JP7359201B2 - Physical condition detection system - Google Patents

Description

The present invention relates to a physical condition detection system.

Conventionally, drivers of commercial vehicles have had their health monitored through a roll call before driving. For example, the roll call management system disclosed in Patent Document 1 below acquires vital data such as breath alcohol, and the vital data is managed on a network.

Japanese Patent Application Publication No. 2012-159959

In addition to accidents caused by driving errors caused by blood alcohol, there are cases where drivers develop cerebral hemorrhage, cerebral infarction, or myocardial infarction while driving, leading to accidents. Therefore, it is necessary to detect symptoms such as cerebral hemorrhage, cerebral infarction, or myocardial infarction at an early stage while driving, and to stop riding or operating a vehicle if such symptoms occur. However, in Patent Document 1, it is possible to determine whether driving is possible by detecting alcohol in breath, but it is not possible to recognize symptoms such as cerebral hemorrhage, cerebral infarction, or myocardial infarction while driving. It is not effective for drivers to wear devices such as blood pressure monitors to detect symptoms such as cerebral hemorrhage, cerebral infarction, or myocardial infarction, as this may interfere with driving.

In addition to cerebral hemorrhage, cerebral infarction, or myocardial infarction, a driver's physical condition may change while driving. Even if a driver is not sick, he or she may become drowsy due to fatigue while driving. Even in such cases, it is necessary to detect changes in the driver's physical condition at an early stage, stop the driver from operating the vehicle, and if necessary change or cancel the driving duties.

An object of the present invention is to provide a physical condition detection system that can respond to changes in physical condition during driving.

In order to solve the above problems, a physical condition detection system according to the present invention has a configuration as described below.

In the physical condition detection system of the present invention, a voice analysis server, a driver terminal, and a manager terminal are communicably connected via a network. The voice analysis server includes a question storage unit that stores questions for the driver, a model answer storage unit that stores model answers made by the driver's voice to the questions, and a driver terminal that stores the questions stored in the question storage unit. It functions as a voice analysis section that compares the model answer with the driver's actual answer and analyzes the driver's physical condition based on the difference between the model answer and the actual answer. The driver terminal functions as a speaker that issues the question and a microphone that obtains the driver's actual voice response. The above-mentioned question is asked one or more times in a certain period of time, and the above-mentioned manager terminal obtains the analysis result of the driver's physical condition from the voice analysis server.

According to the present invention, a driver is asked and answered questions by voice while driving, and can recognize changes in physical condition that can be determined by voice. Accidents can be prevented by early detection of changes in physical condition caused by illness or drowsiness, and by stopping driving.

1 is a diagram showing the configuration of a physical condition detection system of the present application. FIG. 3 is a diagram of an audio waveform showing an example of a model answer. FIG. 6 is a diagram of a voice waveform showing an example of an actual answer in which the time from question to answer is long. FIG. 7 is a diagram of a voice waveform showing an example of a real answer in which the time from the start to the end of the answer is longer. FIG. 7 is a diagram of an audio waveform showing an example of a real answer in which the amplitude of the answer is large. FIG. 7 is a diagram of an audio waveform showing an example in which the sound of an actual answer is different from the sound of a model answer. 1 is a diagram showing the configuration of a physical condition detection system of the present application that is linked with an automatic stopping device of a vehicle. 1 is a diagram showing the configuration of a physical condition detection system of the present application equipped with a learning function.

A physical condition detection system according to an embodiment of the present invention will be described with reference to the drawings. The physical condition detection system is one in which the hardware and software function as described. Although a plurality of embodiments will be described, even in different embodiments, the same means may be given the same reference numerals and the explanation thereof may be omitted.

[Embodiment 1]
The physical condition detection system 10 of the present application shown in FIG. 1 includes a voice analysis server 12, a driver terminal 14, and an administrator terminal 16. The voice analysis server 12, driver terminal 14, and administrator terminal 16 are communicably connected via a network 18. The voice analysis server 12, the driver terminal 14, and the administrator terminal 16 are equipped with means for communicating with each other, and a description thereof may be omitted in some cases.

[Speech analysis server]
The voice analysis server 12 includes a question storage section 20 that stores questions, a model answer storage section 22 that stores model answers, a question transmission section 24 that questions the driver, and a voice analysis section 26 that analyzes the physical condition of the driver.

Questions for the driver are stored in the question storage section 20. The questions are audio data. The content of the question is arbitrary, and it is preferable that the question be answered without fail by the driver. For example, questions include the driver's name or date of birth. There may be one question or multiple questions.

The model answer storage unit 22 asks questions to the driver in advance and stores model answers obtained from the driver. The model response is voice data uttered by the driver. For example, the model answer 28 is a sound vibration waveform of air (FIG. 2), and specifically, the model answer 28 is converted into a digital format and stored. The amplitude varies depending on the size of the voice, the time varies depending on the length of the voice, and the waveform varies depending on the content of the voice. The amplitude in FIG. 2 is stored as voltage data.

The data of the model answer 28 includes a time t1 from the question to the answer, and a time t2 from the start to the end of the answer. The time t1 may be the time from the start of the question to the start of the answer, or may be the time from the end of the question to the start of the answer.

The model answer 28 is broken down into phonemes 30a to 30f for each waveform with the same characteristics. The amplitude B1 of each phoneme 30a to 30f and the waveform of each phoneme 30a to 30f are included in the model answer data. The amplitude B1 of each of the phonemes 30a to 30f indicates the magnitude of the sound and serves as intonation data (including intonation and accent). Although FIG. 2 only shows the amplitude B1 for the phoneme 30f, data on the amplitude B1 is individually stored for all the phonemes 30a to 30f.

Since the waveform of each phoneme 30a to 30f differs depending on the sound, the waveform of each phoneme 30a to 30f becomes data indicating what kind of sound it is. Sounds may be identified from the waveforms of the phonemes 30a to 30f, and each phoneme 30a to 30f may be stored as character data.

The data t1, t2, and B1 in the model answer 28 are set to include a certain tolerance range. This is because it is difficult for the driver to utter exactly the same voice as the model answer 28 at any time.

The question transmitting section 24 transmits the questions stored in the question storage section 20 to the driver terminal 14. Questions are sent multiple times over a certain period of time. The fixed time can be set to one day, half a day, eight hours, etc. Questions may be sent at predetermined time intervals or at multiple predetermined times. For example, the question transmitter 24 transmits a question once every hour. The driver will answer questions multiple times during the driving task. Note that if the driver's driving duties become short, the number of times the question is sent may be reduced to one. If there are multiple questions, you may send one question at a time or multiple questions.

The voice analysis section 26 compares the actual answer given by the driver during driving work with the model answer 28 stored in the model answer storage section 22, and analyzes the physical condition of the driver based on the difference. The voice analysis unit 26 determines the time from the question to the start of the answer, the time from the start to the end of the answer, the intonation of the driver's voice giving the answer, and the content of the answer.

The judgment methods used by the voice analysis unit 26 are (A) to (D) below. (A) The voice analysis unit 26 obtains the time from the question to the start of the answer in the actual answer, and compares the time with the time t1 of the model answer 28. If the actual response time is within the above-mentioned allowable range of the model response 28, it is determined that the driver's physical condition is normal. For example, as shown in FIG. 3, if the time t1x from the question to the start of the answer in the actual answer 32a is longer than the time t1 in the model answer 28 and is not within the above-mentioned allowable range for the model answer 28, the driver's physical condition is It is judged as abnormal.

If there is no answer even after a certain period of time has passed since the start of the question, the time t1x of the actual answer 32a cannot be determined. Therefore, the voice analysis unit 26 determines that the driver's physical condition is abnormal if no response is received within a predetermined time. The predetermined time is set to be longer than the above-mentioned allowable range.

(B) The voice analysis unit 26 obtains the time from the start to the end of the actual answer, and compares that time with the time t2 of the model answer 28. If the time of the actual answer falls within the above shared range of the model answer 28, it is determined that the driver's physical condition is normal. For example, as shown in FIG. 4, in the actual answer 32b, there is a period of no sound between the phonemes 30c and 30d and between the phonemes 30d and 30e. Therefore, the time t2x from the start to the end of the answer in the actual answer 32b becomes long, and if it is not within the above-mentioned allowable range of the model answer 28, it is determined that the driver's physical condition is abnormal.

(C) Divide the actual answer into phonemes and find the amplitude of the waveform for each phoneme. If the obtained amplitude of each phoneme is within the above-mentioned allowable range of the amplitude B1 of the model answer 28, it is determined that the driver's physical condition is normal. For example, in the actual answer 32c of FIG. 5, the amplitude B1x of the phoneme 30d' between the phonemes 30c and 30e is larger than the amplitude B1 of the phoneme 30d of the model answer 28. If the amplitude B1x of the phoneme 30d' in the actual answer 32c is not within the above-mentioned allowable range of the model answer 28, it is determined that the driver's physical condition is abnormal.

(D) Compare the waveforms of each phoneme between the model answer 28 and the actual answer. As a result of the comparison, if the waveforms are the same, it is determined that the driver's physical condition is normal. Rather than comparing the waveforms themselves, the sounds obtained from the waveforms may be converted into character data, and the waveforms may be compared based on whether or not the character data match. For example, the phoneme 30d of the model answer 28 in FIG. 2 becomes the phoneme 30g in the actual answer 32d of FIG. Since the actual response 32d in FIG. 6 has a different waveform and sound from the model response 28 in FIG. 2, it is determined that the driver's physical condition is abnormal.

If the driver's physical condition is normal in all of the above (A) to (D), it is finally determined that the driver's physical condition is normal. If at least one of the driver's physical conditions is abnormal, it is determined that the driver's physical condition is abnormal.

The analysis results are transmitted to the administrator terminal 16 via the network 18. The analysis result may be character data displayed on the display of the administrator terminal 16, or may be audio data emitted from a speaker. The analysis results may be sent to the driver terminal 14 in addition to being sent to the administrator terminal 16.

[Driver terminal]
The driver terminal 14 is a portable terminal or a device installed in the vehicle 34. Portable devices include smartphones, mobile phones, and tablet computers. The portable terminal may be removably attached to the vehicle 34. As the device installed in the vehicle 34, a car navigation system (automotive navigation system) device may be used, or the car navigation system and a smartphone may be linked. The driver terminal 14 functions as a speaker 36 that issues questions and a microphone 38 that captures audio.

The driver terminal 14 reproduces the questions sent from the voice analysis server 12 using the speaker 36. The driver listens to the questions uttered from the speaker 36 and answers them. The answer is collected by the microphone 38 and transmitted to the voice analysis server 12.

The driver simply listens and answers the questions. The driver is still holding the steering wheel and there is no problem with driving. If the driver becomes ill or becomes drowsy, the actual answer may not fall within the allowable range of the model answer 28, or the driver may give an inaccurate answer.

If there are multiple drivers, each driver carries a driver terminal 14. Questions are stored in the question storage section 20 for each driver terminal 14, and model answers 28 are stored in the model answer storage section 22. That is, the question and model answer 28 are linked for each driver terminal 14. The question transmitting unit 24 may transmit questions at different times for each driver terminal 14.

[Administrator terminal]
Administrator terminal 16 includes a portable terminal or a stationary computer. Examples of the administrator terminal 16 include a smartphone, a mobile phone, a tablet computer, a laptop computer, and a desktop computer. The administrator terminal 16 receives the analysis results from the voice analysis server 12 and displays the analysis results on a display provided in the administrator terminal 16. The analysis results may be announced over a loudspeaker. If the driver's physical condition is abnormal, the manager terminal 16 may emit an alarm through a speaker. From these facts, the manager can recognize changes in the driver's physical condition.

The manager looks at the analysis results displayed on the display of the manager terminal 16 and, if the driver's physical condition is abnormal, instructs the driver to stop riding and driving. If the manager terminal 16 and the driver terminal 14 are smartphones or mobile phones, they may be used to issue an instruction to stop boarding or operation. Further, a stop instruction unit 40 may be provided to send a command to stop crewing/operation to the manager terminal 16, and the stop command unit 40 may transmit the command to stop crew/operation to the driver terminal 14. A stop instruction is issued from the speaker 36 of the driver terminal 14, or the stop instruction is displayed on the display.

[Physical condition detection method]
Next, a method for detecting physical condition using the physical condition detection system 10 of the present application will be explained. (1) Questions for the driver are stored in the question storage section 20 in advance. Furthermore, the driver is asked questions in advance, and the driver's answers are stored in the model answer storage section 22 as model answers 28. The voice analysis server 12 may acquire the model answer 28 by sending question data from the voice analysis server 12 to the driver terminal 14 and having the driver give an answer.

(2) The question transmitting unit 24 transmits the question to the driver terminal 14 while the driver is driving. When the driver terminal 14 is activated and communicates with the voice analysis server 12 as a signal, the question transmitting section 24 starts transmitting questions. Transmission is performed multiple times at a fixed time. When the power of the driver terminal 14 is turned off and communication with the voice analysis server 12 is cut off, the question transmitting section 24 stops transmitting questions.

(3) When the driver terminal 14 receives the question, the speaker 36 emits the question, and the microphone 38 collects the driver's answer. The driver listens to the questions and answers them vocally.

(4) When the voice analysis server 12 receives the real answer, the voice analysis unit 26 analyzes the real answer. As described above, the determination is made based on whether the actual answer is within the allowable range of the model answer 28 and whether the content of the actual answer is correct. If the actual answer is not within the allowable range of the model answer 28 or the content of the actual answer is inaccurate, the analysis result is determined to be an abnormal physical condition. If the actual answer falls within the allowable range of the model answer 28 and the content of the actual answer is correct, the analysis result is determined to be in normal physical condition.

(5) The analysis results are sent from the voice analysis server 12 to the administrator terminal 16. The analysis results are displayed on a display provided in the administrator terminal 16. The administrator can recognize changes in the driver's physical condition by looking at the analysis results displayed on the display. If the analysis results indicate that the driver is in poor physical condition, the manager will contact the driver. If the driver has developed an illness such as a cerebral hemorrhage, cerebral infarction or myocardial infarction, the system will instruct the driver to stop driving or driving, and if the driver is drowsy, the system will instruct the driver to take a break. The operation stop instruction may be issued using both terminals 14 and 16, or may be issued using another mobile phone or the like.

As described above, the present application can detect abnormal physical condition of the driver while driving and instruct the driver to stop driving or riding. This makes it easier to prevent vehicle accidents by recognizing changes in the driver's physical condition, such as the onset of illness or drowsiness. Questions and answers are provided by voice, so there is no problem with driving.

[Embodiment 2]
As shown in FIG. 7, an automatic stop device 44 for the vehicle 34 may be used in the physical condition detection system 42. A stop signal to the automatic stop device 44 is generated by a stop signal generation unit 46 of the administrator terminal 16. When the analysis result indicates that the physical condition is abnormal, the stop signal generation unit 46 generates a stop signal by the manager operating the manager terminal 16, and the stop signal is sent from the manager terminal 16 to the automatic stop device 44 of the vehicle 34. Send. The vehicle 34 automatically stops due to the stop signal. The administrator may transmit a stop instruction to the driver terminal 14 using the stop instruction section 40 before transmitting the stop signal. The manager and driver may use other communication methods to communicate with each other to check for any abnormalities. Sends a stop signal if there is something really wrong with the driver.

Depending on the driving situation, the driver may not be able to respond immediately or the intonation of the voice may change. In this case, even if the driver's physical condition is normal, the analysis result of the voice analysis unit 26 may indicate that the driver's physical condition is abnormal. Therefore, the driver terminal 14 may include a canceling unit 48 that generates a signal to cancel the actual answer. The signal generated by the canceling unit 48 is transmitted to the voice analysis server 12. The cancellation unit 48 may be configured such that the driver emits a specific voice toward the microphone 38, analyzes the voice, and generates a cancellation signal.

If the voice analysis server 26 receives the cancellation signal within a predetermined time after receiving the actual answer, the voice analysis unit 26 cancels the analysis result. Even if the analysis result indicates an abnormal physical condition, the analysis result will be canceled. The canceled information is also transmitted to the administrator terminal 16. The manager can also stop the operation before instructing the driver to stop.

[Embodiment 3]
Questions and model answers 28 may be changed periodically. For example, the questions and model answers 28 are changed every month and are stored in the question storage section 20 and the model answer storage section 22. Even if the question is the same, only the model answer 28 may be rewritten. This is because the sound produced by the driver changes over time, and the model answer 28 may change.

[Embodiment 4]
Like the physical condition detection system 50 in FIG. 8, the voice analysis server 12 may accumulate real answers, learn the accumulated real answers, and rewrite the model answer 28 in the model answer storage unit 22. The speech analysis server 12 includes a real answer storage section 52 that stores real answers, and an answer learning section 54 that learns real answers. Every time the voice analysis server 12 receives a real answer, it stores the real answer in the real answer storage section 52.

The answer learning section 54 uses AI (Artificial Intelligence) to create model answers 28 and allowable ranges for the driver from a large number of actual answers, and rewrites the model answers 28 and allowable ranges in the model answer storage section 22. It is impossible for a driver to always give a fixed answer, and each answer is slightly different, but by using AI, the answer learning unit 54 learns a large number of real answers uttered by one driver during normal times. It is possible to improve the accuracy of the driver's judgment as to whether something is normal or abnormal.

(Section 1) In the physical condition detection system according to one embodiment, a voice analysis server, a driver terminal, and an administrator terminal are communicably connected via a network. The voice analysis server includes a question storage section that stores questions for the driver, a model answer storage section that stores model answers to the questions by the driver's voice, and a driver terminal that stores the questions stored in the question storage section. It functions as a voice analysis section that compares the model answer with the driver's actual answer and analyzes the driver's physical condition based on the difference between the model answer and the actual answer. The driver terminal functions as a speaker that issues the questions and a microphone that obtains real answers in the driver's voice. The above-mentioned question is asked one or more times in a certain period of time, and the above-mentioned manager terminal obtains the analysis result of the driver's physical condition from the voice analysis server.

According to the physical condition detection system described in item 1, the driver only answers questions by voice while driving, and does not interfere with driving. Changes in the driver's physical condition are detected early while driving, and the results are displayed on the administrator's terminal, allowing the administrator to contact the driver and stop driving. It is also possible for the driver to stop the vehicle before an accident occurs.

(Section 2) Comparison of the model answer and the driver's actual answer is based on the time from the question to the start of the answer, the time from the start to the end of the answer, the intonation of the driver's voice, and the response time. including at least one of accuracy.

According to the physical condition detection system described in item 2, it is possible to detect that the driver's vocalization has become strange due to a change in the driver's physical condition.

(Section 3) The vehicle driven by the driver is equipped with an automatic stop device, and the administrator terminal includes a stop signal generation unit that transmits a stop signal to the automatic stop device.

According to the physical condition detection system described in Section 3, the vehicle can be stopped at the discretion of the administrator without the driver having to stop the vehicle.

(Section 4) The driver terminal includes a cancellation unit that generates a signal to cancel the actual answer, and when the voice analysis server receives the signal to cancel the actual answer within a predetermined time after receiving the actual answer, the voice analysis unit Cancel the analyzed results.

According to the physical condition detection system described in Section 4, even if the driver is unable to give an accurate answer due to the driving situation and is determined to be in poor physical condition, the result can be canceled so that the driver is no longer determined to be in poor physical condition.

In addition, the present invention can be implemented with various improvements, modifications, and changes based on the knowledge of those skilled in the art without departing from the spirit thereof.

10, 42, 50: Physical condition detection system 12: Voice analysis server 14: Driver terminal 16: Administrator terminal 18: Network 20: Question storage section 22: Model answer storage section 24: Question transmission section 26: Voice analysis section 28: Model answers 30a, 30b, 30c, 30d, 30d', 30e, 30f, 30d: Phonemes 32a, 32b, 32c, 32d: Real answer with abnormality 34: Vehicle 36: Speaker 38: Microphone 40: Stop instruction section 44: Automatic Stop device 46: Stop signal generation section 48: Cancellation section 52: Real answer storage section 54: Answer learning section

Claims (10)

a question storage unit that stores questions for the driver;
a model answer storage unit that stores a model answer that is a voice uttered by the driver in response to the question;
a question output unit that outputs the question stored in the question storage unit as voice;
an answer acquisition unit that acquires a model answer and a real answer based on the driver's voice in response to the question;
a voice analysis unit that compares the model answer with the driver's real answer and analyzes the driver's physical condition based on the difference between the model answer and the real answer;
When storing a model answer, the question output section outputs the question stored in the question storage section as a voice, and the answer acquisition section acquires the model answer in response to the question by the driver's voice, and a model answer storage unit stores the model answer acquired by the answer acquisition unit,
When analyzing the driver's physical condition, the question output unit outputs the questions stored in the question storage unit as voice during the driver's driving time, and the answer acquisition unit outputs the questions stored in the question storage unit as voice during the driver's driving time. The physical condition detection system obtains the driver's actual voice response to the question , and the voice analysis unit compares the model response with the driver's actual response . The voice analysis server, driver terminal, and administrator terminal are connected via a network to enable communication.
The voice analysis server
the question storage unit;
the model answer storage unit;
a question transmitting unit that transmits the question stored in the question storage unit to a driver terminal;
functioning as the voice analysis section;
The driver terminal is
a speaker as the question output section;
Functions as a microphone as the answer acquisition unit,
The question transmitting unit transmits the question one or more times in a certain period of time,
The physical condition detection system according to claim 1, wherein the administrator terminal obtains an analysis result of the driver's physical condition from a voice analysis server. The model answer includes the time from the question to the start of the answer and its tolerance range,
The voice analysis unit calculates the time from the question to the start of the answer in the real answer, compares the calculated time with the time from the question to the start of the answer in the model answer, and if it is within the allowable range, determines the physical condition of the driver. The physical condition detection system according to claim 1 or 2 , wherein the physical condition detection system determines that the driver's physical condition is normal, and determines that the driver's physical condition is not normal if it is outside the allowable range. The model answer includes the time from the start of the answer to the end of the answer and its tolerance range,
The voice analysis section calculates the time from the start of the answer to the end of the answer in the real answer, compares the obtained time with the time from the start of the answer to the end of the answer in the model answer, and if it is within the allowable range, the driver 4. The physical condition detection system according to claim 1 , wherein the physical condition of the driver is determined to be normal, and if the physical condition of the driver is outside the allowable range, it is determined that the physical condition of the driver is not normal. The model answer includes amplitudes and their tolerance ranges for the plurality of divided phonemes,
The voice analysis section divides the actual answer into phonemes, finds the amplitude of the waveform for each phoneme, compares the obtained amplitude with the amplitude of the phoneme in the model answer, and if it is within the allowable range, determines the physical condition of the driver. The physical condition detection system according to any one of claims 1 to 4 , which determines that the driver's physical condition is normal, and determines that the driver's physical condition is not normal if it is outside the allowable range. The model answer includes waveforms of a plurality of divided phonemes and their tolerance ranges,
The voice analysis unit divides the actual answer into phonemes, compares the waveform of each phoneme with the waveform of the phoneme in the model answer, and determines that the driver's physical condition is normal if it is within the acceptable range, The physical condition detection system according to any one of claims 1 to 5 , which determines that the driver's physical condition is not normal if the physical condition is outside the allowable range. The vehicle driven by the driver is equipped with an automatic stop device,
7. The physical condition detection system according to claim 1 , further comprising a stop signal generation section that transmits a stop signal to an automatic stop device when the voice analysis section determines that the physical condition of the driver is not normal. Equipped with a cancellation section that generates a signal that cancels the actual answer,
The physical condition detection system according to any one of claims 1 to 7 , wherein the voice analysis section cancels the analyzed result if a signal for canceling the response is received within a predetermined time after receiving the actual answer. 9. The physical condition detection system according to claim 1 , further comprising an answer learning section that learns model answers from actual answers and rewrites the model answers stored in the model answer storage section. A voice analysis server in the physical condition detection system according to any one of claims 1 to 9 ,
the question storage unit;
the model answer storage unit;
a question transmitting unit that transmits the question stored in the question storage unit to a driver terminal;
A voice analysis server that functions as the voice analysis section.

Images