone switching networks which may not contain the digital signal processing circuitry required to avoid third tone problems.

By avoiding the use of the numbers/symbols 3, 6, a, b, c, and d in the above described manner, the vast majority of third tone problems can be avoided without the need for digital signal processing circuitry in a DTMF detector and without modifying the DTMF signal generator.

ii. Generation of Tone Pairs Wherein the Arithmetic Difference of the Generated Tones Comprising Each Tone Pair Will Fall Outside the Range of the Bandpass Filters Included in Standard DTMF Detectors

The second approach of the present invention to avoiding third tone error problems involves the generation of tone pairs wherein the arithmetic difference between the generated tones comprising each tone pair will fall outside the range of the bandpass filters included in standard DTMF generators.

Because of manufacturing tolerances and component differences, DTMF generators will vary slightly in output frequency from one DTMF signal generator to another even though the nominal frequencies, which represent the frequencies the DTMF generators are designed to produce, will normally be the same. Thus, DTMF detectors are designed to detect, i.e., accept as valid, a range of frequencies corresponding to the Lo-band and Hi-band DTMF tone frequencies.

Referring now to FIG. 8B, there is illustrated a chart of the frequency accept range of a standard DTMF detector circuit. As illustrated, the accept range is .+-.1.5%+2 Hz of the nominal frequency illustrated in the center column of the chart shown in FIG. 8B. Referring now to FIG. 8C there is illustrated a chart of the frequency reject range of a standard DTMF detector circuit. As illustrated, the reject range is .+-.3.5% of the nominal frequency illustrated in the center column of FIG.
8C. Referring now to FIG. 8D, there is illustrated a chart of the out of range frequencies of a standard DTMF detector circuit. These frequencies represent frequencies outside the accept/reject ranges of a standard DTMF detector which cannot be recognized by a standard DTMF detector circuit in a reliable manner. Referring now to FIG. 8E, there is illustrated a composite chart of standard DTMF tone detector circuit reject and accept ranges. In particular, the chart of FIG. 8E illustrates the standard accept/reject ranges associated with each of the eight nominal tone frequencies listed in the center column of the chart.

Because the Lo-band and Hi-band tones which will be accepted by a DTMF detector are permitted to vary over a limited range, e.g., as illustrated in the charts of FIGS. 8A-8E, it is possible to control the generation of Lo-band and Hi-band tones so that acceptable tones are generated while the arithmetic difference between the tones of any generated tone pair will be such that it will fall outside the accept range of a standard DTMF detector.

This may be done, by, e.g., providing the DTMF tone generator which produces tones having a nominal frequency that is closer to the outside limits of the acceptable frequency range of those symbols/numbers that cause third tone problems. For example, in one embodiment the auto-dialer 100 is designed to generate Lo-band frequency and Hi-band frequency tones, for the tone pairs representing, e.g., the digits 3 and 6, that will fall within a standard DTMF detector's accept range, e.g., the accept range of a MITEL.TM. model number MT8870D DTMF detector circuit, but will have an arithmetic difference that falls outside the detector's accept range. This can be achieved by selecting the nominal center frequencies of the DTMF tones generated by the auto-dialer 100 of the present invention, for the tones used to represent the numbers 3 and 6, towards the outer edge of the "accept range" of standard DTMF detector circuits. It should be noted that in accordance with the present invention, the microprocessor 104 can be programmed to select and control the generation of DTMF tones of various tone pairs, so that the tones of a tone pair will fall within the accept range of conventional detectors, but create an arithmetic difference which is outside the tolerance range of such detectors.

Referring now to FIG. 8F, there is illustrated a graph of the high tone output levels from carbon and electret telephone microphones with 100-102 dB (SPL) input for the twelve acoustically dialed digits 0-9, * and #. As illustrated, the output levels for each acoustically dialed digit is different depending on whether a carbon or electret microphone is used. In the graph, the output level varies from approximately -6 dBm to approximately -3 dBm for electret microphones and from approximately -13.5 dBm to approximately -9 dBm for carbon microphones.

Referring now to FIG. 8G, there is illustrated a graph of the low tone output levels from carbon and electret telephone microphones with 100-102 dB (SPL) input for the twelve acoustically dialed digits 0-9, * and #. As illustrated, the output levels for each acoustically dialed digit is different depending on whether a carbon or electret microphone is used. In the graph, the output level varies from approximately -6 dBm to approximately -3 dBm for electret microphones and from approximately -13.5 dBm to approximately -12 dBm for carbon microphones. The relative intensities of the high and low tones vary according to which digit is encoded. The output levels of the carbon microphone low tones vary less than -2 dBm among the twelve different acoustically dialed digits, whereas the output levels of the carbon microphone high tones vary approximately 4 dBm among the twelve different acoustically dialed digits.

Referring now to FIG. 9A, there is illustrated a chart indicating the Lo-band frequency and Hi-band frequency of the tones which the auto-dialer 100, in accordance with one embodiment is programmed to generate, for each of the listed tone pairs. As will be noted, the values in FIG. 9A vary from those of FIG. 8A as the result of the intentional use of tones which will produce valid DTMF tone pairs while avoiding third tone problems by generating fundamental differences between the Hi and Lo tones that will fall outside the frequency accept range of most DTMF signal detectors.

The significance of the fundamental frequency differences is particularly significant in light of the relative power of the third tone noise (margin) which accompanies the digits 3, 6, a, b, c, d. These noise levels approach the lower acceptable limits of standard detectors, and can, because of their frequency, be interpreted by the detectors as being considered valid tones, thus providing the detector with two Lo tone frequencies to decode.

While the generation of tones towards the outer limits of the acceptable frequency ranges in accordance with the present invention offers a method of reducing third tone problems, it also requires that the DTMF generator of the present invention be more accurate and stable than would otherwise be required to insure that the tone generator only produces DTMF tones that will fall within the acceptance range of a standard DTMF detector circuit. As will be discussed below, calibration features of the present invention help insure that the required accuracy in frequency output is achieved.

In accordance with the second approach to reducing third tone problems, the third tones generated by a carbon microphone will not be interpreted as valid Lo-frequency tones because of their frequency. In this manner, third tone problems are substantially reduced without the need for the DTMF detector circuitry to include digital signal processing circuitry.

It should be noted that while this approach provides a suitable method for eliminating or reducing third tone problems, it requires the auto-dialer 100 of the present invention to generate DTMF tones with a higher degree of accuracy then would otherwise be required. In a device that is designed to incorporate automatic calibration features, such as the auto-dialer 100 of the present invention, it may be possible to achieve the required higher standards at little or no additional cost.

B. Amplitude Variation between Low-band and Hi-Band Frequencies Errors

In addition to errors caused by the presence of a third tone, errors may also be caused by power differences between the Lo-band and the Hi-band tone signals comprising a tone pair. These differences depend largely on the type of microphone used. Referring now to FIGS. 8F and 8G which are charts illustrating exemplary output levels of both carbon and electric microphone, for the high-tones and low-tones, respective, it can be seen that for the carbon microphone in particular, there are significant differences between Hi-band tone signal power output levels and Lo-band tone signal power output levels of many DTMF tone pairs. This difference in signal output levels results in the introduction of twist into the received signal. As can be seen, the twist which represents distortion, that is introduced by carbon microphones can be significant.

As discussed above, each of the numbers/symbols of a standard telephone keypad is represented by a DTMF tone signal, i.e., a tone pair, comprising one Lo-band tone and one Hi-band tone. Furthermore, for such a signal to be detected as a valid signal, the difference, referred to as twist, in the power level between the Hi-band and Lo-band tone signals of any particular received tone pair must fall within a specific range for the signal to be considered valid. The acceptable range of power levels received by standard DTMF receivers in a tone pair requires that the Hi-band tone signal power level is not more than 4 dBm greater or 8 dBm less than the Lo-band tone signal power level. If these power level conditions are not met, a received tone pair will be rejected.

It has been found that carbon microphones tend to be less efficient at converting low frequency sound waves, e.g., acoustic Lo-band frequency tones, into electrical signals then they are at converting high frequency sound waves, e.g., acoustic Hi-band frequency tones, into electrical signals. This disparity in conversion efficiency introduces twist or power level differences into received tone pairs with a predictable bias generally in favor of the Hi-band tone signal.

This introduced twist, resulting from the use of carbon microphones, adds to any twist that may exist in a tone pair signal generated by an auto-dialer. While, in some cases the twist introduced by a carbon microphone may act to counter twist existing in a received tone pair, in other cases it will simply add to the degree of twist. In such a case, it may cause a tone pair which would otherwise have an acceptable degree of twist to be rejected because of the twist introduced by the use of a carbon microphone.

Because the vast majority of public telephones presently in use include carbon microphones (due to their relative ruggedness and low cost), the twist introduced by such microphones presents a potentially significant source of errors for the acoustic transmission of DTMF signals to telephone systems.

To counter this problem, in one embodiment of the present invention, the Lo-band and Hi-band tone signals are amplified separately before being supplied to the audio output device, e.g., speaker. In such an embodiment, microprocessor 104
controls the independent amplification of the Lo-band and Hi-band tones via the amplitude control signals supplied the amplifiers 458, 460 with the Lo-band tones being amplified to a greater extent than the Hi-band tones. The difference in amplification is of such a degree that the auto-dialer 100 is designed to compensate for the varying efficiency of carbon microphones in converting acoustic tone signals of different frequencies into electrical signals.

Accordingly by intentionally introducing twist into the tone pairs, e.g., by the separate amplification of each individual tone in a tone pair being generated by the auto-dialer 100 of the present invention, it is possible to counter the predictable non-linear signal conversion caused by the use of carbon microphones.

In a sense, this approach to error avoidance may be thought of as introducing intentional distortion into the relative power levels of a DTMF tone pair to compensate for the distortion expected to be introduced into the signal upon reception by a carbon microphone.

While the twist introduced by the use of carbon microphones is somewhat predictable, twist introduced by, e.g., line loss introduced by lines coupling the DTMF signal receiver to the microphone receiving acoustic DTMF signals from the auto-dialer
100 are somewhat less predictable. As the result of empirical field tests conducted using a variety of telephones, it has been found that the introduction of certain twist values into a DTMF signal will generally produce better tone recognition then will be achieved without the introduction of twist. The varying amplification of Lo-band and Hi-band tones in the tone pairs being generated is to cause the electrical signal leaving the receiving telephone instrument to be in accordance with industry standards established for optimum performance after the conversion of the received signals into electrical signals.

The amplification of the low tone at a higher gain level than that of the high tone will also tend to cause any resulting third tone to be of a lower level than that which would be achieved by amplifying both signals at a higher rate.

Referring now to FIG. 9B there is illustrated a chart which shows twist values for each of the ten DTMF tone pairs normally used for telephone dialing which have shown to produce satisfactory results under a wide range of field conditions, e.g. both urban and rural telephone conditions.

While FIG. 9B does not include the extended character set (a,b,c,d), it should be noted that these tone-pairs are not included on standard dialing pads and, therefore, are not recognized by common switching systems for the placement of calls. The use of these digits is generally limited to post-access data collection where one can be reasonably certain that digital processing equipment will be available to decode such tones, including their otherwise errant third tones.

While, as described above, it is often desirable to intentionally vary the level of amplification of each tone of a tone pair, the amount of amplification intentionally introduced should be small enough so that the maximum twist detected by a standard receiving DTMF decoder will be such that the Hi tone level is not more than 4 dB greater or 8 dB less than the Lo tone, i.e., such that the twist falls within the range of what is defined as a standard valid DTMF signal. Since different telephone microphones (e.g. carbon, electric) will transmit Hi and Lo tones with varying efficiency, because line losses can vary significantly, and because varying speakers will have different spectral output curves, the determination of optimum twist levels should be based on field tests including the generation of a plurality of twist levels, in conjunction with the use of level detection equipment at a central receiving site to determine optimum twist levels for a particular device or type of speaker that will provide for the accurate detection of the tones by a wide variety of telephones under varying conditions.

Referring now briefly to FIG. 9B, audio output signal levels and twist levels which have provided satisfactory results for the indicated tone pairs are illustrated. As can be seen from the chart in FIG. 9B, the introduction of twist levels of between -8 and +2 dBm, e.g., through the separate amplification of the Lo and Hi tones, has provided satisfactory results.

Referring now to FIG. 9C acoustic output levels of Hi and Lo tones corresponding to the digits 0-9 are shown. As illustrated, in order to compensate for e.g., the uneven energy conversion of Lo and Hi acoustic tones into electrical signals by carbon microphones, it is desirable to generate Lo and Hi tones having different acoustic sound pressure levels. As illustrated, by varying the sound pressure levels by as much as 9 dB (SPL), e.g., for the Lo and Hi tones corresponding to the digit 1
has proven to provide satisfactory results.

It has been found that when testing to determine the twist levels that are appropriate to be used with a device, the determination should be based on the signal characteristics received following transmission through the public telephone network or those received at the line end of one of more telephone instruments, and not simply the characteristics of locally generated/received signals.

As discussed above, the auto-dialer 100 includes separate Hi-band and Lo-band frequency signal generation circuits 401, 403 for independently amplifying both fundamental frequencies of a DTMF tone pair. This amplification of individual Lo-band and Hi-band tones may be done to pre-established, e.g., pre-programmed, and remotely alterable levels for any selected tone-pair. In addition, the auto-dialer's microprocessor 104 and RAM 108 can be used to store and modify parameters used to control amplification levels for each fundamental frequency, i.e., tone, of a tone-pair. As discussed further below with regard to security features of the present invention, the ability to reprogram or modify the stored amplification values may be enabled or disabled in response to the receipt by the auto-dialer 100 of a pre-selected group of tones which can serve as an access key, instruction or reprogramming command.

Just as the amplification levels of the DTMF signal, e.g., tone pairs generated by the auto-dialer 100, can be controlled by stored values or parameters, the auto-dialer 100 can adjust the levels of amplification which are applied to signals received by the auto-dialer's audio transducer, e.g., speaker/microphone 114. In addition, the stored information for controlling amplification levels may be reprogrammed in response to signals received by the auto-dialer 100 with the reprogramming feature being enabled/disabled in response to the receipt of a pre-selected series or group of tones that may be stored in the RAM 108.

C. Errors Resulting from Insufficient Signal Power

In addition to third tones and excessive twist, errors may also result from insufficient signal strength, i.e., power, at the microphone of the receiver. As discussed above, one of the conditions for a tone pair signal being interpreted as a valid DTMF signal is that each tone in the tone pair received at the DTMF detector have a signal power level that falls within the range of 0 to -25 dBm.

As a practical matter, the minimum sound pressure level of an acoustically coupled and generated tone that will be recognized by a standard DTMF detector is a function of the distance of the tone source, e.g., audio output device 114 of the auto-dialer 100, to the inductor, e.g., microphone 118 of a receiving device. In addition, the power level of the received signal will depend on the energy of the signal output by the auto-dialer 100, the directionality of the sound waves, and any apparatus provided to focus the sound waves towards the microphone. Because of these factors, the movement of the speaker/microphone 114, of the auto-dialer 100 away from a microphone during detonation, i.e., output of the tones being generated, can result in a sufficient decrease in the power level of the signal received at the handset's microphone 118 to cause a tone-pair to be rejected.

The present invention is designed to insure that the audio signal received at the microphone 118 of a telephone 122 has sufficient acoustical power that the electrical signal produced therefrom will be detectable as a valid DTMF tone signal.

In one embodiment of the present invention, a proximity detector 228, such as a pressure switch or light sensor, is incorporated into the audio dialer 100 of the present invention to detect the auto-dialer's proximity to a telephone handset's microphone 118. In such an embodiment, when the auto-dialer's speaker 114 is placed in close proximity to the handset's microphone 118, the tone signal output of the auto-dialer 100 is enabled. In the event that a user moves the auto-dialer away from the handset's microphone 118, as indicated by the output of the proximity sensor 228 and detected by the microprocessor 104, the auto-dialer 100 will prevent or cease the output of DTMF tones and indicate, e.g., through the use of a user--noticeable signal, e.g., an audible signal or a message on the display of the auto-dialer 100, that the auto-dialer 100 should be placed closer to the microphone 118. Such a output may be incorporated into the display device 202.

In this manner, tones will be generated only when the speaker/microphone 114 of the auto-dialer 100 is in close enough proximity to the handset's microphone 118 to prevent or limit the number of errors that might otherwise occur due to lower or varying tone characteristics due to a varying or excessive distance between the output of the auto-dialer 100 and the corresponding handset's microphone 118 resulting from a user moving the auto-dialer 100 away from the handset's microphone 118 while tones are being generated.

As discussed above, in order to insure that a user keeps the auto-dialer 100 in close proximity to the microphone of a receiver when pulsing out tones, in one embodiment, as illustrated in FIG. 3, the auto-dialer 100 also includes an audio and/or visual tone output indicator 103, e.g., a light or buzzer, coupled to the DTMF encoder output, to indicate to a user when the auto-dialer 100 is pulsing out tones. In this manner, the auto-dialer 100 provides a signal to the user that the user should keep the auto-dialer 100 in close proximity to the receiver's microphone to avoid errors.

In addition to the use of a proximity sensor 228, the housing of the auto-dialer 100 of the present invention (see, FIGS. 10A-10D) is designed to be relatively small making it easy to visually or manually center the speaker/microphone 114 of the auto-dialer 100 over a telephone handset's microphone 118. Tests have shown that failing to properly align the speaker/microphone 114 with the center of the handset's microphone 118 can result in a wide range in terms of signal intensity as detected by the microphone 116. In addition to being small in size, the housing includes a circular area designed to easily mate with the speaker end of a standard handset. Thus, by the design of auto-dialer's housing, a user can easily align the center of the transducer with the center of an interfacing microphone or speaker so as to provide for the uniform transmission/reception of tones.

Referring briefly to FIG. 10A, a bottom view of the auto-dialer housing 101 is illustrated. As illustrated, the auto-dialer 100 has a housing 101 which has a head portion 130 with audio output openings 150 to allow sound out. From the bottom planar view of FIG. 10A, the head has a generally circular appearance. The head portion 130 is designed to be smaller in diameter than most telephone mouthpieces, e.g., less than 6 cm in diameter. In one embodiment, the circular head portion 130 is designed to be approximately 34 millimeters (mm) in diameter. This is sufficiently small to permit the auto-dialer 100 to be easily centered, by visual inspection, with the microphone of a wide variety of telephone handsets including the handset of a NYNEX public telephone which, in one field test, was found to contain a microphone having a diameter of approximately 6 cm. While other microphones used in mouthpieces will vary in size, the relatively small size of the head portion 130 should permit easy, yet relatively precise, visual and manual alignment, e.g., centering with most telephone handset based microphones in use today.

The ability to visually center the auto-dialer 100 of the present invention with the microphone of a handset is important because mouthpieces on handsets frequently vary in size and shape making it difficult to use a circular gasket or other device to aid in the centering of auto-dialer speakers with the microphone contained in a mouthpiece of an increasing number of telephones, which do not have a locating ring as was the custom with telephones built many years ago.

Thus, because of the shape, e.g., generally circular appearance when viewed from above or below, and small size of the head portion 130 of the auto-dialer 100 it is possible to easily center it using visual techniques which may not be possible using a rectangular shaped housing or other type of housing which makes it difficult to see the handset's microphone 118, and thus its center, when the auto-dialer 100 is placed in close proximity thereto.

Other features of the present invention are also designed to enhance the efficiency of the transfer of and acoustic DTMF signal to the microphone 118 of a telephone handset 121. For example, in one embodiment of the present invention a relatively sound transparent barrier, as opposed to a sound baffling barrier, is used to encircle the area of the auto-dialer's case near audio output openings 150. The barrier is used to reduce the resulting harmonics when the case of the auto-dialer
100 makes contact with an interfacing, e.g., telephone. However, the barrier is not designed to occlude ambient sound from entering the column between the device and microphone.

Testing has shown that when using a baffle in conjunction with an electret microphone it is important to maintain at least a 25% open air flow between the auto-dialer's speaker and the microphone.

Accordingly, while an isolating barrier, such as a gasket, may be used between the auto-dialer's speaker/microphone 114 and the handset's microphone 118 to reduce the level of ambient noise transmitted to the handset's microphone 118, in general, higher device output should be provided in lieu of such an isolating barrier. The use of higher signal output levels is generally more effective than the use of such isolating barriers because there will normally be some ambient noise present during the use of an auto-dialer 100 regardless of the presence of a gasket since, e.g., most handset housings are reasonably good conduits for ambient sound to the microphone 118.

Therefore, while a gasket interface between the auto-dialer 100 and the interfacing microphone may be appropriate to reduce the possibility of harmonics arising from the movement of the auto-dialer 100 on the surface of the microphone, the gasket should not be intended to provide isolation of the generated tones from the ambient noise environment.

The use of high audio output signal levels, e.g., 95 to 114 dB(spl), has proved to present advantages with regard to transmitting a DTMF signal to carbon microphones found in handsets 121. Significantly, testing has shown that such relatively high power levels do not present problems when the auto-dialer 100 is used with ele