DE102019213928A1 - Evaluating a voice of a vehicle occupant - Google Patents

Abstract

Method for evaluating a voice of a vehicle occupant, comprising the following steps: acquiring audio data with a voice sample by means of a microphone; Evaluating the audio data with regard to at least one characteristic of the voice sample; Decide on the vehicle occupant's access rights to the vehicle based on the evaluation of the audio data.

Description

FIELD OF THE INVENTION

The present invention relates to a method for evaluating a voice of a vehicle occupant.

TECHNICAL BACKGROUND

In 2017, 7.3% of all fatally injured road users died as a result of an alcohol accident in Germany alone, i.e. every 14th person killed (source: https://www.dvr.de/u nfallstatistik / de / al kohol /).

The DE 102007047709 from Bosch is an example of the state of the art for an alcohol immobilizer. In the document, a device for preventing a vehicle from being started by vehicle drivers who are under the influence of alcohol is described in more detail. The device thus includes a determination of the alcohol content in the driver's breath, blood, tissue, skin sweat or tissue fluid. A vehicle lock is activated depending on the evaluation result. In some countries such a test may already be compulsory, in other countries it could be offered as a voluntary corrective measure for repeat offenders.

The system requires a lot of sensors and the evaluations with blood samples etc. are very extensive. For example, a humidity, temperature and pressure sensor are necessary in each case just to avoid that an air pump is not released as sample air. Nevertheless, a simple manipulation / bypassing by submitting "samples" by the co-driver or passers-by / friends is possible at any time. In addition, the system is expensive, with an installation about 2000 euros are due at the time of registration (https://www.alkoholtester-infos.de/alcolock/).

SUMMARY OF THE INVENTION

Against this background, the invention is based on the object of specifying a cost-effective method for deciding on access rights to a vehicle.

According to the invention, this object is achieved by a method for evaluating a voice of a vehicle occupant with the features of claim 1.

• - ein Verfahren zum Auswerten einer Stimme eines Fahrzeuginsassen mit den folgenden Schritten: Erfassen von Audiodaten mit einer Stimmprobe mittels eines Mikrofons; Auswerten der Audiodaten hinsichtlich wenigstens einer Charakteristik der Stimmprobe; Entscheiden über Zugriffsrechte des Fahrzeuginsassen auf das Fahrzeug aufgrund der Auswertung der Audiodaten.

Accordingly, it is provided:

• - A method for evaluating a voice of a vehicle occupant with the following steps: recording audio data with a voice sample by means of a microphone; Evaluating the audio data with regard to at least one characteristic of the voice sample; Decide on the vehicle occupant's access rights to the vehicle based on the evaluation of the audio data.

A microphone is a sound transducer that converts airborne sound as alternating sound pressure vibrations into corresponding electrical voltage changes as a microphone signal.

A camera is a photographic apparatus that can record static or moving images on photographic film or electronically on magnetic video tape or digital storage medium or transmit them via an interface.

An artificial neural network (ANN) is, in particular, a network of networked artificial neurons which is simulated in a computer program. The artificial neurons are typically arranged on different layers. The artificial neural network usually comprises an input layer and an output layer, the neuron output of which is the only one of the artificial neural network that is visible. Layers located between the input layer and the output layer are typically referred to as hidden layers. Typically, an architecture or topology of an artificial neural network is first initiated and then trained in a training phase for a special task or for several tasks in a training phase.

The term “topology of an ANN” encompasses all aspects relating to the structure of an ANN. This includes, for example, the number of neurons of the ANN, the division of the neurons into the individual layers of the ANN, the number of layers of an ANN, the networking of the neurons and the weighting of the networking.

The training of the artificial neural network typically includes a change in the weight of a connection between two artificial neurons of the artificial neural network. The weight contains information on the strength of the consideration of an input of a neuron. The training of the artificial neural network can also include developing new connections between artificial neurons, deleting existing connections between artificial neurons, adapting threshold values of the artificial neurons and / or adding or deleting artificial neurons.

An audio signal (also sound signal) is an electrical signal that transports acoustic information or acoustic noise.

Vehicles within the meaning of this patent application are land, air and / or water vehicles.

Computer program products generally comprise a sequence of instructions which, when the program is loaded, cause the hardware to carry out a specific method that leads to a specific result.

Access rights to a vehicle are a release for an action in interaction with the vehicle. For example, the release for starting the vehicle, a read right for data stored in the vehicle and the like are an access right.

In this patent application, a voice sample is speaking and / or singing syllables or texts. Characteristics of a voice can be its biometric characteristics, i.e. permanent characteristics that do not change. Other characteristics are, for example, the effects of the muscle tone of the muscles affecting the voice on the voice. Muscle tone is variable, so these are temporary features of a voice.

Mind-altering substances include alcohol, drugs, prescription drugs such as opiates, psychotropic drugs, and anti-epileptic drugs. These include in particular psychotropic substances that affect the human psyche. Other names for it are also psychoactive substance or psychotropic.

Any externally supplied substance that causes changes in a person's psyche and consciousness is called psychotropic or psychoactive. Such an influence can be subliminal and, for example, be experienced positively as stimulation, relaxation or a pleasant change in mood. However, it can also have a profound effect on the state of consciousness and lead to seizures, impaired consciousness or, in the worst case, to a coma.

States of consciousness are, for example, states of intoxication due to psychotropic substances, fatigue or acute outbreaks of mental illnesses such as schizophrenia or depression.

A vehicle occupant is, for example, a vehicle driver or a passenger.

Commissioning a vehicle includes starting an engine of a vehicle. Commissioning can also include the activation of other functions of the vehicle, such as a radio.

In this patent application, persons authorized to operate a vehicle are e.g. a vehicle owner, vehicle renter or other persons who have received authorization to drive the vehicle from an owner of the vehicle.

In this patent application, the first training phase is interrupted by the second training phase. For example, it can be provided that the vehicle is delivered to a customer between the first training phase and the second training phase.

Recognizing at least one specific voice means identifying a voice.

The basic idea of the invention is to evaluate the voice of a vehicle occupant, in particular a vehicle driver, with regard to the characteristics of a spoken voice sample and to decide on the vehicle occupant's right of access to the vehicle based on the evaluation of the voice sample.

The decision about the access right of the vehicle occupant includes, for example, the decision as to whether a vehicle occupant is authorized to start a vehicle or, for example, to access personal data stored in the vehicle.

For example, in this context it is also conceivable that so-called “last destinations” from a navigation system can only be read by certain people.

Advantageous refinements and developments emerge from the further subclaims and from the description with reference to the figures of the drawing.

According to a preferred development of the invention, the audio data are evaluated with regard to a state of consciousness of the vehicle occupant who spoke the voice sample. In this case, it is useful if the access right is a release to start the vehicle.

Numerous states of consciousness affect muscle tone. A person's voice is determined, among other things, by the muscle tone of the Lip, jaw, palate and tongue muscles.

For example, it can be provided that overtiredness of the vehicle occupant is detected on the basis of the voice test and, in this case, the vehicle occupant is prevented from starting the vehicle.

Alternatively, it is also conceivable to detect acute phases of mental illness on the basis of the voice test and, if necessary, to prevent the vehicle occupant from starting the vehicle.

It is also expedient here if the voice sample is examined for signs of consumption of consciousness-altering substances by the vehicle occupant who spoke the voice sample.

Many mind-altering substances cause intoxication in people, which impair their ability to drive. Intoxication can often be seen in a voice sample. For example, the consumption of alcohol leads to the tongue being poorly coordinated, even with a small amount of alcohol. In this case, one also speaks of the tongue drooping. Such coordination errors can be reliably detected in a voice sample using artificial intelligence.

It goes without saying that the consumption of other mind-altering substances also has an influence on the voice and can be detected in a voice sample.

According to a preferred development of the invention, it is checked whether the vehicle occupant who spoke the voice sample is driving the vehicle or is sitting on a driver's seat, that is to say at the front left in right-hand traffic. This can be checked, for example, with a camera or with a directional microphone.

Furthermore, it can be advantageous to monitor that vehicle occupants do not change places after the voice sample has been submitted.

It is thus possible to prevent the voice sample from being given by a passenger who may not have consumed any mind-altering substances and a right to access the vehicle is granted on the basis of the voice test of the front passenger, although the vehicle is actually used by a vehicle driver who has consumed mind-altering substances has, is controlled.

According to a preferred development of the invention, a vehicle driver is requested to give a voice sample before the vehicle is started up. In this regard, it is conceivable, for example, to ask the vehicle driver a question with a polysyllabic answer. For example, such a question could be, “How much is 700 + 13?”.

It is also useful if the text to be spoken cannot be foreseen by the vehicle occupant. It is thus possible to prevent a vehicle occupant from bypassing the method in that previously recorded voice samples are stored on a data carrier and played back by a playback device. This is particularly advantageous when recognizing characteristics due to mind-altering substances in the voice.

Unpredictable can mean that each text can only be spoken once or that at least one can choose from a large number, e.g. at least 1000, of texts to be spoken.

Alternatively or additionally, it is conceivable to record audio data with a voice sample while driving, without the vehicle occupant being requested to submit a voice sample. For example, the voice sample can be generated from spoken conversations, telephone calls, voice commands and / or the like.

According to a preferred development of the invention, based on the audio data with the voice sample, it is checked whether it is a vehicle occupant who is authorized to drive the vehicle. Effective theft protection can thus be provided on the basis of biometric features of a voice of a vehicle owner or an authorized vehicle owner.

The human voice has biometric features, i. H. Like a fingerprint, each voice is unique and can be clearly assigned to a person. The voice depends on the size of the oral cavity, the throat, the nose, the mouth and the arrangement of the muscles such as the tongue, the jaw muscles or the palate muscles.

It is useful if the function of checking the authorization to drive the vehicle can be deactivated for a predetermined period of time, for a predetermined number of starts of the vehicle and / or for a predetermined number of kilometers. This ensures that, as an exception, third parties can also be given the opportunity to obtain access rights to a vehicle without first saving its voice as an authorized vehicle driver.

It goes without saying that it is advantageous to design a possible functionality for checking the voice sample with regard to consciousness-altering substances in such a way that this functionality cannot be deactivated.

It goes without saying that a training method for an artificial neural network, which is set up to carry out a method as described above, is advantageous. It is provided that the artificial neural network is trained in a first training phase to decide on the basis of voice samples whether a speaker of a voice sample is under the influence of mind-altering substances and the artificial neural network in a second training phase to recognize at least one specific one Voice is trained in such a way that the artificial neural network is able to recognize an influence of mind-altering substances in the specific voice after completion of the training.

Accordingly, the artificial neural network is first trained to recognize characteristics in a voice based on mind-altering substances, in particular before a vehicle is delivered. In a second training phase, the artificial neural network is specifically trained to recognize specific voices. It goes without saying, however, that it is not immediately possible to train the artificial neural network directly to recognize consciousness-altering substances in a specific voice, as it is generally not possible to provide a vehicle owner with a consciousness-altering network as part of the second training phase To administer substance.

It is thus possible to prevent people with a persistent speech disorder from not being granted access rights to a vehicle even though the people are in a clear state of consciousness.

It goes without saying that in this context a distinction is made between persistent and temporary language disorders. A speech disorder is persistent if it lasts for several weeks, for example. Persistent speech disorders can be congenital or triggered by illnesses.

It is also expedient if, during the second training phase, it is ensured by further means that a speaker of the specific voice is not under the influence of mind-altering substances.

It is thus possible to prevent a method for detecting characteristics in a voice due to mind-altering substances from being circumvented in that a vehicle owner ingests a mind-altering substance before the second training phase and the artificial neural network in the second training phase with a voice sample that Characteristics due to a mind-altering substance, has, is trained.

The further means can be provided, for example, as breath tests, measurement of the conductivity of the skin and / or by a further person who carries out corresponding tests manually. The other person can, for example, be commissioned by a certification organization, e.g. TÜV.

The computer program product according to one embodiment of the invention carries out the steps of a method according to the preceding description when the computer program product runs on a computer, in particular a computer internal to the vehicle. When the program in question is used on a computer, the computer program product produces an effect, namely the recognition of characteristics in a voice sample or the training of an artificial neural network.

Figure list

• 1 ein schematisches Blockdiagramm gemäß einer Ausführungsform der Erfindung.

The present invention is explained in more detail below with reference to the exemplary embodiments specified in the schematic figures of the drawings. It shows:

• 1 a schematic block diagram according to an embodiment of the invention.

The accompanying drawings are intended to provide a further understanding of the embodiments of the invention. They illustrate embodiments and, in conjunction with the description, serve to explain principles and concepts of the invention. Other embodiments and many of the advantages mentioned emerge with a view to the drawings. The elements of the drawings are not necessarily shown to scale with one another.

In the figures of the drawings, identical, functionally identical and identically acting elements, features and components - unless stated otherwise - are each provided with the same reference symbols.

DESCRIPTION OF EXEMPLARY EMBODIMENTS

1 Figure 3 shows a schematic block diagram according to an embodiment of the invention. In the step S1 audio data are recorded with a voice sample using a microphone. In the step S2 the audio data are evaluated with regard to at least one characteristic of the voice sample. In the step S3 the vehicle occupant's access rights to the vehicle are decided based on the evaluation of the audio data.

List of reference symbols

S1-S3

Procedural steps

QUOTES INCLUDED IN THE DESCRIPTION

This list of the documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• DE 102007047709 [0003]

Claims (12)

Method for evaluating a voice of a vehicle occupant with the following steps: - Acquisition (S1) of audio data with a voice sample by means of a microphone; - evaluating (S2) the audio data with regard to at least one characteristic of the voice sample; - Decide (S3) about the vehicle occupant's access rights to the vehicle based on the evaluation of the audio data. Procedure according to Claim 1 , the audio data being evaluated with regard to a state of consciousness of the vehicle occupant who spoke the voice sample. Procedure according to Claim 2 wherein the voice sample is examined for evidence of mind-altering substance use by the vehicle occupant who spoke the voice sample. Method according to one of the preceding claims, wherein it is checked whether the vehicle occupant who spoke the voice sample is driving the vehicle. Method according to one of the preceding claims, wherein a vehicle driver is requested to submit a voice sample before the vehicle is put into operation. Procedure according to Claim 5 , whereby a text to be spoken of the voice sample of a request to submit the voice sample cannot be foreseen for the vehicle occupant. Method according to one of the preceding claims, wherein audio data are recorded with a voice sample without the vehicle occupant being requested to submit a voice sample. Method according to one of the preceding claims, wherein, on the basis of the audio data with the voice sample, it is checked whether it is a vehicle occupant who is authorized to drive the vehicle. Procedure according to Claim 8 , wherein the function of checking the authorization to drive the vehicle for a predetermined period of time, for a predetermined number of starts of the vehicle and / or for a predetermined number of kilometers driven can be deactivated. Training method for an artificial neural network, which is set up to carry out a method according to one of the preceding claims, wherein the artificial neural network is trained in a first training phase to decide on the basis of a voice sample whether a speaker of a voice sample is under the influence of mind-altering substances and the artificial neural network is trained in a second training phase to recognize at least one specific voice, such that the artificial neural network is able to recognize an influence of mind-altering substances in the specific voice after the training has ended. Training method according to Claim 9 During the second training phase, further means are used to ensure that a speaker of the specific voice is not under the influence of mind-altering substances. Computer program product which is designed to be loaded into a memory of a computer and which comprises software code sections with which the steps of the method according to one of the preceding Claims 1 - 11 be executed when the computer program product is running on the computer.

Images