DE102006058758B4 - Method and device for controlling a telecommunication terminal - Google Patents

Abstract

A method of controlling a telecommunication terminal (200, 201, 202) comprising the steps of a) registering at least two input signals generated by different first input means (11-17), b) converting the two input signals generated by different first input means (11-17) into one common first record, c) automatically determining a communication ready state of a user of the telecommunications terminal (200, 201, 202) by evaluating the first record (10) in response to predetermined variable rules (30), d) monitoring the communication readiness state of the user e) performing a predetermined function of the telecommunications terminal (200, 201, 202) upon detection of a change in the communication ready state of the user, characterized in that the input signals z for detecting a communication request at least one of the following signals comprise: - a transit time difference (43) from acoustic signals to microphones located spatially on the telecommunication terminal equipment that is equal to or near zero within a specified tolerance, or - if by a predetermined sensor - by a motion sensor detected movement of the telecommunications terminal the user's face appears in the camera image, or by evaluating sensors and / or microphones in smart clothing, the relative position of the terminal is detected to the user.

Description

The invention relates to a method and a device for controlling a telecommunication terminal.

In telecommunications, different methods are used to simplify or improve the operation of terminals and the voice quality for the user. For this purpose, a hands-free function is integrated in terminals, for example, or measures are provided for suppressing background noise.

Out EP 0 930 761 A2 Furthermore, for example, a phone with adaptive speed dialing method, the phone updates the speed dial memory, for example, depending on the call history, for example, frequently called phone numbers added to the speed dial memory and rarely called phone numbers are removed from this again or the phone numbers in the speed dial memory in dependence the number of times they are called.

The operation of terminals can also be simplified by the use of macro commands. A corresponding method for controlling mobile radio terminals, for example, in DE 101 17 654 A1 described, with macro commands are used to control the mobile terminal that can be created by the user or downloaded from the Internet or the mobile network and stored on the mobile station or its smart card and used for execution.

In order to simplify the operation, also control functions are increasingly provided in a manner that they are non-contact executable because, for example, in certain environments such as in a vehicle manual operation of a telecommunication device by a user is not possible or not desirable. As an alternative to manual operation, the operation can take place, for example, by means of voice commands if appropriate methods for speech recognition are provided.

In EP 0 141 289 A2 For example, a subscriber station apparatus is described in which voice signals inputted via a microphone are directly used for telephone dialing, for which a selection control is provided, which consists of a check and switch logic, a selection control computer, a telephone register memory, a display unit, an input keyboard and a bidirectional Bus system consists, wherein the bus system, the said components of the selection control connects to each other and is connected via a first interface with a speech recognition device, a second interface with a data carrier unit and an optional and Sprachchsignalereinrichtung with a subscriber line, wherein the telephone register memory and / or the data carrier unit are provided for the reception of a subscriber vocabulary and the associated acoustic reference pattern data, and wherein the microphone, the speech recognition device and the election and Sprachignigna Lumsetzereinrichtung are connected via signal lines to the testing and switching logic.

Out DE 101 30 852 A1 Furthermore, a voice-controlled remote control system for electrical devices is known, which comprises at least one handset and at least one base unit, wherein the base unit is designed for controlling the electrical devices in response to coded control signals, which are transmitted from the handsets to the base station. The coded control signals are generated by the handheld devices by pattern matching of voice commands input via a microphone of the respective handset and data stored in a voice pattern memory and by assigning voice recognition to a respective coded control signal, wherein the assignment of voice recognition to the coded control signals in the handset by means of a mapping electronics is individually adjustable, wherein a speaker-dependent voice pattern recognition may be provided in the handsets.

In DE 102 51 112 A1 A method and a system for speech recognition is described, in which a speech signal of a user for detecting a speech signal contained in the speech signal is analyzed and converted within a test procedure the most likely matching detection result again into a speech signal and for verification and / or correction to the user is issued. In the analysis, a number of alternative recognition results are generated which match the next largest probabilities to the speech signal to be recognized, the output being made within the test procedure such that the user can interrupt the output in case of an error in the output recognition result, and then automatically for a section of the respective recognition result last output before a break, the respective sections of the alternative recognition results are output for selection by the user. Preferably, for this purpose, during the output of the recognition result, the voice activity of the user is permanently monitored. In a preferred embodiment, all alternatives Recognition results output together with an indicator, and the selection is made by entering the indicator, for which purpose the respective indicator a button signal of a communication terminal can be assigned and the selection is made by pressing the corresponding button.

In DE 10 2004 029 873 B3 Furthermore, an automatic speech dialogue system is described, wherein the speech recognition result of a Nuterzäußerung is stored and output for confirmation by the user, and wherein a procedure is proposed for situations in which the user denies the correctness of the speech recognition result while mentioning a corrective command.

Out US Pat. No. 6,263,216 B1 a voice-controlled radiotelephone intended in particular for use in a motor vehicle is known in which background noise is constantly detected and analyzed in order to determine a mean weighted energy spectrum of the background noise which is subtracted from the detected energy spectrum when the speech recognition is carried out, the speech recognition optionally being selected by pressing a key is initiated or automatically when the detected energy spectrum exceeds a predetermined threshold.

There have also been attempts in the past to interpret portions of a speech recognizer within a system by defining an activation term to be spoken, such as "wake up", to evaluate further utterances for a particular period of time. Such an activation expression is also commonly referred to as a "hot word". However, misinterpreted background noise as well as similar expressions of the user or other people often led to unwanted activations or deactivations of corresponding parts of the speech recognizer.

For the control of a terminal may also be relevant in which position the user is relative to the device. Out DE 101 56 954 B9 In this context, a method for communication is known in which the relative position of a speaker within a vehicle can be determined with the aid of a permanently installed microphone array in order, for example, to preclude disturbing noises. Also, cameras that are typically provided today in mobile terminals such as mobile phones can be used to detect the position of a user to this terminal within the detection angle using appropriately calibrated software.

For example in US 2002/0167392 A1 a motion-activated voice-controlled communication device is described, which is designed for storing messages and which comprises a motion detector and a device for speech recognition, in the presence of stored messages upon detection of a movement is an automatic announcement that messages are available for interrogation, which then by means of Voice commands can be queried.

Out US 5 255 341 A Further, a command input device for a voice-controlled elevator system with a presence detector is known, which detects the presence of a user at a predetermined distance to a microphone, a voice recognition unit a recognized voice command the elevator control as soon as the user no longer detected by the presence detector.

In EP 1 117 076 A2 There is further described a self-service terminal having an acoustic interface for receiving and / or outputting acoustic information, which is formed as an array of individually controllable acoustic elements, and wherein the acoustic elements are controlled by a control unit in response to user location information which receives the control unit from a user locating mechanism to focus each acoustic element on the location of the user. The acoustic elements are microphones or loudspeakers which can be controlled in such a way that only acoustic signals are received in the vicinity of the user or audible signals can only be heard in the local vicinity of the user. The user location mechanism is preferably based on visual capture, but alternatively, an audio capture mechanism may be provided, for example.

Out DE 101 56 954 A1 is known for the hands-free, for speech recognition and for the occupant communication systems, especially in the vehicle, advantageous image-based adaptive acoustics, with individual, distributed microphone / loudspeaker arrangements a plurality of microphones or microphone arrays and speakers or loudspeaker arrays are adapted such that already before Voice activity, an initialization of the array is achieved on the speaker, the acoustic detection is connected to the image processing in order to obtain a better assignment and recognition of the speaker.

In US 2002/0152010 A1 A multimodal identification system is described which serves to recognize persons authorized to activate one or more vehicle devices, such as automatic locks, one Vehicle engine, lighting or a vehicle alarm.

Out US 2005/0134117 A1 For example, an interface for controlling a plurality of in-vehicle devices is known, wherein the device to be controlled is specified by a combination of a gesture and a voice input.

Furthermore, it is known to equip mobile terminals with movement and inclination sensors in order to detect the orientation of the device in the spatial axes, static and dynamic in the movement. Appropriately equipped mobile phones are known in which such sensors are used directly to perform certain control functions. Furthermore, from the DE 202 21 078 U1 For example, a equipped with tilt sensors mobile phone, in which different functions can be performed with the same button with different inclination of the device.

Although the methods known from the prior art thus allow the determination of the position of a speaker within a room or the orientation of a device. However, this information is not suitable for being able to decide with sufficient certainty whether a user wishes to communicate.

The present invention is therefore based on the object to propose a method and a device, with the aid of a turning or turning away a user to a telecommunication terminal can be reliably detected.

This object is achieved by a method according to claim 1, and a telecommunication device according to claim 13. Advantageous embodiments and further developments are described in the respective subclaims.

Accordingly, a method according to the invention for controlling a telecommunication terminal device comprises registering at least one first input signal generated by a first input device, generating a first data record from this at least one first input signal, and automatically determining a communication readiness state of a user of the telecommunication terminal device by evaluating the first data record depending on given variable rules. To be able to detect a change in the communication readiness state, it is monitored. If a change in the communication readiness state of the user is detected, a predetermined function of the telecommunication terminal device or at least one device associated with the telecommunication device via a network is automatically executed. In the simplest case, the communication readiness state of a user consists in the presence or absence of communication readiness of a user.

The monitoring of the readiness for communication state is preferably carried out by registering at least one first input signal at predefined times, generating a first data record therefrom, and by automatically evaluating this first data record, a ready-to-communicate state is determined. Advantageously, these steps are performed cyclically in a predetermined time interval to ensure continuous monitoring.

The telecommunication terminal device, hereinafter also referred to as terminal for short, can advantageously be designed as a mobile telecommunication terminal such as a mobile phone, smartphone or PDA. Depending on the purpose, however, any other telecommunication-trained device can be used as a terminal. In a preferred embodiment of the method, the terminal is designed, for example, as a permanently installed public telephone or as an Internet station, wherein the Internet station comprises, for example, a publicly accessible access terminal for access to the Internet.

A change in the communication readiness state of a user is usually detected whenever a user turns to the terminal or turns away from it.

The method thus advantageously makes it possible to obtain at least one activation or deactivation signal when a user's turn to or from the terminal is detected, the signal serving to activate or deactivate a predetermined component of the telecommunication terminal.

The automatic execution of the predetermined function of the telecommunication terminal device advantageously corresponds to the activation or deactivation of at least one predetermined component of the telecommunication terminal device. Accordingly, an activation or deactivation signal is transmitted to the corresponding component, for example as a function of the position and the movement of the user in relation to the terminal, namely when the user by his actions with the terminal and / or linguistic utterances his communication request or . expresses the conclusion of the communication. A signal is thus always generated when the user turns to the terminal and starts with at least one local or via a network associated application and / or a remote communication partner to communicate via voice or when the user turns away from the terminal and terminates the communication with the application and / or a remote communication partner, the communication partner may be human and / or machine.

In a preferred embodiment of the method, the terminal is controlled by the user by means of voice control, wherein the function to be executed upon detection of a change in the communication readiness state of the user advantageously comprises activating or deactivating at least one application of the terminal or an application associated with the network Communication of the user with an application program of the terminal is used or supported. For this purpose, the application advantageously comprises at least one speech recognition, echo canceling, hands-free, speaker verification, speaker recognition, speaker classification, voice identification, speech synthesis, noise reduction function and / or at least one function for detecting and evaluating prosodic features of the speech input, such as Example word accent, syllabic emphasis or sentence melody.

When using non-contact input devices such as cameras or microphones, the method advantageously allows the use of terminals without direct contact between the user and the terminal, such as the removal and placing of a handset. Appropriately, corresponding non-contact output devices are provided, such as a screen or speakers.

The detection of a communication request of a user of a terminal is thus in principle by means of fixed, deterministic rules that are stored in the terminal or even in a remote server system, the rules specify which conditions must be present to a "Zuwendungs" - or "avoidance "Signal to generate. The conditions are described by the first data set, which represents input signals registered in each case by means of selected input devices. Examples of usable input devices are microphones, microphone arrays, cameras, tilt sensors, motion sensors, styluses, touch-sensitive screens, mice and / or buttons.

Advantageously, registering the at least one first input signal comprises registering at least two input signals generated by means of different input devices, since a communication readiness of the user can be detected more reliably by evaluating different types of input signals due to more complex rules. The conversion of the registered input signals into a common data set represents a preprocessing in which feature vectors are obtained which advantageously serve to form a statistical behavior model or media model to be trained, with the aid of which after a training or adaptation process when certain static and / or or dynamic input signals a Kommunikationsbeginn- or communication end signal can be generated.

A media model, such as a language model, describes the order in which inputs, such as words and / or sensor signals, are expected by a media recognizer, such as a speech recognizer.

A particularly preferred embodiment of the method accordingly provides to store the first data record created from the first input signals, and to register at least one second input signal by means of a second input device and to generate a second data record therefrom. By evaluating the second data set as a function of predetermined fixed rules, a state of communication readiness of a user of the telecommunication terminal device is determined. Depending on this communication readiness state and / or a change in this readiness for communication state, the variable rules for evaluating the first data record are automatically adapted on the basis of the stored first data record.

The behavior model defined by the variable rules can be pre-assigned before a training or adaptation process by values which already recognize a willingness to communicate for certain input signals.

If a terminal is typically always operated by the same user, substantially repetitive automatic adaptation of the variable rules for that specific user advantageously produces user-specific rules for evaluating the first data record.

On the other hand, it may also be particularly advantageous to generate substantially user-unspecific rules, for example to provide them as a default in a terminal for a new user. For this purpose, there is advantageously a repeated automatic adaptation of the variable rules for a multiplicity of different users.

After completion of individual training processes for a larger number of users of a communication network who have access to terminals that are designed to carry out the described method, the trained models of the individual users are therefore advantageously used to derive from these a user-independent behavioral model, which serves as an initial setting or pre-assignment for further training for new users. After a longer period of use with a large number of users, the training of the behavioral model for generating the "communication request" signal can thus be dispensed with for most new users.

As already described above, the at least one first input signal is preferably generated by means of non-contact input devices in order to enable non-contact control by the user.

The training or adaptation of the variable rules is performed in a preferred embodiment of the method in that upon detection of a change in the communication readiness state of the user based on the evaluated second data set, the variable rules are adapted such that in response to the adjusted rules by evaluating the first Record the same communication readiness state is determined as by evaluating the second record. For this purpose, it may also be advantageous to create the first data set from input signals which were registered shortly before or shortly after registering the second input signal.

Accordingly, the method advantageously provides that the first and second data sets are created from input signals registered over a first or second period, wherein the first period is before or after the second period or the first and second period at least partially coincide.

Conveniently, automatically determining a communication ready state includes determining a probability value or similar parameter reflecting the reliability of the determined state for the presence of a communication readiness and comparing the determined probability value or parameter with a predetermined limit value.

For training or adapting the variable rules, it is advantageous if the communication readiness state can be determined with high reliability from the second data record. For this purpose, for example, a narrowly defined, detectable action, such as pressing a predetermined key, may be provided for the second input signal. Accordingly, registering the at least one second input signal advantageously includes manual user input.

In cases where, in addition to an input result, a high reliability or confidence value is additionally supplied, as described, for example, in US Pat. B. may be possible for certain voice inputs, and this value reflects a high level of security of the detected result, registering the at least one second input signal may advantageously also comprise another than safely evaluated input mode.

In order to reliably determine a communication readiness state for different situations or conditions, the first data record is advantageously created from a plurality of input signals which are registered by means of different input devices. The first data set thus advantageously comprises different input signals in digitized form and thus has a multiplicity of analyzable features.

The evaluation of the first and / or second data set advantageously comprises evaluating a mouse / stylus or keyboard input and / or determining a transit time difference of acoustic signals generated by a microphone array and / or voice recognition of acoustic signals and / or speaker recognition of acoustic signals Signals and / or an evaluation of prosodic features of speech inputs and / or an image recognition of an image taken by a camera and / or a face recognition of an image taken by a camera and / or the evaluation of signals of a proximity or inclination sensor. However, the method described is not limited to the listed evaluation methods, but may, depending on the intended use, include further, also future, evaluation methods.

The evaluation of the first and / or second data record can be done by the latter itself or by a separate processing unit, depending on the design of the telecommunication terminal. For example, in mobile devices, the computing power may be insufficient for the complex evaluation of the first and / or second data set, so that the evaluation is advantageously carried out by means of a remote server that is connectable via a communication network with the terminal. For this purpose, the first and / or second data record is transmitted to the processing unit separate from the terminal, which evaluates the first and / or second data set, monitors the communication readiness state determined by evaluation and, upon determining a change in the communication readiness state of the user Executing the predetermined function transmits a control signal to the terminal.

A telecommunication device according to the invention comprises a telecommunication terminal device, at least a first input device for registering first input signals, a preprocessing unit for converting registered input signals into digitized data, an evaluation unit for evaluating the digitized data, a control unit for controlling the telecommunication terminal device, and a memory for storing variable rules, wherein the evaluation unit is configured to automatically determine a communication readiness state of a user of the telecommunication terminal device, to monitor the communication readiness state of the user by evaluating digitized data obtained from first input signals in dependence on the variable rules stored in the memory and upon detection of a change in the Communication ready state of the user to transmit a control signal to the control unit, wherein the Control unit is designed to perform a predetermined function of the telecommunication terminal in response to the receipt of the control signal.

For adapting the variable rules stored in the memory, the device advantageously has a second input device for registering second input signals. In this embodiment, the evaluation unit is further configured to store first digitized data obtained from first input signals in the memory, by automatically evaluating a user's communication readiness state by evaluating second digitized data obtained from second input signals in dependence on fixed rules stored in the memory the telecommunications terminal to detect and monitor, and depending on the detected from the evaluated second digitized data communication readiness state and / or a change in communication readiness state and the stored first digitized data automatically adapt the variable rules and store the customized rules in the memory ,

For example, in order to be able to evaluate acoustic signals, the preprocessing unit is advantageously designed to generate the first and / or second digitized data from input signals which are registered over a predetermined period of time.

The second input device is preferably for confirming a donation or avoidance of a user to the terminal to enable adaptation of the variable rules and thereby to improve the determination of a communication ready state based on input signals of the at least one first input device. For this purpose, the adaptation of the variable rules advantageously takes place as a function of first input signals which are registered by means of the first input device shortly before, shortly after or during a confirmation by the user by means of the second input device.

Accordingly, the preprocessing unit is advantageously designed to generate the first and second digitized data from first and second input signals, respectively, registered over a first and second time period, the first time period before or after the second time period, or at least the first and second time periods partially agree.

The confirmation of a change of a communication readiness state by means of the second input device by the user may preferably be optional, for example if a probabilistic value for the presence of a communication readiness which is close to a predefined limit value is determined on the basis of the first input signals. In this case, the confirmation by means of the second input device preferably triggers the adaptation of the variable rules. Accordingly, determining a change in the communication readiness state of the user determined by evaluating the second digitized data by the evaluation unit advantageously effects the adaptation of the variable rules.

Preferably, a communication ready state is determined on the basis of a plurality of input signals registered by different input devices. Accordingly, the at least one first input device advantageously comprises at least two first input devices, each of which is designed to register different types of input signals, which advantageously have at least one non-contact input device, such as a microphone, a microphone array, a camera, or a motion detector. or tilt sensor.

The second input device provided for confirming a change in the communication readiness state is advantageously designed as an input device to be operated manually by the user, for example as a stylus, mouse, key or touchscreen, in order to ensure reliable evaluation of the corresponding second input signals. Also, the second input device can be advantageously designed as a microphone to an acoustic signal such as a "Hot Word ", to register. Furthermore, the first input device can be used simultaneously as a second input device.

The evaluation unit is advantageous for evaluating a mouse / stylus or keyboard input and / or for determining a transit time difference of acoustic signals generated by a microphone array and / or for speech recognition of acoustic signals and / or for evaluation of prosodic features of speech inputs and / or Speaker recognition of acoustic signals as a function of deposited speaker profiles and / or for image recognition of an image taken by a camera and / or face recognition of an image taken by a camera depending deposited facial profiles and / or for evaluating signals of at least one proximity or inclination sensor formed.

As described above, determining a communication ready state advantageously involves determining a probability value or similar parameter that reflects the reliability of the determined state. For this purpose, a limit value for this parameter is preferably stored in the memory, wherein the evaluation unit is designed to determine a probability value or a similar parameter for the presence of a communication readiness and to compare the determined value with the stored limit value.

Depending on the specific design and intended use of the terminal, the evaluation unit can be advantageously integrated into the terminal. Alternatively, the evaluation unit is arranged in a separate processing unit, which can be connected to the terminal by means of a communication connection.

Furthermore, in an advantageous embodiment, the first and / or second input device are embodied as input units separate from the telecommunication terminal device, wherein the input units are preferably designed to communicate wirelessly with the telecommunication terminal device, for example by means of Bluetooth.

A preferred embodiment of a telecommunications device comprises a public telephone as a terminal, which is advantageously equipped with non-contact input devices and corresponding output devices. In order to enable communication that can not be overheard by bystanders, the device in this embodiment preferably includes an output device for outputting directional acoustic signals, for example formed as an ultrasonic speaker system, by means of a modulated ultrasonic signal, an acoustic beam to a predetermined spatial area can be directed.

The invention will now be described in more detail by way of example with reference to preferred embodiments and with reference to the accompanying drawings. The same reference numerals in the drawings designate the same or similar parts.

Show it:

1 : a schematic representation of the basic mode of operation of a preferred embodiment of the method according to the invention,

2 FIG. 1 shows a schematic representation of a first preferred embodiment of a telecommunication device according to the invention, FIG.

3 a schematic representation of a second preferred embodiment of a telecommunications device according to the invention, and

4 : A schematic representation of another preferred embodiment of a telecommunication device according to the invention, which comprises a public telephone.

1 schematically shows the basic operation of a preferred embodiment of the method.

At least one first input signal is registered by means of at least one first input device. As usable input devices are in 1 exemplified as a pen or mouse trained pointing device 11 , Touch screen and / or buttons 12 , a microphone array 13 , a camera 14 , a proximity sensor 15 , a motion sensor 17 , as well as a tilt sensor 16 shown. Input signals of one or more of the input devices 11 to 17 become at least a first record 10 converted and pre-processed if necessary. Accordingly, the digital record includes 10 in the illustrated embodiment, a selection of digitized data, the pen / mouse inputs 21 , Keys and / or Touchscreen Inputs 22 , one between different microphones of the microphone array 13 determined transit time difference 23 , an audio sequence recorded by means of at least one microphone as a speech sample 24 , a digital camera image as a picture sample 25 , as well as sensor values 26 the proximity, tilt, and motion sensors 15 to 17 may include.

The record 10 becomes an evaluation unit 40 fed in by evaluating the Record in dependence of given variable rules 30 automatically a communication readiness state of a user of a terminal is detected and monitored. Upon change of the communication readiness state of the user, an activation or deactivation signal is automatically generated and supplied to a control unit of the terminal in order to activate or deactivate a predetermined function of the terminal or of an application associated via a network. This function can be, for example, the establishment or removal of a communication connection with another communication user or the activation or deactivation of an application of the terminal, such as a voice control.

For the evaluation of the first data set are in the illustrated embodiment, various evaluation modules 41 to 49 in the evaluation unit 40 intended. These include, for example, at least one pen / mouse analysis 41 , a button and / or touchscreen evaluation 42 , an evaluation of a transit time difference 43 , a sensor evaluation 44 , a speech recognition 45 , a speaker recognition 46 , an evaluation of prosodic features of speech input 49 , an image recognition 47 , as well as a face recognition 48 , To execute the different evaluation modules 41 to 49 For comparison, user-specific voice or picture profiles can advantageously be stored.

The variable rules 30 which may advantageously include a media model are dynamically adapted in a preferred embodiment.

For this purpose, a first record 10 stored and at least created a second record, wherein the first and second data set is typically created from input signals of different groups of input devices. In this case, input signals are preferably selected for the second data set on the basis of which a communication readiness state of a user is dependent on predefined fixed rules 20 can be reliably detected.

Is by means of the evaluation unit based on the second data set in dependence of the predetermined fixed rules 20 detects a change in communication readiness state of a user, so there is an automatic adjustment of the variable rules 30 depending on the stored first record 10 ,

2 shows a preferred embodiment of a telecommunication device 110 , which as a telecommunications terminal 200 is trained. The terminal 200 can be designed as a mobile or stationary device, for example as a mobile phone, smartphone, PDA, car phone or public telephone. Accordingly, the terminal comprises 200 means not shown, such as a transceiver, for communicating with another communication party via a communication network.

Furthermore, the terminal comprises 200 in the illustrated embodiment, input devices 210 which, for example, one or more of the in 1 illustrated facilities 11 to 17 may include at least one preprocessing unit 220 for converting registered input signals into digitized data, as well as at least one evaluation unit 240 for evaluating the digitized data, which is essentially the evaluation unit 40 in 1 equivalent. The terminal is controlled by the control unit 230 that access a memory 260 in which, for example, a control program is deposited, and suitable output devices 250 controls.

The evaluation unit 240 is with the control unit 230 connected thereto to transmit a corresponding control signal to this upon detection of a change in the communication readiness state of a user. The rules used for the evaluation are also in memory 260 stored. Accordingly, there is also a direct access of the evaluation unit 240 on the memory 260 intended.

In 3 is a further advantageous embodiment of a telecommunications device 120 represented, in which the evaluation unit 440 in a server remote from the terminal 400 is arranged. In this embodiment, preprocessed input signals are received by the control unit 230 of the terminal 201 and the control unit 430 of the server 400 to the evaluation unit 440 transferred, and there using appropriate variable rules, which in memory 460 are stored, evaluated to a communication readiness state of a user of the terminal 201 to identify and monitor. Upon detection of a change in the communication readiness state by the evaluation unit 440 corresponding control signals on the same way to the control unit 230 of the terminal 201 transfer. For communication of the terminal 201 via a communication connection 330 with the server 400 are appropriate interfaces 310 and 320 intended.

4 shows a further preferred embodiment of a telecommunications device 130 that is a public telephone 202 includes, which can be operated by a user without contact. For this purpose, the phone points 202 different non-contact Input devices on. These include a camera 212 , as well as an array of microphones 213a to 213d , As a further input device is a proximity sensor 211 provided with the phone 202 connected is.

The output devices are a screen 251 as well as a speaker system 252 , or tactile perceptible variable surfaces. In order to prevent others from listening, the speaker system is 252 formed in this embodiment as an ultrasonic speaker system, by means of which a directed to the user sound can be generated.

Optionally, the public telephone can have a button 214 by means of which a user can confirm his willingness to communicate or terminate his communication readiness, in the event of a lack of automatic recognition of his communication readiness state in a memory of the telephone 202 deposited variable rules for determining a communication ready state automatically adjust.

Below are some examples of sensor and speech signals of derivable variable rules.

If, for example, the propagation time difference between speech signals and microphones spatially arranged differently on the terminal within a specified tolerance is equal to or near zero, it can already be assumed in this simple case that the user is using a local and / or network-based application and / or a another person wants to communicate via the terminal, so that a corresponding activation signal is generated. It is also possible to use distance measurements as a criterion, which are based on the measurements of the angles to the user via a camera array.

The evaluation of the transit times can be coupled with an evaluation of an image signal such that it can be determined by evaluating this signal on the terminal or a remote server system, if the user's face appears in one or one of several possible predefined angles to the camera, so that In this case, a corresponding activation signal is generated. About an inclination sensor can be determined in the context of preset tolerances, whether the terminal is angled with the front side down or to the extent that it can be assumed that the user of the terminal, which makes a communication request less likely. The motion sensor serves to record the movements dynamically. If, for example, the user's face appears after a certain movement in the camera image, it may possibly be concluded with a relatively high probability or weighting on a communication request.

If, in addition to the sensor signals indicating a communication request with a high probability, a "hot word" is also detected with a high degree of probability, the activation signal is generated.

In a laterally vertical orientation of the terminal with a camera image without a face can be assumed, for example, with a certain probability that no communication request or no communication request is. Accordingly, an optionally existing activation state is deactivated or a deactivation signal is transmitted.

In this way, taking into account sensor signals and speech recognition, a set of rules, which can also be referred to as behavioral grammar, are defined, which is stored as default in the terminal and / or on a server connected to the communication network terminal and / or user-related. In addition to the sensors and / or microphones installed in the terminal, these can also be found in intelligent clothing and connected to the terminal wirelessly, for. B. communicate via Bluetooth connections, which allows to detect the relative position of the user to the terminal, for example, standing, leaning, sitting or lying on the side. Thus, for example, it can be defined as a rule that a communication request signal is generated only in the case of a user lying to the side and a lateral position of the terminal and / or other conditions. If the user is standing, and the terminal is held in a lateral position, then the probability of a communication request is rated as low.

The signals of the input devices installed in and / or on the terminal and / or in smart clothing can also be timed to form feature vectors, which in turn serve as a basis for training probability models such as a Hidden Markov model, which in this case also serves as Media models or behavioral models can be called.

With this trained, user-dependent behavioral model, the current user behavior detected via the input devices is constantly compared when the terminal is switched on, and the communication request signal or deactivation signal is generated at a certain calculated reliability value or probability.

The training of the behavioral model can be done both by a software on the terminal, as well as on a server accessible via a communication network. For this purpose, for example, the user in the initial phase can be made to actuate a key for each communication request and / or to perform another prescribed and narrowly defined, detectable action. The same applies analogously to the conclusion of the communication. The feature vectors previously detected and possibly shortly thereafter within a certain period of time via the input devices in a plurality of training processes then serve to form a reference behavioral model of a user.

Via a communication network, the sensor data or feature vectors of a large number of users acquired via the input devices during the training phases can also serve to form an averaged, user-independent behavioral model. For this purpose, the users in the initial phase, for example, in turn, prompted to press a button with each communication request and / or perform another prescribed and narrowly defined, detectable action. The feature vectors or data acquired beforehand and possibly shortly thereafter within a certain period of time via the input devices in a plurality of training processes are then transmitted to at least one server system connected to the communication network. Is a sufficient z. B. defined as a limit number of recorded data of a large number of different users, they can be used to form a user-independent behavioral model for a communication request.

It may also be advantageous to provide different behavioral models for a communication request, which are assigned to different applications, whereby these are stored accordingly depending on the application.

To improve the non-contact operation stationary and / or quasi-stationary, for example, built in vehicles, terminals and especially public telephones is in addition to the non-contact activation and / or deactivation of the signals that express a communication desire, nor an output mode, possibly with hands-free functionality directed sound, which prevents bystanders from overhearing the outputs of at least one terminal loudspeaker. For this purpose, an ultrasonic method known from the prior art can be used.

Furthermore, in public telephones advantageously movably installed and / or distributed over the device height sensors, such as proximity sensors, camera systems and microphone arrays, both the location of the user and the accurate tracking in the direction of the user, as well as the control of a likewise movable and / or over the device height distributed speaker system serve z. B. Adjustments to the size of the user (handicapped execution). Also distributed over the device height and / or area devices for magnetic coupling of hearing aids can be distributed, which are controlled according to the user localization.

An advantage of the method described is that not only voice data alone or only other input modes, but also other data of the respective input devices are subjected to a common evaluation, whereby the security for generating or deactivating communication request signals and thus the user acceptance is substantially increased. Users can, if necessary after completion of training phases, without additional keystrokes via a terminal of an application and / or other person as a communication partner "turn". In this way, public telephones are made possible without cable-connected handsets equipped with a built-in loudspeaker system and microphone array and / or camera and / or proximity sensors, in which the signal for communication is no longer activated or deactivated by removing and hanging up the receiver, but an activation and deactivation by monitoring the non-contact and / or unimodal and / or multimodal input options is done. This makes the devices more compact and less susceptible to vandalism. Also, a separate key, for example, to activate and / or cancel the communication, the operation of which may jokingly be made or forgotten, not necessarily required.

Claims (27)

Method for controlling a telecommunications terminal ( 200 . 201 . 202 ) with the steps a) register at least two by means of different first input devices ( 11 - 17 b) converting the two by means of different first input devices ( 11 - 17 ) generated input signals in a common first record, c) automatically determining a communication readiness state of a user of the telecommunications terminal ( 200 . 201 . 202 ) by evaluating the first data record ( 10 ) depending on predetermined variable rules ( 30 d) monitoring the communication readiness state of the user in order to detect a change in the communication readiness state, e) performing a predetermined function of the telecommunications terminal ( 200 . 201 . 202 ) upon detection of a change in the communication readiness state of the user, characterized in that the input signals for detecting a communication request include at least one of the following signals: - a transit time difference ( 43 ) from acoustical signals to microphones located in a different manner spatially on the telecommunication terminal equipment, which is equal to or near zero within a specified tolerance, or - when the user's face appears in the camera image after a predetermined movement of the telecommunication terminal equipment detected by a motion sensor, or via the evaluation Sensors and / or microphones in smart clothing, the relative position of the terminal is detected to the user. The method of claim 1, wherein monitoring the communication ready state by performing steps a) through c) at predetermined times. Method according to one of the preceding claims, further comprising the steps - storing the first data record ( 10 ), - Register at least one by means of a second input device ( 11 - 17 ) generated second input signal, - generating a second data set ( 10 ) from the at least one second input signal, - automatically determining a communication readiness state of a user of the telecommunication terminal device ( 200 . 201 . 202 ) by evaluating the second data set ( 10 ) in dependence of given fixed rules ( 20 ), - Automatic adjustment of the variable rules ( 30 ) for evaluating the first data set as a function of the communication readiness state of the user recognized from the evaluated second data record and of the stored first data record. Method according to claim 3, wherein substantially user-specific rules for evaluating the first data set ( 10 ) by repeatedly automatically adjusting the variable rules ( 30 ) for a specific user. Method according to claim 3, wherein substantially user-unspecific rules for evaluating the first data set ( 10 ) by repeatedly automatically adjusting the variable rules ( 30 ) for a variety of different users. Method according to one of claims 3 to 5, wherein the first and second data set ( 10 ) are created from input signals that are registered over a first or second period, wherein the first period before or after the second period or the first and second periods at least partially agree. Method according to one of the preceding claims, wherein the at least one first input signal by means of non-contact input devices ( 13 - 15 . 17 ) is generated. The method of any one of claims 3 to 7, wherein registering the at least one second input signal comprises manual user input or other input of a safe mode input mode. Method according to one of the preceding claims, wherein the evaluation of the first and / or second data set - the evaluation of a mouse / stylus ( 41 ), Keyboard or touch screen input ( 42 ) and / or - determining a transit time difference ( 43 ) acoustic signals generated by a microphone array and / or - a speech recognition ( 45 ) acoustic signals and / or - an evaluation of prosodic characteristics of acoustic signals ( 49 ) and / or - a speaker recognition ( 46 ) acoustic signals and / or - an image recognition ( 47 ) of an image taken by a camera and / or - a face recognition ( 48 ) of an image captured by a camera and / or - the evaluation of signals of a proximity, motion or inclination sensor ( 44 ). The method of any one of the preceding claims, wherein automatically determining a communication ready state comprises the steps of: Determining at least one probability value or a parameter which reflects the reliability of the determined state for the presence of a willingness to communicate, - comparing the determined probability value or parameter with a predetermined limit value. Method according to one of the preceding claims, wherein the evaluation of the first and / or second data set ( 10 ) and the monitoring by the telecommunication terminal device of the communication readiness state determined from the respectively evaluated data record ( 200 ) he follows. Method according to one of the preceding claims, wherein - the first and / or second data set ( 10 ) to one of the telecommunications terminal equipment ( 201 ) separate processing unit ( 400 ), - the evaluation of the first and / or second data record ( 10 ) and monitoring the communication readiness state determined from the respectively evaluated data record by means of the processing unit ( 400 ), and - upon detection of a change in the communication readiness state of the user for performing the predetermined function, a control signal from the processing unit ( 400 ) to the telecommunications terminal ( 201 ) is transmitted. Telecommunication device ( 110 . 120 . 130 ), comprising - a telecommunications terminal ( 200 . 201 . 202 ), - at least two different first input devices ( 11 - 17 ) each adapted to register different types of input signals, - a preprocessing unit ( 220 ) for converting registered input signals into digitized data, wherein the preprocessing unit ( 220 ) is adapted to convert the at least two registered input signals into a common data record, - an evaluation unit ( 240 . 440 ) for analyzing the digitized data, - a control unit ( 230 ) for controlling the telecommunication terminal equipment ( 200 . 201 ), and - a memory ( 260 . 460 ) for storing variable rules, wherein the evaluation unit ( 240 . 440 ) is designed to be able to obtain digitized data obtained by evaluating from first input signals in dependence on that in the memory ( 260 . 460 ) stored variable rules ( 30 ) automatically a communication readiness state of a user of the telecommunications terminal ( 200 . 201 . 202 ), to monitor the communication ready state of the user, and upon detection of a change in the communication readiness state of the user, a control signal to the control unit ( 230 ), and wherein the control unit ( 230 ) is adapted, in response to the receipt of the control signal, a predetermined function of the telecommunications terminal ( 200 . 201 . 202 ), characterized in that the input signals for detecting a communication request comprise at least one of the following signals: - a transit time difference ( 43 ) from acoustical signals to microphones spatially arranged on the telecommunication terminal equipment which is equal or close to zero within a specified tolerance, or - after a predetermined movement of the telecommunication terminal equipment detected by a motion sensor, the user's face appears in the camera image of a camera, or Signals from sensors and / or microphones in smart clothing of the user, with which the relative position of the user is detected to the terminal. Apparatus according to claim 13, further comprising a second input device ( 11 - 17 . 214 ) for registering second input signals, wherein the evaluation unit ( 240 . 440 ) is further configured to - first digitized data obtained from first input signals in the memory ( 260 ) by evaluating second digitized data obtained from second input signals as a function of fixed rules stored in the memory ( 20 ) automatically a communication readiness state of a user of the telecommunications terminal ( 200 . 201 . 202 ) to determine and monitor, depending on the communication readiness state determined from the evaluated second digitized data and / or a change of the communication readiness state and the stored first digitized data automatically the variable rules ( 30 ) and the custom rules in the memory ( 260 . 460 ). Device according to one of claims 13 or 14, wherein the preprocessing unit ( 220 ) is adapted to create first and / or second digitized data from input signals registered over a predetermined period of time. Apparatus according to claim 15, wherein the preprocessing unit ( 220 ) is adapted to generate the first and second digitized data from first and second input signals, respectively, registered over a first and second time period, wherein the first time period is before or after the second time period, or the first and second time periods are at least partially coincident. Device according to one of claims 14 to 16, wherein the determination of a change of the by evaluating the second digitized data by the evaluation unit ( 240 . 440 ) of the user's readiness for communication readjusting the variable rules ( 30 ) causes. Device according to one of claims 13 to 17, wherein the at least one first Input device at least one non-contact input device ( 13 - 15 . 17 . 211 . 212 . 213a - 213d ). Device according to one of claims 13 to 18, wherein the second input device as manually by the user to be operated input device ( 11 . 12 . 214 ) or is designed as a device for voice input. Device according to one of claims 13 to 19, wherein the evaluation unit ( 240 . 440 ) - to evaluate a mouse / stylus, keyboard or touchscreen input ( 21 . 22 ) and / or - for determining a transit time difference acoustic, from a microphone array ( 13 ) generated signals and / or - for speech recognition of acoustic signals and / or - for the evaluation of prosodic characteristics of acoustic signals ( 49 ) and / or - for speaker recognition of acoustic signals as a function of deposited speaker profiles and / or - for image recognition of a camera ( 14 ) and / or - for face detection of a camera ( 14 ) depending on stored facial profiles and / or - for evaluating signals of an approximation ( 15 ) or inclination sensor ( 16 ) is trained. Device according to one of claims 13 to 20, wherein in memory a limit value is stored and the evaluation unit ( 240 . 440 ) is designed to determine a probability value for the presence of a communication readiness and to compare the determined value with the stored limit value. Device according to one of claims 13 to 21, wherein the evaluation unit ( 240 ) in the telecommunications terminal ( 200 ) is arranged. Device according to one of claims 13 to 22, wherein the evaluation unit ( 440 ) in a separate processing unit ( 400 ) arranged by means of a communication link ( 330 ) with the telecommunication terminal equipment ( 201 ) is connectable. Device according to one of claims 13 to 23, wherein the first and / or second input device ( 211 ) than the telecommunications terminal equipment ( 202 ) separate input units are formed. The apparatus of claim 24, wherein the input units are configured to wirelessly communicate with the telecommunications terminal. Device according to one of Claims 13 to 25, designed as a public telephone ( 202 ), further comprising an output device ( 252 ) for the output of directional acoustic signals. Device according to claim 26, wherein the output device ( 252 ) is an ultrasonic speaker system, which is adapted to direct an acoustic beam to a predetermined spatial area by means of a modulated ultrasonic signal.

Images