CN108903958A - Driver attention evaluates early warning system and its implementation method - Google Patents

Abstract

The invention discloses a driver attention evaluation and early warning solution for single-point collection, which monitors the driver's driving state in real time so as to give reliable reminders when the driver is in a fatigued state. It mainly includes EEG acquisition subsystem, attention evaluation subsystem and driving warning subsystem. Brainwave earphones are used to record the EEG signals from the driver's forehead in real time, and the signals are transmitted to the attention evaluation subsystem via Bluetooth for analysis; the attention evaluation subsystem uses the data of theta and alpha rhythm waves collected in the first 1 minute to construct their respective sample spaces . The driver's driving fatigue status is classified and the analysis results are transmitted to the driving warning subsystem. The driver's smartphone displays the analysis results and automatically implements corresponding warning operations. Applying the driver attention evaluation and early warning scheme of the present invention, the device has a simple structure, is easy to carry and wear, and has timely early warning, while reducing algorithm complexity and reducing the traffic accident rate caused by fatigue driving.

Description

Driver's Attention Evaluation Early Warning System and Its Implementation Method

technical field

The invention relates to a fatigue driving early warning system based on electroencephalogram signal analysis, which belongs to the field of combination of electromechanical and communication.

Background technique

Research studies have shown that fatigue is an important factor affecting drivers' safe driving. When the driver is tired, there will be blurred vision, sore back pain, dull movements, swollen hands and feet, or lack of concentration, slow response, incomplete thinking, lack of energy, anxiety, and impatience. If you still drive the vehicle reluctantly, it is likely to cause a traffic accident. According to incomplete statistics, there are more than 1 billion traffic accidents in the world every year, and the accidents caused by driver fatigue account for about 20% to 30% of the total number of accidents, and the fatality rate of accidents caused by fatigue driving accounts for 20% of all traffic accidents. About 70% of the mortality rate. Therefore, in recent years, the problem of driver fatigue driving has attracted the attention of more and more researchers from all over the world, and the research on fatigue driving detection methods has more important practical significance. In order to control motor vehicle driving accidents, the solution to fatigue driving has become a top priority.

In recent years, with the rapid development of brain-computer interface (BCI) technology, it has become possible to control the external environment only through the thinking activities of the human brain. A large number of literatures point out that the rhythmic waves (δ wave, θ wave, α wave, β wave) of the electroencephalogram signal (EEG for short) are closely related to the states of the human brain such as alertness and fatigue. When the human body is in a state of alertness, the activity of the α wave in the EEG signal in the frontal area is significantly stronger than that of other wave bands; and when the human body is in a state of fatigue, the α wave will be suppressed, while the activity of the θ wave in the temporal lobe area becomes most notably. Therefore, it has become possible to judge the state of the driver by utilizing the changes in the rhythm wave of the EEG signal.

Contents of the invention

In order to solve the above-mentioned technical problems, the object of the present invention is to propose a single-point acquisition driver's attention evaluation and early warning scheme, to detect the driver's fatigue degree based on the attention analysis of EEG signals, and to solve the problem of the driver's fatigue when the driver feels tired. Issues that are automatically and promptly reminded.

A kind of technical solution that the present invention realizes above-mentioned purpose is, the driver's attention evaluation early warning system is associated with the driver's smart phone, and it is characterized in that the system composition comprises:

EEG acquisition subsystem: used for real-time recording of the EEG signal of the driver in the driving state, and signal processing as the source data of attention evaluation;

Attention evaluation subsystem: used to filter the source data, extract the time-domain waveform of the required band, and perform algorithm analysis and classification identification to obtain the fatigue state;

Driving warning subsystem: it is an APP based on smart phone hardware, which is used to feed back early warning information corresponding to the state of weak concentration through the smart phone according to the obtained driver's fatigue state and combined with the set driving mode;

The three subsystems communicate and communicate with each other successively and exchange data.

Further, the EEG acquisition subsystem is oriented to the one-way communication of the attention evaluation subsystem, and the attention evaluation subsystem is oriented to the one-way communication of the driving warning subsystem, and the one-way communication is realized based on a Bluetooth device.

Further, the EEG acquisition subsystem is a single-channel acquisition device that dynamically compensates for EEG signal fluctuations within the normal range and individual differences. side earlobes.

Further, the composition of described attention evaluation subsystem comprises:

The feature signal extraction module is used to extract the characteristic rhythm waves θ and α from the EEG signal by wavelet transform;

A characteristic sample module is used to construct a characteristic sample space using the characteristic rhythm waves θ and α collected during the initial setting period;

The distance calculation weighting module is used to calculate the distance between the real-time signal and the feature sample space and weight the summation;

The pattern matching module is used to compare the relationship between the obtained distance and the preset threshold, and match to obtain the judgment results of the fatigue state in two or more degrees from concentration to distraction.

Further, the driving warning subsystem is provided with a driving mode input interface and an acousto-optic unit, a vibration unit and an automatic navigation unit for outputting warning signals based on a smart phone.

Another technical solution for the present invention to achieve the above-mentioned purpose is that the driver's attention evaluation and early warning method is implemented in association with the driver's smart phone, and is characterized in that it includes steps:

EEG acquisition: Use the EEG acquisition subsystem to record the driver's EEG signal in real time while driving, and process the signal as the source data for attention evaluation;

Attention evaluation: use the attention evaluation subsystem to filter the source data, extract the time-domain waveform of the required band, and perform algorithm analysis and classification identification to obtain the fatigue state;

Driving warning: use the APP of the driving warning subsystem based on the hardware of the smart phone, according to the obtained driver fatigue state, combined with the set driving mode to feed back the warning information corresponding to the state of weak concentration through the smart phone;

The three subsystems communicate and communicate with each other successively and exchange data.

Further, the EEG acquisition subsystem is one-way communication oriented to the attention evaluation subsystem, and the attention evaluation subsystem is unidirectional communication oriented to the driving warning subsystem, and the one-way communication is realized by Bluetooth.

Further, the EEG acquisition is aimed at dynamically compensating for the fluctuation trend and individual differences of the EEG signal within the normal range, and the headgear method is used to set the acquisition point as the driver's side forehead and the reference point as the driver's side earlobe. channel to collect EEG signals.

Further, it is characterized in that the attention evaluation includes steps:

Extract characteristic signals, and use wavelet transform to extract characteristic rhythm waves θ and α from EEG signals;

Establish characteristic samples, and use the characteristic rhythm waves θ and α collected during the initial setting period to construct the characteristic sample space;

Distance calculation weighting, calculating the distance between the real-time signal and the feature sample space and weighting the summation;

Pattern matching, compare the relationship between the obtained distance and the preset threshold, and match to obtain the fatigue state judgment results of two or more degrees from concentration to distraction.

Further, in the driving warning, the driving mode input interface and the sound and light unit, the vibration unit and the automatic navigation unit are set based on the smart phone, the driver manually inputs or the smart phone intelligently recognizes the driving mode, and the smart phone responds to output voice prompts according to the judgment result of the driver's fatigue state , light color reminder, vibration stimulation or automatic navigation to guide the driver into the nearest service area to rest.

The application of the driver's attention evaluation and early warning scheme of the present invention has outstanding substantive features and remarkable progress: the present invention adopts portable EEG acquisition and analysis equipment, and the EEG acquisition equipment adopts a single-channel EEG acquisition method, so that the structure It is simpler and easier to carry, reduces the complexity of the algorithm, and improves the driver's comfort. Filtering, signal processing, and algorithm analysis are performed based on the collected EEG signals to obtain driver fatigue status and early warning, which realizes the function of preventing fatigue driving, and has high judgment accuracy, timely early warning, and easy operation. The problem of fatigue driving has high practical value.

Description of drawings

FIG. 1 is a schematic diagram of the structure and communication of the driver's attention evaluation and early warning system of the present invention.

Figure 2 is a schematic diagram of the installation of the EEG acquisition equipment.

Fig. 3 is a schematic diagram of installation from another perspective of Fig. 2 .

Fig. 4 is an algorithm flow chart of the driver's attention evaluation and early warning method of the present invention.

Figure 5 is a schematic diagram of the initial use interface of the smart phone APP.

Fig. 6 is a schematic diagram of the general mode user interface of the smart phone APP.

Detailed ways

The specific implementation of the present invention will be further described in detail below in conjunction with the accompanying drawings of the embodiments, so as to make the technical solution of the present invention easier to understand and grasp, so as to define and support the protection scope of the present invention more clearly.

The present invention is a fatigue driving early warning system based on EEG analysis. As shown in FIG. 1 , the system mainly includes three parts: an EEG acquisition subsystem, an attention evaluation subsystem and a driving early warning subsystem. Each subsystem is independent and complementary to each other, can carry out data transmission, and carry out data interaction according to a specific transmission protocol. The EEG acquisition subsystem and the attention evaluation subsystem are connected through a Bluetooth device for real-time acquisition of the driver's EEG signals in the driving state; the attention evaluation subsystem performs algorithm analysis based on the data obtained by the EEG acquisition subsystem, and The analysis results are sent to the driving warning subsystem through the bluetooth transmission device. The driving warning subsystem is an APP software independently developed in the mobile phone. After receiving the data from the attention evaluation subsystem, the driving warning subsystem judges whether the driver is in a state of inattention according to the received data. When it detects that the driver is in a state of inattention, the mobile phone vibrates and sends out an alarm to remind the driver. In order to facilitate the understanding of the implementation, the implementation details of each subsystem are described in detail as follows.

1. The EEG acquisition subsystem adopts the "slow self-adaptive" algorithm. Dynamically compensate for the fluctuation trend and individual differences of the brain wave signals of different users within the normal range, so as to calibrate the signals. It can be applied to different groups of people and different surrounding environments. It can have very good accuracy and reliability in different application scenarios. Using single-conductor dry electrode technology, EEG signal acquisition becomes easy to use, safe and comfortable to wear. ThinkGear™ separates the brain wave signal from the noisy environment and amplifies it to produce a clear brain wave signal. Moreover, the single-channel acquisition method is used to make the structure simpler, easy to carry, reduce the complexity of the algorithm, and improve the comfort of the driver.

ThinkGear™ ASIC chip technical features: 1. Using single-lead dry electrode measurement technology, 2. Integrated filtering and amplification functions of physiological signals, 3. Integrated eSense™ algorithm (users do not need to perform additional signal processing), 4. Adopting industrial standards 5. Output the original EEG waveform data, the data output frequency can reach 512Hz, low power consumption and low power consumption.

As shown in Figures 2 and 3, the EEG acquisition subsystem acquires EEG signals from four leads, and the acquisition point is located on the left forehead. The reference point is the earlobe. Put on the collection headgear, the headgear will be close to the corresponding pole. Of course, in actual implementation, the corresponding left and right directions can be flexibly selected according to actual conditions.

2. Attention evaluation subsystem, which consists of several parts such as EEG signal processing, classification recognition algorithm, and early warning evaluation. The signal processing algorithm in the system is efficient, small in size, and easy to operate. The method of channel acquisition, so the intelligent classification algorithm has the advantages of short recognition time and small amount of learning data required.

The role of EEG signal preprocessing is to obtain the required EEG signal data and suppress power frequency interference and noise. Since the EEG signal strength is extremely weak at the microvolt level, it is easily affected by external power frequency noise at 50 Hz and polarization levels below 0.5 Hz. In order to ensure the accuracy of the results, the EEG signal was preprocessed first, and the 4-30 Hz frequency band of the EEG signal was extracted by FIR digital filter, and the baseline drift was corrected.

The analysis subsystem also has an intelligent classification and recognition algorithm, which converts the EEG signals for processing and analysis, and divides the attention in the analysis results into different levels. Through machine learning, continuous improvement of the recognition algorithm will improve the stability and accuracy of recognition. Finally, the analysis result, that is, the degree of normal state, is transmitted to the driving warning subsystem in real time through the Bluetooth transmission device. The algorithm is mainly implemented in the following steps: 1) Extracting characteristic signals. The characteristic rhythm wave is extracted from the original EEG signal by wavelet transform. 2) Establish feature sample space. The characteristic rhythm waves θ and α collected in the first 1 min were used to construct the respective characteristic sample spaces for subsequent calculations. 3) Calculate the distance between the real-time observation data and the feature sample space, and perform a weighted sum. 4) Pattern matching. Determine whether you are fatigued through the relationship between the above sample space distance and the threshold value. If the matching is successful, it will be judged as fatigue and an alarm will be issued, otherwise continue to monitor.

The specific operation steps include: data receiving, algorithm identification and data transmission.

Step 1: Turn on the attention evaluation subsystem, and the Bluetooth device in the attention evaluation subsystem is automatically connected to the EEG acquisition subsystem to receive data.

Step 2: The states of the subjects in the experiment are mainly divided into two types of events, one is awake state events, and the other is fatigue state events. The main research of the present invention is the state of fatigue, in which both θ and β rhythmic waves will change greatly when entering the fatigued state from the awake state, so the θ and β rhythmic waves are used to form respective corresponding sample spaces, wherein Composition of theta and beta characteristic rhythm waves collected during the initial setting period (preferably 1 minute in the embodiment) , sample space, where the mathematical expectations of the sample values of θ and β are respectively , , the covariance matrix is , then the real-time observation data x to the population sample The Mahalanobis distance is: d(x, )= .

Through the above formula, the Mahalanobis distance from the real-time feature signal to the sample space can be obtained , , and because the Mahalanobis distance of θ and α rhythm waves will change significantly with the change of fatigue state, so the weighting of the Mahalanobis distance of θ and α rhythm waves is considered as the basis for determining the degree of fatigue. Assuming that Mda is the Mahalanobis distance of the θ rhythm wave and Mdb is the Mahalanobis distance of the α rhythm wave, the weighted distance =λ a+(1－λ) b, where 0≤λ≤1, and the threshold for fatigue judgment Then the experimental results are obtained according to the ROC curve analysis when satisfying ≥ , it can be determined that it is in a fatigue state, otherwise it is not fatigued, and the algorithm flow chart is shown in Figure 4.

Step 3: Wireless transmission, transmitting the attention level to the early warning device. The sending end is a Bluetooth host, using a one-to-many sending mode, so that the smart early warning bracelet and the vehicle system can obtain attention levels at the same time. The Bluetooth transmission module uses CC2540 as the main chip, embedded with Bluetooth4.0 protocol, which can directly communicate with smart phones.

3. The driving warning subsystem, which is a self-written mobile phone APP. Refer to Figure 5 for the initial interface of the APP, and the APP can be selected as the general mode or the highway assistance mode. In normal mode, refer to Figure 6, which shows the current normal driving status of the driver, and the bottom side of the screen is the Bluetooth connection refresh button, if the Bluetooth connection is abnormal, you can press this button to reconnect. Its main function is to transmit the driver status report generated by the attention evaluation subsystem to the smart phone or vehicle intelligent system via Bluetooth to feed back to the driver, and to remind the driver to pay attention not to fatigue driving. When in normal state, the phone does not vibrate or ring. When the driver is slightly fatigued, the mobile phone vibrates intermittently, the interface lights up yellow, showing a slightly distracted state and rings intermittently to remind the driver to pay attention; when the driver is seriously fatigued, the phone continues to vibrate, the interface lights red, showing serious Fatigue and continuous ringing reminds the driver to stop driving immediately and take a proper rest.

When the APP is in the highway assist mode, has the functions in the general mode, and has the navigation function, the APP will intelligently find the service area closest to the driver, and navigate the driver so that the driver can go to the service area and get full Rest and avoid traffic accidents.

The beneficial effects of the present invention are: the present invention adopts portable EEG acquisition and analysis equipment, and the EEG acquisition equipment adopts a single-channel EEG acquisition mode, which makes the structure simpler, easy to carry, reduces the complexity of the algorithm, and improves the efficiency of the driver. comfort. Send the EEG signals collected by the EEG acquisition equipment to the attention evaluation subsystem for feature extraction and intelligent algorithm analysis, and use the classification algorithm to divide the driving state of the driver into three levels: normal state, mild fatigue and severe fatigue , and finally transmit the classification results to the driver's smart phone or vehicle system via Bluetooth, and adapt the warning action. The invention truly realizes a practical fatigue driving early warning subsystem, and has high judgment accuracy, timely early warning, convenient portability, easy operation, and can effectively solve the problem of fatigue driving on expressways, and has high practical value.

The preferred embodiments of the present invention have been described in detail above, but the present invention is not limited to the above-mentioned specific embodiments, and those skilled in the art can make modifications or equivalent transformations within the scope of the claims, which should be included in the protection scope of the present invention within.

Claims (10)

1. The driver's attention evaluation and early warning system is associated with the driver's smart phone, and is characterized in that the system composition includes: EEG acquisition subsystem: used for real-time recording of the EEG signal of the driver in the driving state, and signal processing as the source data of attention evaluation; Attention evaluation subsystem: used to filter the source data, extract the time-domain waveform of the required band, and perform algorithm analysis and classification identification to obtain the fatigue state; Driving warning subsystem: it is an APP based on smart phone hardware, which is used to feed back early warning information corresponding to the state of weak concentration through the smart phone according to the obtained driver's fatigue state and combined with the set driving mode; The three subsystems communicate and communicate with each other successively and exchange data. 2. according to the described driver's attention evaluation early warning system of claim 1, it is characterized in that: described EEG acquisition subsystem is unidirectional communication facing attention evaluation subsystem, and described attention evaluation subsystem is unidirectional facing driving warning subsystem Communication, and one-way communication is based on Bluetooth devices. 3. according to the described driver's attention evaluation early warning system of claim 1, it is characterized in that: the EEG acquisition subsystem is a single-channel collector for the fluctuation trend and individual difference dynamic compensation of the EEG signal within the normal range, and the acquisition point Set to the driver's side forehead and the reference point to the driver's side earlobe. 4. according to the described driver's attention evaluation early warning system of claim 1, it is characterized in that the composition of described attention evaluation subsystem comprises: The feature signal extraction module is used to extract the characteristic rhythm waves θ and α from the EEG signal by wavelet transform; A characteristic sample module is used to construct a characteristic sample space using the characteristic rhythm waves θ and α collected during the initial setting period; The distance calculation weighting module is used to calculate the distance between the real-time signal and the feature sample space and weight the summation; The pattern matching module is used to compare the relationship between the obtained distance and the preset threshold, and match to obtain the judgment results of the fatigue state in two or more degrees from concentration to distraction. 5. according to the described driver's attention evaluation early warning system of claim 1, it is characterized in that: described driving early warning subsystem is provided with driving mode input interface and is used to output the acousto-optic unit of early warning signal, vibration unit and automatic navigation based on smart phone unit. 6. The driver's attention evaluation and early warning method is realized by associating with the driver's smart phone, and is characterized in that it includes steps: EEG acquisition: Use the EEG acquisition subsystem to record the driver's EEG signal in real time while driving, and process the signal as the source data for attention evaluation; Attention evaluation: use the attention evaluation subsystem to filter the source data, extract the time-domain waveform of the required band, and perform algorithm analysis and classification identification to obtain the fatigue state; Driving warning: use the APP of the driving warning subsystem based on the hardware of the smart phone, according to the obtained driver fatigue state, combined with the set driving mode to feed back the warning information corresponding to the state of weak concentration through the smart phone; The three subsystems communicate and communicate with each other successively and exchange data. 7. according to the described driver's attention evaluation early warning method of claim 1, it is characterized in that: described EEG acquisition subsystem is unidirectional communication facing attention evaluation subsystem, and described attention evaluation subsystem is unidirectional facing driving warning subsystem Communication, and the one-way communication is realized by Bluetooth. 8. The driver's attention evaluation and early warning method according to claim 1 is characterized in that: the EEG acquisition is aimed at the fluctuation trend and individual differences of the EEG signal within the normal range and is dynamically compensated and adopts a headgear mode to set the acquisition point as the driver's The lateral forehead and the reference point are set to the driver's earlobe, and single-channel acquisition of EEG signals is performed. 9. according to the described driver attention evaluation early warning method of claim 1, it is characterized in that described attention evaluation comprises steps: Extract characteristic signals, and use wavelet transform to extract characteristic rhythm waves θ and α from EEG signals; Establish characteristic samples, and use the characteristic rhythm waves θ and α collected during the initial setting period to construct the characteristic sample space; Distance calculation weighting, calculating the distance between the real-time signal and the feature sample space and weighting the summation; Pattern matching, compare the relationship between the obtained distance and the preset threshold, and match to obtain the fatigue state judgment results of two or more degrees from concentration to distraction. 10. The driver's attention evaluation and early warning method according to claim 1 is characterized in that: in the driving early warning, the driving mode input interface and the acousto-optic unit, the vibration unit and the automatic navigation unit are set based on the smart phone, and the driver's manual input or smart phone intelligent recognition In the driving mode, the smart phone responds to output voice prompts, light and color reminders, vibration stimulation or automatic navigation to guide the driver to the nearest service area to rest according to the judgment result of the driver's fatigue state.