Patent Application
20130044867
The present invention generally relates to telecommunications monitoring and, more specifically, to a system and method for monitoring phone calls to and from an inmate within a correctional facility through the use of speech and voice recognition software in order to detect unauthorized call connecting.
There exists a need to control and monitor access to telephone lines in a correctional facility, such as a prison or jail. By and large, such control and monitoring systems are implemented for security considerations. One such security concern is an inmate's attempt to place threatening or harassing phone calls to certain individuals. An early solution to this problem was to limit the inmate's phone access to only allow the inmate to place collect calls in the hope that forcing the inmate to provide his/her identity would give the call recipient an opportunity to deny the call. However, such a solution does not effectively address the issue, as inmates can easily provide a false identity hoping that the recipient will accept the collect call.
The problem of an inmate placing threatening or harassing phone calls is further complicated by the fact that the inmate callers often work in concert with individuals outside of the correctional facility. For example, a certain individual may have a restraining order against the inmate such that the inmate is not allowed to contact that individual in any way, including by phone. However, the inmate may contact an accomplice who then sets up a 3-way call with (or forwards the call to) the individual, thereby allowing the inmate to have an unlawful conversation with the individual. Though the 3-way call or call forwarding may be prohibited by law, it is difficult for the correctional institution to prevent it because the inmate's phone call to the accomplice is lawful and the accomplice has unrestricted telephone access. Therefore, the accomplice's ability to set up 3-way phone calls and/or call forwarding to bridge telephone connections essentially provides the inmate with unrestricted telephone access.
There are a variety of known attempts to solve the problem of unauthorized inmate telephone call connecting. In the context of call bridging, one common solution is to detect the hook flash signal produced when such call bridging is attempted. Pursuant to one known solution, a detection apparatus includes a low pass band filter for passing energy having frequencies below a certain frequency and an energy detector for detecting a specific electrical energy pulse having been passed by the low pass filter. Therefore, the detection apparatus is dependent upon the analog signals generated during the call bridging attempt. Another known solution detects the presence of certain tones within the telephone signal, such as dual tone multifrequency tones, special information tones, dial tones, and other call progress tones. The identification of these analog tones is used to determine if an attempted call bridging has taken place.
As shown through the above discussion, the prior art solutions have been singularly focused on the identification of attempted phone call connecting in the context of standard, analog telephone systems. However, like all technologies, the way people send and receive phone calls is continuing to evolve. More and more companies and individuals are utilizing voice over internet protocol (VoIP) services. VoIP is a general term for a range of communication transmission technologies designed to deliver voice and other forms of communication over IP networks, such as the internet or other packet-switched networks, as opposed to the public switched telephone network (PSTN). Because of the bandwidth efficiency and low costs that VoIP technology can provide, businesses and other institutions (such as correctional facilities) are beginning to utilize VoIP services in place of their traditional copper-wire telephone systems.
In order to allow the voice communications to be transferred over an IP network, the voice communications are converted from an analog signal into a digital format. As a result, the communications transferred and delivered over VoIP do not contain many of the same signals and tones utilized by the prior art systems to detect attempted phone call connecting. Therefore, known systems would be unable to detect inmate call connecting in a VoIP delivery context.
Therefore, there is a need for a system and method for detecting attempted and unauthorized call connecting for inmate communications delivered via VoIP.
The present invention provides an improved inmate call monitoring system and method. The claims, and only the claims, define the invention.
The principles of the present disclosure provide a system and method for monitoring and controlling inmate telephone calls delivered via VoIP or other means. By using voice biometrics software to identify call participant mood and/or voice prints, the disclosed monitoring and control system can determine whether an unauthorized call connecting attempt has been made or when the call is being used for an improper purpose. When such a determination is made, the call can be terminated, recorded, or monitored by an operator, depending on the preference of the correctional facility and/or to whom the inmate has called.
In one aspect of the present disclosure, a method for managing and controlling telephone activity in a correctional facility is provided. The method comprises of providing a first communicative connection between a caller and a recipient and delivering the conversation between the caller and the recipient over the first communicative connection. The method continues by executing voice biometrics software to evaluate the conversation. A detection response is then executed based upon the evaluation of the biometrics software.
It is an object of certain embodiments of the present disclosure to provide an improved inmate call monitoring system and method.
Further forms, objects, features, aspects, benefits, advantages, and embodiments of the present invention will become apparent from a detailed description and drawings provided herewith.
For the purpose of promoting an understanding of the principles of the invention, reference will now be made to the embodiments illustrated in the drawings and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of the invention is thereby intended. Any alterations and further modifications in the described embodiments, and any further applications of the principles of the invention as described herein are contemplated as would normally occur to one skilled in the art to which the invention relates. One embodiment of the invention is shown in great detail, although it will be apparent to those skilled in the relevant art that some features that are not relevant to the present invention may not be shown for the sake of clarity.
As illustrated, monitoring and control system 20 comprises a voice biometrics module 22, a memory 24, a controller 26, and a database 28. Each of these components is communicatively and operatively connected to one another. The voice biometrics module 22 comprises computer software designed to identify, among other things, unique voice prints and mood(s) of those participating in the telephone conversation. As used throughout this application, a “voice print” is understood to refer to the unique spectrogram of a particular voice. Voice biometrics module 22 may exist as a single module (as depicted) or may in other embodiments be separate and distinct components. When executed, the voice biometrics module 22 can identify call participant mood and/or voice prints in either a real-time fashion, i.e., during the on-going call, or in a post-processing fashion, i.e., communications stored into memory 24. Examples of the software utilized by voice biometrics module 22 is the Voice Print Analysis System available from AVM Software Ltd having a development office in the Netherlands, the CMU Sphinx speech recognition engines available at http://cmusphinx sourceforge.net/, a project by Carnegie Mellon University in Pittsburg, Pennsylvania, and the Modular Audio Recognition Framework, an open-source research platform available at www.marf.sourceforge.net.
Memory 24 is constructed and arranged to optionally store the digital voice communications received from and delivered to inmate telephones 12, 14. In some embodiments, memory 24 is a standard memory component, such as, but not limited to, a hard disk drive, RAM, or ROM, or any other known electrical storage device.
Controller 26 is constructed and arranged to dictate and monitor the operations of monitoring and control system 20 and execute an appropriate detection response if it is determined that unauthorized call connecting has been attempted. As used herein, unauthorized call connecting is understood to mean call bridging, call forwarding, and/or any other known means for creating a communicative connection between one party of a telephone call and a third party to the telephone call. Certain call participant moods, the number of unique voice prints identified during a call, and the presence of prohibited voice prints may indicate that one party is attempting, or has attempted, to initiate unauthorized call connecting with a third party.
Controller 26 controls the recording of the digital voice communications received from and delivered to inmate telephones 12, 14 onto memory 24. Controller 26 may instruct the memory to record an entire phone call, a portion of the phone call, or to act as a memory buffer, recording a predetermined portion of the call before a potential issue is recognized by the voice biometrics module 22. The controller 26 may also be operational to terminate the connection between an inmate telephone 12, 14 and an outside phone line. Additionally, the controller may be operational to decide when the voice biometrics module 22 is to be executed.
Database 28 stores and maintains mood indicators, inmate and third-party voice prints, and voice print thresholds which are indicative of a call connecting attempt. As illustrated, a single database 28 is provided within monitoring and control system 20. In such a situation, the database 28 is constructed and arranged as a global database against which caller moods and/or voice prints in all inmate telephone conversations are evaluated. In other embodiments, the database 28 may be constructed and arranged as separate databases specific to each individual inmate.
Monitoring and control system 20 is also communicatively coupled to a hub 40. Hub 40 is configured to communicatively and operationally connect correctional facility computers 42, 43, 44 with monitoring and control system 20. When system 20 identifies an unauthorized call connecting attempt, system 20 is constructed and arranged to notify the correctional facility computers 42, 43, 44 via email, audio and/or visual indication. In turn, computers 42, 43, 44 can access memory 24 to allow the computer operators to listen to the recorded phone calls. In some embodiments, the operators of computers 42, 43, 44 can listen to the telephone conversations live. The correctional facility computers 42, 43, 44 may also have the capability of updating database 28 with additional names, words, numbers, voice prints, or moods which will flag attempted and unauthorized call connecting. While three correctional facility computers are depicted, other embodiments of the present disclosure may comprise fewer or more than three correctional facility computers.
While voice biometrics module 22, memory 24, controller 26, and database 28 are depicted as residing within the same location, it is within the scope of the present disclosure that the components of system 20 do not need to be physically proximate. For example, the database 28 may be physically stored remotely from the other components and accessed via the internet. Monitoring and control system 20 and VoIP service provider 32 are connected to the internet 30 through known techniques.
A monitoring and control system 70 is communicatively and operationally connected to the internet 65. Monitoring and control system 70 is constructed and arranged to receive the VoIP communication from ATA 60 and relay that communication to outside phone lines 80, and vice versa. The telephone communications sent to and received from outside phone lines 80 may be in VoIP or standard, analog format.
Like the monitoring and control system 20 of
A hub 85 is configured to communicatively and operationally connect correctional facility computers 87, 88 and 89 with the internet 65. When system 70 identifies an unauthorized call connecting attempt, system 70 is constructed and arranged to notify the correctional facility computers 87, 88, 89 through a notification sent via the internet 65. Through that connection to the internet 65, computers 87, 88, 89 can access memory 74 to allow the computer operators to listen to the recorded phone calls. In other embodiments, the operators of computers 87, 88, 89 can listen to the telephone conversations live. Similarly, the correctional facility has the capability of updating caller moods and/or voice prints which will flag attempted and unauthorized call connecting. In a broader sense, the components of the monitoring and control system 70 may be configured remotely by the correctional facility computers 87, 88, 89 by known techniques, including, but not limited to, GUI or CLI-type interfaces.
ATA 60, monitoring and control system 70, and hub 85 are connected to the internet 65 through known techniques. As depicted in
ATA 60 may be used with the telephones to convert the analog communications into a digital format. In other embodiments, the individual telephones are VoIP devices thereby eliminating the need for an ATA within the system.
At some time later, the voice biometrics software is run (step 108). As previously discussed, the voice biometrics software identifies the mood(s) of the call participants. Moods indicative of an unauthorized call connecting attempt include anger, frustration, and fear, just to name a few.
While analyzing the conversation, the control and monitoring system will determine if the inmate's mood is a possible concern (decision step 110). In another embodiment, the mood of the other call participant(s) is evaluated as well. If no concerning mood is identified, then the stored call recording may be deleted (step 112). If a concerning mood is identified, then that is indicative of an unauthorized call connecting attempt and a detection response may be executed. At step 114, the phone call record is maintained. At step 116, the correctional facility is notified of the possible unauthorized call connecting attempt.
At some time later, the voice biometrics software is run (step 148). As previously discussed, the voice biometrics software is capable of identifying a unique voice print for each call participants. While analyzing the conversation, the control and monitoring system will determine the number of voice prints identified (decision step 150). If two voice prints are identified, then the stored call recording may be deleted (step 152). If more than two voice prints are identified, then that is indicative of an unauthorized call connecting attempt and a detection response may be executed. At step 154, the phone call record is maintained. At step 156, the correctional facility is notified of the possible unauthorized call connecting attempt. In other embodiments, the threshold number of voice prints indicative of an unauthorized call connecting attempt may be changed to three or more.
That comparison determines if the number of voice prints is greater than the voice print threshold (decision step 176). In the illustrated embodiment, the voice print threshold is two. If the number of voice prints identified is greater than two, then that is indicative of an unauthorized call connecting attempt and a detection response may be executed (step 178). As used throughout this application, the detection response is understood to be any desired action the correctional facility determines to take including, but not limited to, terminating the phone call, alerting the correctional facility personnel of the unauthorized call connecting attempt, archiving a recording of the conversation, routing the call to a live operator, disabling the inmate telephone (i.e., the mouthpiece and/or earpiece) while the possible unauthorized call connecting is being investigated, flagging the call as potential fraud, blocking future calls to be placed to that number by the inmate, etc. For the call blocking, the monitoring and control system may optionally prohibit a particular inmate from placing a call to the telephone number in which an unauthorized call connecting attempt has previously been identified. The call blocking may be applied to an inmate for a particular phone number for a specified period of time or permanently.
If the number of voice prints identified is two or less, then the monitoring and control system determines if the telephone call has been ended (decision step 180). If the call has not been ended, then the conversation will continue to be monitored (step 174). If the call has been ended, the connection between the caller and the recipient is terminated (step 182).
The voice biometrics software identifies the number of unique voice prints present during the call and allows the call to be monitored (step 198). As the number of voice prints are identified, they are compared to a voice print threshold. That comparison determines if the number of voice prints is greater than the voice print threshold (decision step 200). In the illustrated embodiment, the voice print threshold is two. If the number of voice prints identified is greater than two, then that is indicative of an unauthorized call connecting attempt and a detection response may be executed. Therefore, at step 202, the buffer memory is stored as a call record memory. Further, at step 204, the remainder of the phone call is optionally recorded. The buffer memory and the recording of the remainder of the phone call constitute a complete call record.
If the number of voice prints identified is two or less, then the monitoring and control system determines if the telephone call has been ended (decision step 206). If the call has not been ended, then the conversation will continue to be monitored (step 198). If the call has been ended, the connection between the caller and the recipient is terminated (step 208) and the buffer may optionally be deleted (step 210).
In this embodiment, the call is monitored to evaluate whether an inmate voice print is identified which does not correspond with the name provided in the input step (decision step 230). Again, the database of inmate voice prints is utilized. If a different inmate voice print is identified, then that is indicative of an unauthorized call connecting attempt and/or improper phone usage and a detection response may be executed (step 232).
If a different inmate voice print is not identified, then the monitoring and control system determines if the telephone call has been ended (decision step 234). If the call has not been ended, then the conversation will continue to be monitored (step 228). If the call has been ended, the connection between the caller and the recipient is terminated (step 236).
In this embodiment, the call is monitored to evaluate whether a third party voice print is identified which is associated with a third party the inmate placing the phone call is prohibited from calling (decision step 260). If a prohibited third party voice print is identified, then that is indicative of an unauthorized call connecting attempt and/or improper phone usage and a detection response may be executed (step 262).
If a prohibited third party voice print is not identified, then the monitoring and control system determines if the telephone call has been ended (decision step 264). If the call has not been ended, then the conversation will continue to be monitored (step 258). If the call has been ended, the connection between the caller and the recipient is terminated (step 266).
If the dialed telephone number is not found in the database, the phone call is placed (step 286) and at least one of the caller and recipient are notified that the call will be recorded (step 288). At step 290, the conversation between the inmate caller and the outside recipient is recorded in digital form. At step 292, the telephone call is completed.
At some time later, the voice biometrics software is run (step 294) and the control and monitoring system will determine if a potential issue is identified (decision step 296). Potential issues include, but are not limited to, the determination of more than two voice prints during the call, call participant mood, unauthorized voice prints detected, etc. If no potential issue is identified, then the call recording may be deleted (step 298). If at least one potential issue is identified, then that is indicative of an unauthorized call connecting attempt and a detection response may be executed. At step 300, the phone call record is maintained. At step 302, the correctional facility is notified of the possible unauthorized call connecting attempt.
The system and method of the present disclosure allows the attorney-client conversation to be “passively monitored”. For the purposes of the present disclosure, the term “passively monitor” is intended to mean that conversation is not capable of being listened to by an individual outside of the attorney-inmate relationship and no recording of that conversation is made or maintained. However, the conversation is monitored solely by the voice biometrics software to identify call participant mood(s) and/or voice prints. In turn, the monitoring and control system is executed to determine if the call participant mood, number of call voice prints, or particular call voice prints suggest that an unauthorized call connecting attempt has been made. The monitoring and control system can therefore terminate the phone call to thwart such a connecting attempt, or execute other appropriate detection responses.
Such a process is depicted in
If the dialed telephone number is found in the database, the conversation between the inmate and the outside recipient is passively monitored (step 326). In one embodiment, the passive monitoring is performed pursuant to process 168 as depicted in
It will be appreciated by those skilled in the art that all of the presently disclosed embodiments may also be implemented for calls originating outside of the correctional facility and placed to an inmate. It will also be appreciated by those skilled in the art that the control and monitoring systems presently disclosed would be equally applicable to phone calls being delivered primarily over a PSTN rather than a VoIP communication system. In some embodiments, the control and monitoring system can tap into the analog phone signal and convert the communications into a digital format, at which point the various components of the control and monitoring system may be implemented as disclosed herein.
While the invention has been illustrated and described in detail in the drawings and foregoing description, the same is to be considered as illustrative and not restrictive in character, it being understood that only the preferred embodiments have been shown and described and that all changes and modifications that come within the spirit of the invention are desired to be protected. It is also contemplated that structures and features embodied in the present examples can be altered, rearranged, substituted, deleted, duplicated, combined, or added to each other. The articles “the”, “a” and “an” are not necessarily limited to mean only one, but rather are inclusive and open ended so as to include, optionally, multiple such elements.
