WO2022095985A1 - 一种智能驾驶汽车乘员舒适性评价方法和*** - Google Patents

一种智能驾驶汽车乘员舒适性评价方法和*** Download PDF

Info

Publication number
WO2022095985A1
WO2022095985A1 PCT/CN2021/129177 CN2021129177W WO2022095985A1 WO 2022095985 A1 WO2022095985 A1 WO 2022095985A1 CN 2021129177 W CN2021129177 W CN 2021129177W WO 2022095985 A1 WO2022095985 A1 WO 2022095985A1
Authority
WO
WIPO (PCT)
Prior art keywords
occupant
comfort
evaluation
vehicle
intelligent driving
Prior art date
Application number
PCT/CN2021/129177
Other languages
English (en)
French (fr)
Inventor
刘亚辉
陶书鑫
仇斌
余志超
Original Assignee
清华大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from CN202011237853.9A external-priority patent/CN112353392B/zh
Priority claimed from CN202011237862.8A external-priority patent/CN112353393B/zh
Application filed by 清华大学 filed Critical 清华大学
Publication of WO2022095985A1 publication Critical patent/WO2022095985A1/zh

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state

Definitions

  • the invention relates to the fields of automobile ergonomics, intelligent driving, and human-computer interaction, and in particular relates to an intelligent driving vehicle occupant comfort evaluation that comprehensively considers vehicle driving state data, occupant EMG, ECG, EEG signals and comfort scores method and system.
  • the occupant comfort of smart cars is mostly based on the subjective evaluation of the occupants, and it is impossible to effectively, scientifically and accurately identify the occupant's comfort feeling. Therefore, there is a need for an occupant comfort evaluation method to objectively and quantitatively evaluate the occupant's subjective comfort.
  • Patent CN108742610A discloses a steering comfort evaluation method that realizes the correlation between myoelectricity and subjectivity.
  • a multi-channel electromyographic signal physiological test recorder was used as the signal acquisition facility to measure the electromyographic information of each muscle when each subject performed the steering action.
  • the test driver scored the subjective evaluation form. Perform root mean square processing on the collected EMG signals and set the weights, and normalize the subjective scores of the tested drivers.
  • the corresponding relationship model between physiological information and subjective evaluation is constructed to comprehensively determine the comfort of the driver during the steering process.
  • the muscle parts selected in this patent are only for the parts with obvious changes in muscle activity when the vehicle moves laterally. Therefore, the comfort studied is only for the lateral comfort when changing lanes, which has great limitations.
  • the patent It is one-sided to only measure the EMG signals of the human body.
  • the future development trend of occupant detection is the acquisition of multiple physiological signals of occupants, so as to improve the accuracy of occupant state detection.
  • the purpose of the present invention is to provide a method and system for evaluating the comfort of an occupant of an intelligent driving vehicle, which realizes intelligent driving through the collection of occupant EMG, ECG and EEG signals and the correlation of their subjective comfort scores.
  • the subjective and objective unification of the comfort evaluation of the car occupants reveals the internal connection between the dynamic control of the steering, braking, acceleration, etc. of the intelligent driving car and the subjective comfort of the occupants, and further establishes the intelligent driving car dynamics based on the physiological signals of the occupants.
  • the boundary of control, and based on this, the prediction of the riding comfort of intelligent driving vehicles is realized, and a comprehensive evaluation system for occupant comfort is constructed.
  • the evaluation results are fed back to intelligent driving vehicles to improve the riding comfort of intelligent driving vehicles.
  • the present invention adopts the following technical solutions:
  • a method for evaluating the comfort of an occupant of an intelligent driving vehicle comprising the following steps:
  • the vehicle dynamics-based information is obtained.
  • a comprehensive evaluation model of occupant comfort is constructed.
  • the occupant comfort comprehensive evaluation index predicted by the above occupant comfort comprehensive evaluation model and the vehicle three-degree-of-freedom model are used to establish a vehicle dynamics control domain based on occupant comfort as the corresponding index of intelligent driving vehicle dynamic control to ensure ride comfort. sex.
  • the method for obtaining experimental data comprises the following steps:
  • each tested occupant is seated on the seat of the smart vehicle, the smart vehicle is driving in a specified scene, and each sign signal collection electrode collects the sign signal of the tested occupant. While the experiment is in progress, each tested occupant will be scored in real time according to the comfort evaluation table .
  • the EMG collection electrodes are respectively placed on the sternocleidomastoid muscle, trapezius muscle, rectus abdominis, external oblique muscle and latissimus dorsi of the tested passenger. muscle.
  • the method for establishing an objective evaluation model of comfort based on physical sign information includes the following steps:
  • the extracted comfort evaluation index includes an electromyography evaluation index, an electrocardiographic evaluation index and an electroencephalogram evaluation index; the electromyography evaluation index includes an average muscle activation degree and an electromyographic signal fluctuation degree index,
  • the ECG evaluation index includes R-R interval standard deviation and R-R difference standard deviation index, and the EEG evaluation index includes alpha wave energy ratio and delta wave to alpha wave energy ratio index.
  • N represents the number of selected muscle parts
  • MA i represents the muscle activation degree of the ith detection part
  • E i represents the root mean square value of the muscle EMG signal of the ith detection part
  • R i represents the ith detection part
  • the muscle of the part occupies the weight of the EMG information
  • R ⁇ is the energy ratio of ⁇ wave
  • E ⁇ is the energy of ⁇ wave
  • E ⁇ is the energy of ⁇ wave
  • E ⁇ is the energy of ⁇ wave
  • K ⁇ - ⁇ is the energy ratio of ⁇ wave to ⁇ wave
  • E ⁇ is the energy of ⁇ wave
  • E ⁇ is the energy of ⁇ wave.
  • a 1 , a 2 , and a 3 are the weight coefficient of the subjective evaluation index of the occupant, the weight coefficient of the objective evaluation index based on the occupant's sign information, and the weight coefficient of the occupant comfort prediction index based on the vehicle dynamics
  • x 1 (n), x 2 (n), x 3 (n) are the calculated values of the above three evaluation indexes at the current moment
  • x 1 (nm), x 2 (nm), x 3 (nm) are the first m of the above three evaluation indexes Calculated value at time
  • f() is a dynamic weight function.
  • the basic description of the occupant comfort comprehensive evaluation model is as follows:
  • K is the comprehensive evaluation index value of occupant comfort
  • a 1 , a 2 , and a 3 are the weight coefficient of the subjective evaluation index of the occupant, the weight coefficient of the objective evaluation index based on the occupant physical information, and the prediction of the occupant comfort based on the vehicle dynamics.
  • Index weight coefficient; x 1 (n) is the occupant subjective score, time series array; x 2 (n) is the objective evaluation score based on occupant physical information, time series array; x 3 (n) is the occupant comfort prediction based on vehicle dynamics scores, time series array.
  • An intelligent driving vehicle occupant comfort evaluation system comprising:
  • Occupant physical sign detection system Occupant physical sign detection system, occupant comfort evaluation feedback system, vehicle state acquisition system and computer;
  • the occupant sign detection system is used to measure the occupant's sign information during the driving of the intelligent driving vehicle, including the occupant's EMG, EEG and ECG information;
  • the occupant comfort evaluation feedback system is used to collect the subjective comfort evaluation scores of the occupants during the driving process of the intelligent driving vehicle;
  • the vehicle state acquisition system is used to collect and measure the dynamic information of the intelligent driving vehicle during the intelligent driving process, including the three-axis speed and the three-axis acceleration of the intelligent driving vehicle;
  • the computer is used to predict the occupant comfort according to the received occupant physical information, subjective comfort evaluation score and vehicle dynamics information, and feed it back to the intelligent driving system of the intelligent driving vehicle to provide automatic driving of the intelligent driving vehicle. refer to.
  • the computer is provided with a communication module, a data receiving and recording module, a data processing and analysis module, and an interaction and display module;
  • the communication module is used to realize the computer and the occupant physical sign detection system, the occupant comfort evaluation feedback system, and the vehicle status.
  • the communication between the acquisition system and the intelligent driving system of the intelligent driving vehicle; the data receiving and recording module is used for real-time, synchronous acquisition and storage of occupant physical information, occupant subjective comfort evaluation information and dynamic information of the intelligent driving vehicle;
  • the data processing and analysis module is used to process the acquired occupant physical information, occupant subjective comfort evaluation information and vehicle dynamics information, and combined with the preset evaluation model, identify the occupant comfort state at the current moment and predict the future preset time.
  • the occupant state according to the occupant, and the recognition accuracy rate of the evaluation model is analyzed in combination with the occupant's subjective comfort evaluation information; the interaction and display module is used to realize the interaction between the evaluation system and the operator.
  • the data processing and analysis module includes a preprocessing module, a feature extraction module, a model application module, a vehicle state prediction module, an accuracy rate analysis and a model correction module;
  • the preprocessing module uses resampling and comprehensive filtering to preprocess the collected occupant physical information and vehicle dynamics information;
  • the feature extraction module is used for extracting physiological indexes and vehicle dynamics indexes from the preprocessed occupant physical information and vehicle dynamics information;
  • the model application module is used for inputting the extracted physiological indicators, vehicle dynamics indicators and occupant subjective comfort evaluation results into three preset evaluation models, and finally obtains a comprehensive evaluation index of occupant comfort;
  • the vehicle state prediction module It includes a vehicle dynamics model, which is used to predict the vehicle state in the future;
  • the accuracy analysis and model correction module is used to compare the results of the three evaluation models with the subjective comfort evaluation results of the occupants, and analyze the preset time window And the correct rate of occupant comfort recognition in the global scope, and through the model parameter modification interface, the operator can modify the model parameters by himself or automatically adjust the model parameters according to the program.
  • the feature extraction module extracts the feature value, for the occupant sign signal data, the average muscle activation degree of the electromyographic signal and the fluctuation degree of the electromyographic signal, the R-R of the electrocardiogram signal are extracted according to the feature extraction method peculiar to different physiological information.
  • the standard deviation of the interval and the standard deviation of the R-R difference, the energy ratio of the alpha wave and the energy ratio of the delta wave to the alpha wave in the power spectrum of the EEG signal are used as the eigenvalues of the occupant's physical information; for the vehicle state data, the corresponding time window is extracted
  • the longitudinal acceleration and the mean square value of the acceleration are taken as the characteristic value of the vehicle state.
  • the three preset evaluation models in the model applicable module are:
  • the objective evaluation model of occupant comfort based on physical information the input is the occupant physiological index extracted at the current moment, and the output is the objective evaluation index of occupant comfort at the next moment;
  • the occupant comfort prediction model based on vehicle dynamics consists of two parts.
  • the first part is the input of vehicle dynamics index and control information, and the output is the vehicle dynamics index at the next moment;
  • the second part of the input is the vehicle power at the next moment.
  • the comprehensive evaluation index of the occupant comfort at the previous moment is obtained, and the output is the occupant comfort prediction result at the next moment;
  • the comprehensive evaluation model of occupant comfort whose input is the objective evaluation index of occupant comfort at the next moment, the prediction result of occupant comfort at the next moment, and the subjective comfort evaluation result of the occupant at the current moment, and the output is the comprehensive evaluation index of occupant comfort at the next moment .
  • the occupant physical sign detection system includes a human electromyographic signal detection system, a human electrocardiographic signal detection system, and a human electroencephalographic signal detection system; the human electromyographic signal detection system is used to measure the generation of the detected occupants when they ride in the intelligent driving car.
  • the human electromyographic signal; the human electrocardiographic signal detection system is used to measure the human electrocardiographic signal generated when the occupant under test rides in the intelligent driving car; The human brain electrical signal generated during the time; finally sent to the computer through the computer port.
  • the human EMG signal detection system includes an EMG acquisition electrode and an EMG instrument; the EMG acquisition electrode is placed at the position of the corresponding muscle group of the human body, including the sternocleidomastoid muscle, the trapezius muscle, the rectus abdominis, and the rectus abdominis.
  • the latissimus dorsi is used to measure human EMG signals, and is connected with the EMG instrument through wireless transmission equipment; the EMG instrument is connected with the computer set in the intelligent driving vehicle through its microcomputer interface, so as to collect the collected data.
  • the EMG signal is transmitted to the computer.
  • the human body ECG signal detection system includes an ECG acquisition electrode and an electrocardiograph; the ECG acquisition electrode is placed at a corresponding position on the chest of the human body to measure the human body ECG signal, and communicates with the ECG through a wireless transmission device.
  • the electrocardiograph is connected with the computer through its microcomputer interface, so as to transmit the collected electrocardiographic signal to the computer.
  • the human EEG signal detection system includes an EEG acquisition electrode and an EEG instrument; the EEG acquisition electrode is placed at a corresponding position on the head of the human body to measure the human EEG signal, and communicates with the brain through a wireless transmission device.
  • the EEG is connected to the computer through its microcomputer interface, so as to transmit the collected EEG signals to the computer.
  • the vehicle state acquisition system includes a six-axis acceleration sensor and a vehicle state synchronization module, the six-axis acceleration sensor is arranged at the center of mass of the intelligent driving vehicle, and is used for collecting the lateral and longitudinal acceleration data of the intelligent driving vehicle; the The vehicle status synchronization module is connected with the vehicle OBD interface, and is used to obtain the vehicle driving status data on the CAN bus of the intelligent driving vehicle.
  • the occupant comfort evaluation method of the present invention comprehensively analyzes and processes the occupant's ECG, EMG, and EEG signals, extracts the corresponding ECG comfort index, EMG comfort index, and EEG comfort index, and finally The three types of indicators are used to objectively evaluate the occupant's riding comfort. Compared with the identification of the occupant's comfort through a single physiological signal, the accuracy and reliability of the objective evaluation of the occupant's comfort are improved.
  • the present invention integrates subjective evaluation of occupants, objective evaluation based on physiological information, and prediction based on vehicle dynamic characteristics, including three dimensions: subjective, physiological, and vehicle, and includes the consideration of the current moment, the past moment, and the future moment. more systematic and perfect in the evaluation of occupant comfort.
  • the comfort evaluation method of the present invention is not a comfort evaluation method for a typical working condition, but an occupant comfort identification method for intelligent driving under all working conditions. When the vehicle accelerates, brakes, turns, and changes lanes can be evaluated by the comfort evaluation method of the present invention.
  • the present invention can be widely used in the field of intelligent vehicle occupant comfort evaluation.
  • Fig. 1 is a system block diagram of an intelligent driving vehicle occupant comfort evaluation system of the present invention
  • Fig. 2 is a structural diagram of an intelligent driving vehicle occupant comfort evaluation system of the present invention
  • Fig. 3 is the electromyographic signal acquisition electrode setting position diagram of a kind of intelligent driving vehicle occupant state detection system of the present invention
  • Fig. 4 is the scoring software interface of a kind of intelligent driving vehicle occupant state detection system of the present invention
  • FIG. 5 is a flow chart of a method for evaluating the comfort of an intelligent driving vehicle occupant according to the present invention
  • Fig. 6 is the physical sign index of a kind of intelligent driving vehicle occupant comfort evaluation method of the present invention.
  • FIG. 7 is a schematic diagram of the composition of the comprehensive evaluation index of the occupant comfort of the smart car according to the present invention.
  • FIG. 8 is a schematic structural diagram of a vehicle dynamics-based member comfort prediction model of the present invention.
  • the marks in the figure are as follows: 1. Intelligent vehicle controller; 2. Intelligent equipment; 3. Electroencephalogram acquisition electrode; 4. EEG; 5. Electrocardiograph; 6. EMG; 7. Computer; Driving a car; 9. Seat; 10. EMG acquisition electrode; 11. ECG acquisition electrode.
  • this embodiment provides an evaluation system for passenger comfort of an intelligent driving car, which can be installed in the intelligent driving car and connected with the intelligent driving system of the intelligent driving car.
  • the evaluation system provided in this embodiment can not only obtain the vehicle state information during the intelligent driving process by using the sensors of the intelligent vehicle itself, but also measure the occupant's vital sign information during the intelligent driving process, so as to detect and analyze the occupant state in the intelligent vehicle, in order to improve the intelligent
  • the occupant comfort provides a theoretical basis. Specifically, it includes an occupant physical sign detection system, an occupant comfort evaluation feedback system, a vehicle state acquisition system, and a computer.
  • the occupant physical sign detection system is used to measure the occupant's physical information during the driving process of the intelligent driving car, including the occupant's EMG, EEG and ECG information;
  • the occupant comfort evaluation feedback system is used to collect the occupant's information during the driving process of the intelligent driving car.
  • Subjective comfort evaluation score is used to collect and measure the dynamic information of the intelligent driving vehicle during the intelligent driving process, including the three-axis speed and acceleration of the intelligent driving vehicle;
  • the comfort evaluation score and vehicle dynamics information can predict the comfort of the occupants, and feed them back to the intelligent driving system of the intelligent driving vehicle to provide a reference for the automatic driving of the intelligent driving vehicle.
  • the computer 7 is provided with a communication module, a data receiving and recording module, a data processing and analysis module, and an interaction and display module.
  • the communication module is used to realize the communication between the computer and the occupant physical sign detection system, the occupant comfort evaluation feedback system, the vehicle state acquisition system and the intelligent driving system of the intelligent driving car;
  • the data receiving and recording module is used for real-time, synchronous collection and Store occupant physical information, occupant subjective comfort evaluation information, and dynamic information of intelligent driving vehicles;
  • the data processing and analysis module is used to process the acquired occupant physical information, occupant subjective comfort evaluation information and vehicle dynamics information, combined with preset
  • the evaluation model of the occupant can identify the occupant's comfort state at the current moment and predict the occupant's state at a preset time in the future.
  • the recognition accuracy of the evaluation model is analyzed; the interaction and display module is used to realize The interaction between the evaluation system and the operator includes two parts: one is to display all the collected information and key variables and results in the process of data processing and analysis in real time; the other is to provide system control buttons and parameter adjustment interfaces, and the operator can The results control the operation of the evaluation system or adjust the parameters of the evaluation model.
  • the data processing and analysis module includes a preprocessing module, a feature extraction module, a model application module, a vehicle state prediction module, an accuracy rate analysis and a model correction module.
  • the preprocessing module uses resampling and comprehensive filtering to preprocess the collected occupant sign information and vehicle dynamics information to improve the signal-to-noise ratio and reduce interference;
  • the feature extraction module is used to extract the preprocessed occupant sign information and vehicle dynamics information.
  • Physiological indexes and vehicle dynamics indexes are extracted from the dynamics information;
  • the model application module is used to input the extracted physiological indexes, vehicle dynamics indexes and occupant subjective comfort evaluation results into three preset evaluation models, and finally obtain occupant comfort.
  • the vehicle state prediction module includes the vehicle dynamics model, which is used to predict the vehicle state at the future time to further predict the future occupant comfort; the accuracy rate analysis and model correction module is used for the above three evaluations.
  • the results of the model are compared with the subjective comfort evaluation results of the occupants, and the preset time window and the correct rate of occupant comfort recognition in the global range are analyzed, and the model parameter modification interface is used to allow the operator to modify the model parameters by himself or automatically adjust the model according to the program. parameter.
  • the feature extraction module when the feature extraction module extracts physiological indicators and vehicle dynamics indicators, the occupant physical information is extracted according to the characteristic extraction methods unique to different physiological information, such as extracting the R-R interval standard deviation of the ECG signal, and the EEG.
  • the ratio of ⁇ -band energy to ⁇ -band energy in the power spectrum of the signal is used as a physiological index; for vehicle dynamics information, the longitudinal acceleration and acceleration mean square value in the corresponding time window are extracted as vehicle dynamics indicators.
  • model application module which are:
  • An objective evaluation model of occupant comfort based on physical information whose input is the occupant physiological index extracted at the current moment, and the output is the objective evaluation index of occupant comfort at the next moment;
  • the occupant comfort prediction model based on vehicle dynamics, which consists of two parts, the first part is the input of vehicle dynamics index and control information, and the output is the vehicle dynamics index at the next moment; the second part is the input of the next moment.
  • the vehicle dynamics index and the comprehensive evaluation index of the occupant comfort at the previous moment, and the output is the occupant comfort prediction result at the next moment;
  • the comprehensive evaluation model of occupant comfort whose input is the objective evaluation index of occupant comfort at the next moment, the prediction result of occupant comfort at the next moment, and the subjective comfort evaluation result of the occupant at the current moment, and the output is the comprehensive occupant comfort at the next moment. evaluation indicators.
  • the occupant sign detection system includes a human electromyographic signal detection system, a human electrocardiographic signal detection system, and a human electroencephalographic signal detection system; the human electromyographic signal detection system is used to measure the occupant under test.
  • the human EMG signal detection system includes an EMG acquisition electrode 9 and an EMG meter 6; wherein, the EMG acquisition electrode is placed in the position of the corresponding muscle group of the human body, including the sternocleidomastoid.
  • Muscle SCM
  • TraP trapezius
  • RA rectus abdominis
  • LD latissimus dorsi
  • the microcomputer interface is connected with the computer equipment arranged in the intelligent vehicle, so as to transmit the collected EMG signals to the computer equipment.
  • the human electrocardiogram signal detection system includes an electrocardiogram acquisition electrode 10 and an electrocardiograph 5 .
  • the ECG collection electrode is placed at the corresponding position of the human chest (the placement position is a well-known technology to those skilled in the art, and the present invention will not repeat it here) to measure the human body ECG signal, and is connected to the electrocardiograph through a wireless transmission device;
  • the instrument is connected with the computer equipment through its microcomputer interface, so as to transmit the collected ECG signals to the computer equipment.
  • the human EEG signal detection system includes an EEG acquisition electrode 3 and an EEG instrument 4; This is not repeated here) to measure the human EEG signal, and connect with the EEG instrument through the wireless transmission device, and the EEG instrument is connected with the computer device through its microcomputer interface, thereby transmitting the collected EEG signal to the computer device.
  • the occupant comfort evaluation feedback system adopts a smart device (such as a smart phone) installed with a scoring App, and the smart device is connected to the computer 7 through WiFi.
  • a smart device such as a smart phone
  • the occupant holds the smart device and sends each scoring situation to the computer 7 according to the preset cycle.
  • the occupants can also perform real-time scoring according to real-time strong stimulation when scoring.
  • the vehicle state acquisition system includes a six-axis acceleration sensor and a vehicle state synchronization module, and the six-axis acceleration sensor is arranged near the center of mass of the intelligent driving vehicle and is used to collect the horizontal and longitudinal acceleration data of the intelligent driving vehicle; the vehicle state The synchronization module is connected to the OBD interface of the vehicle, and is used to obtain vehicle driving status data on the CAN bus of the intelligent driving vehicle, such as vehicle dynamics information and control information such as vehicle speed, steering wheel angle, throttle opening, brake master cylinder pressure, etc.
  • the present invention provides a method for evaluating the occupant comfort of an intelligent driving vehicle, forming a subjective-physiological-vehicle integrated occupant comfort evaluation, which mainly includes the following four parts: 1. The occupant’s subjective evaluation 2. Objective evaluation based on occupant physical information; 3. Occupant comfort prediction based on vehicle dynamics; 4. Dynamic weight function. Specifically, it includes the following steps:
  • the corresponding dynamic index is extracted, and based on the dynamic index and the pre-established vehicle dynamics-based occupant comfort prediction model, the vehicle dynamics-based information is obtained.
  • the occupant comfort prediction evaluation index is used to characterize the response characteristics and sensitivity of various occupant physiological indicators to vehicle dynamics indicators;
  • a comprehensive evaluation model of occupant comfort is constructed.
  • the occupant comfort comprehensive evaluation index predicted by the occupant comfort comprehensive evaluation model and the vehicle three-degree-of-freedom model are used to establish a vehicle dynamics control domain based on occupant comfort, including the comfort domain, transition domain and discomfort domain, as an intelligent driving vehicle. Corresponding indicators of dynamic control to ensure ride comfort.
  • step 1) the method for obtaining experimental data, comprises the following steps:
  • the sign signal acquisition equipment includes electromyography, electrocardiography, and electroencephalography, and the electromyography, electrocardiography, and electroencephalography are respectively connected to the electromyography, electrocardiography, and EEG acquisition electrodes are connected.
  • the collected sign signal data includes EMG signal, ECG signal and EEG signal data.
  • each sign signal collection electrode collects the physical information of the tested occupant. While the experiment is in progress, each tested occupant is comfortable according to the comfort evaluation table. Sexual evaluation scores.
  • the EMG collection electrodes are respectively placed on the sternocleidomastoid muscle (SCM), trapezius muscle (TRAP), rectus abdominis (RA) of the tested occupant in this application. ) and latissimus dorsi (LD).
  • SCM sternocleidomastoid muscle
  • TRAP trapezius muscle
  • RA rectus abdominis
  • LD latissimus dorsi
  • the process of scoring the tested occupant according to the comfort evaluation table is: during the driving process of the intelligent vehicle, every 15s or after the tested occupant receives a certain stimulus (steering, acceleration, braking), through the installation Smart devices (smartphones) with scoring software score the changes in their own comfort caused by stimulation. , severe, very poor, poor, dividing line, barely acceptable, average, good, very good, excellent, the occupant physical data and comfort score within 15s constitute a sample, when a sufficient number of samples are collected, the experiment ends.
  • the comfort evaluation table in the present invention is shown in Table 1 below.
  • the subjective comfort score can also introduce different intervals according to actual needs, such as introducing a 0.25-point interval, and the occupants can make subtle adjustments to the score according to the actual situation to obtain more and more accurate comfort scores.
  • the method for establishing an objective evaluation model of comfort based on physical sign information includes the following steps:
  • the preprocessing of the sign information includes resampling and noise reduction filtering
  • the extracted objective indicators of the occupant comfort evaluation include the electromyography evaluation index, the electrocardiographic evaluation index and the EEG evaluation.
  • the EMG evaluation index includes the average muscle activation degree and the EMG signal fluctuation index
  • the ECG evaluation index includes the R-R interval standard deviation and the R-R difference standard deviation index
  • the EEG evaluation index includes the alpha wave energy ratio and the delta wave. and alpha wave energy ratio indicator.
  • Each indicator is defined as follows:
  • the degree of muscle activation reflects the strength of each muscle when the occupant rides the smart vehicle and resists vibration from all directions of the body.
  • the formula for calculating the degree of muscle activation MA is:
  • RMS Test is the RMS value of the currently measured EMG signal
  • RMS MVC is the RMS value of the EMG signal during the maximum voluntary muscle contraction
  • MA is the degree of muscle activation.
  • EMG(t) is the voltage value measured by the electromyographic signal collection
  • T is the length of the time window, which is 0.05s.
  • the muscle activation degree of each detection site is calculated, and the average muscle activation degree MA mean is obtained by adding up:
  • N represents the number of selected muscle parts
  • MA i is the muscle activation degree of the ith detection part.
  • the fluctuation range F of the EMG signal refers to the ratio of the maximum value of the root mean square value of each part of the muscle to the average value in one evaluation period.
  • RMS max is the maximum value of the RMS of the muscle part in one evaluation cycle
  • RMS mean is the mean value of the RMS of the muscle part in one evaluation cycle.
  • each muscle signal of the tested occupant average the root mean square value of the muscle signals of different tested occupants at the same position, and each calculated average value can be recorded as the sternocleidomastoid muscle RMS 1 , trapezius RMS 2 , rectus abdominis RMS 3 and latissimus dorsi RMS 4 , the weight of EMG information occupied by each muscle can be recorded as:
  • E i represents the root mean square value of each muscle EMG signal.
  • the standard deviation of the R-R interval is used to describe the degree of change in the heart rate of the tested occupant when he encounters an uncomfortable event stimulus during the ride in an intelligent vehicle.
  • the formula for calculating the standard deviation of the R-R interval is as follows:
  • SDNN is the standard deviation of the RR interval
  • N is the total number of heartbeats
  • RR i is the ith RR interval
  • RR mean is the average of N RR intervals.
  • Adjacent R-R difference standard deviation RMSSD represents the root mean square of the adjacent R-R interval difference, which reflects the change of adjacent R-R interval and represents the rapid change degree of HRV signal.
  • the calculation formula is:
  • the energy ratio of ⁇ wave represents the energy ratio of ⁇ wave (7 ⁇ 13Hz) to the total frequency band, and its calculation formula is:
  • R ⁇ is the energy ratio of the ⁇ wave; E ⁇ is the energy of the ⁇ wave (1-4 Hz); E ⁇ is the energy of the theta wave (4-7 Hz); E ⁇ is the energy of the ⁇ -wave; E ⁇ is the ⁇ -wave (13-25 Hz) )energy.
  • the ratio of delta wave to alpha wave energy represents the ratio of delta wave energy to alpha wave energy, and its calculation formula is:
  • K ⁇ - ⁇ is the energy ratio of ⁇ wave to ⁇ wave
  • E ⁇ is the energy of ⁇ wave (1 ⁇ 4Hz)
  • E ⁇ is the energy of ⁇ wave.
  • the extracted dynamic index of the intelligent driving vehicle includes the average vehicle speed, the root mean square value of the three-axis acceleration, the root mean square value of the yaw angular velocity, and the root mean square value of the yaw angular acceleration.
  • a dynamic weighting function is designed, according to the occupant's subjective comfort evaluation index obtained in step 1), the objective evaluation index of comfort based on the sign signal obtained in step 2), and Step 3)
  • the obtained real-time change of the occupant comfort prediction evaluation index based on the vehicle dynamics information calculates the corresponding weight coefficient, and the weighted calculation obtains the comprehensive evaluation index of the occupant comfort. This index will be fed back to the intelligent driving system as Refer to the regulation of dynamic control to improve the ride comfort of the vehicle.
  • a 1 , a 2 , and a 3 are the weight coefficient of the occupant's subjective evaluation index, the weight coefficient of the objective evaluation index based on the occupant's physical information, and the weight coefficient of the occupant's comfort prediction index based on vehicle dynamics
  • x 1 ( n), x 2 (n), x 3 (n) are the calculated values of the above three evaluation indexes at the current moment
  • x 1 (nm), x 2 (nm), x 3 (nm) are the above three evaluation indexes
  • the calculated value of the first m time; f() is a dynamic weight function.
  • x 1 (n) is the subjective score of the occupant, time series array
  • x 2 (n) is the objective evaluation score based on the occupant's physical information, time series array
  • x 3 (n) is the occupant comfort prediction based on vehicle dynamics information Score, time series array; according to the three comfort scores at the current moment and the comfort score at the previous moment, determine the weight ratio of the three scores. Over time, the weights also change.
  • a three-degree-of-freedom model of the vehicle is constructed according to the characteristics of the intelligent vehicle's own vehicle, and a state space is established.
  • This control domain and the comprehensive evaluation index of occupant comfort at the current moment predict the occupant comfort in the future, and form an occupant comfort prediction model based on vehicle dynamics, and the output results are incorporated into the occupant comfort evaluation system.
  • the three-degree-of-freedom model of the vehicle includes longitudinal translation, lateral translation and yaw motion of the vehicle, but does not include vertical motion, roll motion and pitch motion; state quantities include longitudinal velocity, lateral velocity and yaw angular velocity; control quantity Including steering wheel angle, accelerator opening, braking intensity, etc.
  • control domain based on ride comfort is the range of intelligent vehicle dynamics characteristics that will result in comfort or discomfort for the occupants.
  • the construction of the three-degree-of-freedom model of the vehicle is a well-known technology by those skilled in the art, and the present invention will not repeat them here.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Hospice & Palliative Care (AREA)
  • Pathology (AREA)
  • Developmental Disabilities (AREA)
  • Psychiatry (AREA)
  • Psychology (AREA)
  • Social Psychology (AREA)
  • Physics & Mathematics (AREA)
  • Child & Adolescent Psychology (AREA)
  • Biophysics (AREA)
  • Educational Technology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

一种智能驾驶汽车（8）乘员舒适性评价方法和***，方法包括以下步骤：1）获取实验数据，包括各被测乘员在静止状态下和智能驾驶汽车（8）行驶状态下的体征信息以及被测乘员的主观舒适性评价指标；2）基于获取的体征信息和预先建立的基于体征信息的舒适性客观评价模型，计算基于体征信号的舒适性客观评价指标；3）基于动力学指标和基于车辆动力学的乘员舒适性预测模型，得到基于车辆动力学信息的乘员舒适性预测评价指标；4）构建乘员舒适性综合评价模型，基于乘员舒适性综合评价模型预测得到的乘员舒适性综合评价指标以及车辆三自由度模型，建立基于乘员舒适性的车辆动力学控制域，以保证乘坐舒适性。该方法可以广泛应用于智能车辆乘员舒适性评价领域。

Description

一种智能驾驶汽车乘员舒适性评价方法和*** 技术领域
本发明涉及汽车人机工效学、智能驾驶、人机交互领域，具体涉及一种综合考虑车辆行驶状态数据，乘员肌电、心电、脑电信号以及舒适性评分的智能驾驶汽车乘员舒适性评价方法和***。
背景技术
随着智能驾驶技术的发展，除了安全性以外，人们的关注点越来越集中在智能驾驶汽车的舒适性上。造成乘员舒适性问题的诱因是由于智能驾驶功能代替了传统的驾驶员控制，缺少驾驶人根据自身舒适性对车辆进行主动调控的人机环。为了弥补上缺失的人机环首先需要对智能驾驶车辆内的乘员状态进行准确的检测。传统车辆乘员舒适性研究大多集中于对乘员姿势舒适性研究，驾驶员转向操纵舒适性研究。智能驾驶汽车的舒适性主要是研究智能驾驶汽车加速，制动，转向中乘员舒适性感受。目前对于智能车乘员舒适性多为通过乘员主观评价进行，无法对乘员舒适性感受进行有效，科学，准确的识别。因此需要一种乘员舒适性评价方法，对乘员的主观舒适性进行客观量化的评价。
专利CN108742610A公开了一种实现肌电和主观相关联的转向舒适度评价方法，首先选择受试驾驶员在转头时颈部及上背部主要发力肌肉和驾驶员转动方向盘时肩部发力肌肉作为待测肌肉,采用多通道肌电信号生理测试记录仪为信号采集设施,测量各受试者进行转向动作时每块肌肉的肌电信息其次,受试驾驶员进行主观评价表的打分而后,对采集到的肌电信号进行均方根处理并设置权重，对受试驾驶员的主观评分进行归一化处理最终,构建生理信息与主观评价对应关系模型，综合确定驾驶员转向过程中的舒适度，该专利选取的肌肉部位只是针对在车辆横向运动时，肌肉活性变化比较明显的部位，因而所研究的舒适性也仅是针对换道时的横向舒适性，局限性较大，同时该专利只测量人体的肌电信号进行研究，具有片面性，未来乘员检测的发展趋势是乘员多生理信号采集，以此来提高乘员状态检测的准确性。
发明内容
针对上述问题，本发明的目的是提供一种智能驾驶汽车乘员舒适性评价方法和***，通过对乘员肌电、心电和脑电信号的采集以及其主观舒适性评分相关联，实现了智能驾驶汽车乘员的舒适性评价的主客观统一，揭示了智能驾驶汽车转向、制动、加速等动力学控制与乘员主观舒适性之间的内在联系，进一步建立了基于乘员生理信号的智能驾驶汽车动力学控制的边界，并以此为依据实现了对智能驾驶车辆乘坐舒适性 的预测，构建了关于乘员舒适性的综合评价***，评价结果反馈于智能驾驶汽车，提升智能驾驶汽车的乘坐舒适性。
为实现上述目的，本发明采取以下技术方案：
一种智能驾驶汽车乘员舒适性评价方法，其包括以下步骤：
1)获取实验数据，包括各被测乘员在静止状态下的体征信息、智能驾驶汽车行驶状态下的体征信息以及被测乘员的主观舒适性评价指标以及车辆动力学信息；
2)对获取的体征信息进行预处理后提取相应的生理指标，并基于该生理指标和预先建立的基于体征信息的舒适性客观评价模型，得到基于体征信号的舒适性客观评价指标；
3)对获取的智能驾驶汽车的动力学信息进行预处理后提取相应的动力学指标，并基于该动力学指标和预先建立的基于车辆动力学的乘员舒适性预测模型，得到基于车辆动力学信息的乘员舒适性预测评价指标；
4)根据得到的乘员的主观舒适性评价指标、基于体征信号的舒适性客观评价指标、基于车辆动力学信息的乘员舒适性预测评价指标以及动态权重函数，构建乘员舒适性综合评价模型，基于所述乘员舒适性综合评价模型预测得到的乘员舒适性综合评价指标以及车辆三自由度模型，建立基于乘员舒适性的车辆动力学控制域，作为智能驾驶汽车动力学控制的相应指标，以保证乘坐舒适性。
进一步，所述步骤1)中，获取实验数据的方法，包括以下步骤：
1.1)选取若干乘员，并在每一被测乘员的相应位置处安放体征信号采集电极，开启测试体征信号采集设备，确定各采集设备工作正常，并记录下各被测乘员处于静止状态下的体征信号数据；
1.2)被测乘员用力晃动头部，耸动肩膀，使得颈部、肩部、背部肌肉进行最大自主收缩运动，确保各体征信号采集电极不发生脱落，同时记录肌电信号最大自主收缩状态下的峰值；
1.3)被测乘员在智能车辆座位上落座，智能车辆在规定场景下行驶，各体征信号采集电极采集被测乘员的体征信号，在实验进行同时，各被测乘员根据舒适性评价表进行实时打分。
进一步，所述步骤1.1)中，对肌电信号进行采集时，将肌电采集电极分别放置在被测乘员的胸锁乳突肌，斜方肌，腹直肌，腹外斜肌和背阔肌。
进一步，所述步骤2)中，建立基于体征信息的舒适性客观评价模型的方法，包括以下步骤：
2.1)对获取的体征信息进行预处理后提取相应的生理指标作为乘员舒适性评价的客观指标；
2.2)对提取的乘员舒适性评价客观指标以及舒适性评分数据进行多元回归分析，得到舒适性评价客观指标与舒适性评分数据之间的映射关系，即基于体征信息的舒适性客观评价模型。
进一步，所述步骤2.1)中，提取的舒适性评价指标包括肌电评价指标、心电评价指标和脑电评价指标；所述肌电评价指标包括平均肌肉激活程度和肌电信号波动程度指标，所述心电评价指标包括R-R间期标准差和R-R差值标准差指标，所述脑电评价指标包括α波能量比和δ波与α波能量比指标。
进一步，所述平均肌肉激活程度指标MA mean和所述肌电信号的波动范围F的计算公式分别为：
Figure PCTCN2021129177-appb-000001
Figure PCTCN2021129177-appb-000002
其中，N表示选取的肌肉部位个数，MA i表示第i个检测部位的肌肉激活程度；E i表示第i个检测部位的肌肉肌电信号的均方根值，R i表示第i个检测部位的肌肉所占肌电信息权重；
所述R-R间期标准差和所述相邻R-R差值标准差的计算公式分别为：
Figure PCTCN2021129177-appb-000003
Figure PCTCN2021129177-appb-000004
式中，SDNN为R-R间期标准差；N为心搏总数；RR i为第i个R-R间期；RR mean为N个R-R间期的平均值；RR i、RR i-1分别为相邻的R-R间期，i为整数且i>＝1；
所述α波能量比和所述δ波与α波能量比的计算公式分别为：
Figure PCTCN2021129177-appb-000005
Figure PCTCN2021129177-appb-000006
其中,R α为α波能量比；E σ为σ波能量；E θ为θ波能量；E α为α波能量；E β为β波能量；K δ-α为δ波与α波能量比；E σ为σ波能量；E α为α波能量。
进一步，所述步骤4)中，动态权重函数的基本描述如下：
[a 1(n)，a 2(n)，a 3(n)]
＝f(x 1(n)，x 2(n)，x 3(n)，x 1(n-1)，x 2(n-1)，x 3(n-1)，...，x 1(n-m)，x 2(n-m)，x 3(n-m))
式中，a 1，a 2，a 3分别为乘员主观评价指标权重系数、基于乘员体征信息的客观评价指标权重系数以及基于车辆动力学的乘员舒适性预测指标权重系数，x 1(n)，x 2(n)，x 3(n)为上述三种评价指标的当前时刻的计算值，x 1(n-m)，x 2(n-m)，x 3(n-m)为上述三种评价指标的前m时刻的计算值；f()为动态权重函数。
进一步，所述步骤4)中，所述乘员舒适性综合评价模型的基本描述如下：
K＝a 1(n)·x 1(n)+a 2(n)·x 2(n)+a 3(n)·x 3(n)
式中，K为乘员舒适性综合评价指标值；a 1，a 2，a 3分别为乘员主观评价指标权重系数、基于乘员体征信息的客观评价指标权重系数以及基于车辆动力学的乘员舒适性预测指标权重系数；x 1(n)为乘员主观评分,时序数组；x 2(n)为基于乘员体征信息的客观评价得分，时序数组；x 3(n)为基于车辆动力学的乘员舒适性预测得分，时序数组。
一种智能驾驶汽车乘员舒适性评价***，其包括：
乘员体征检测***、乘员舒适性评价反馈***、车辆状态采集***以及计算机；
所述乘员体征检测***用于测量智能驾驶汽车行驶过程中乘员的体征信息，包括乘员的肌电、脑电和心电信息；
所述乘员舒适性评价反馈***用于收集智能驾驶汽车行驶过程中乘员的主观舒适性评价得分；
所述车辆状态采集***用于收集测量智能驾驶过程中智能驾驶汽车的动力学信息，包括智能驾驶汽车的三轴速度和三轴加速度；
所述计算机用于根据接收到的乘员体征信息、主观舒适性评价得分以及车辆动力学信息，对乘员舒适性进行预测，并反馈于智能驾驶车辆的智能驾驶***，为智能驾驶车辆的自动驾驶提供参考。
进一步，所述计算机中设置有通讯模块、数据接收与记录模块、数据处理分析模块以及交互与显示模块；所述通讯模块用于实现计算机与乘员体征检测***、乘员舒适性评价反馈***、车辆状态采集***以及智能驾驶汽车的智能驾驶***之间的通信；所述数据接收与记录模块用于实时、同步采集并存储乘员体征信息、乘员主观舒适性评价信息以及智能驾驶汽车的动力学信息；所述数据处理分析模块用于对获取的乘员体征信息、乘员主观舒适性评价信息以及车辆动力学信息进行处理，结合预设的评价模型，识别当前时刻的乘员舒适性状态并预测未来预设时刻内的乘员状态，同时结合乘员的主观舒适性评价信息，对评价模型的识别正确率进行分析；所述交互与显示模 块用于实现本评价***与操作员的交互。
进一步，所述数据处理分析模块包括预处理模块、特征提取模块、模型适用模块、车辆状态预测模块、正确率分析及模型修正模块；
所述预处理模块采用重采样与综合滤波对采集的乘员体征信息与车辆动力学信息进行预处理；
所述特征提取模块用于从预处理后的乘员体征信息以及车辆动力学信息中提取生理指标和车辆动力学指标；
所述模型适用模块用于将提取到的生理指标、车辆动力学指标以及乘员主观舒适性评价结果输入预设的三种评价模型，并最终得到乘员舒适性综合评价指标；所述车辆状态预测模块包含车辆动力学模型，用于预测未来时刻的车辆状态；所述正确率分析及模型修正模块用于将所述三种评价模型的结果与乘员主观舒适性评价结果相对比，分析预设时间窗以及全局范围内的乘员舒适性识别正确率，并通过模型参数修改接口，实现操作员自行修改模型参数或根据程序自动调整模型参数。
进一步，所述特征提取模块提取特征值时，对于乘员体征信号数据，则依据不同生理信息特有的特征提取方法，提取肌电信号的平均肌肉激活程度和肌电信号波动程度，心电信号的R-R间期标准差和R-R差值标准差，脑电信号的功率谱中的α波能量比和δ波与α波能量比，作为乘员体征信息的特征值；对于车辆状态数据，则提取相应时间窗内的纵向加速度以及加速度均方值作为车辆状态特征值。
进一步，所述模型适用模块中预设的三种评价模型分别为：
基于体征信息的乘员舒适性客观评价模型，其输入为当前时刻提取的乘员生理指标，输出为下一时刻乘员舒适性客观评价指标；
基于车辆动力学的乘员舒适性预测模型，其包含两部分，第一部分输入为车辆动力学指标以及控制信息，输出为下一时刻的车辆动力学指标；第二部分输入为下一时刻的车辆动力学指标与上一时刻乘员舒适性综合评价指标，输出为下一时刻的乘员舒适性预测结果；
乘员舒适性综合评价模型，其输入为下一时刻乘员舒适性客观评价指标、下一时刻的乘员舒适性预测结果、当前时刻乘员主观舒适性评价结果，输出为下一时刻乘员舒适性综合评价指标。
进一步，所述乘员体征检测***包括人体肌电信号检测***、人体心电信号检测***和人体脑电信号检测***；所述人体肌电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体肌电信号；所述人体心电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体心电信号；所述人体脑电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体脑电信号；最后通过计算机端口发送给所述计算机。
进一步，所述人体肌电信号检测***包括肌电采集电极和肌电仪；所述肌电采集电极安放在人体的相应肌群位置,包括胸锁乳突肌，斜方肌，腹直肌和背阔肌，以测量人体肌电信号，并通过无线传输设备与所述肌电仪相连；所述肌电仪通过其微机接口与设置在智能驾驶车辆中的所述计算机相连，从而将采集到的肌电信号传给所述计算机。
进一步，所述人体心电信号检测***包括心电采集电极和心电仪；所述心电采集电极安放在人体胸口的相应位置以测量人体心电信号，并通过无线传输设备与所述心电仪相连；所述心电仪通过其微机接口与所述计算机相连，从而将采集到的心电信号传给所述计算机。
进一步，所述人体脑电信号检测***包括脑电采集电极和脑电仪；所述脑电采集电极安放在人体头部的相应位置以测量人体脑电信号，并通过无线传输设备与所述脑电仪相连，所述脑电仪通过其微机接口与所述计算机相连，从而将采集到的脑电信号传给所述计算机。
进一步，所述车辆状态采集***包括六轴加速度传感器以及车辆状态同步模块，所述六轴加速度传感器设置在智能驾驶汽车的质心处，用于对智能驾驶汽车的横纵向加速度数据进行采集；所述车辆状态同步模块与车辆OBD接口相连，用于获取智能驾驶汽车CAN总线上的车辆行驶状态数据。
本发明由于采取以上技术方案，其具有以下优点：
1、本发明的乘员舒适性评价方法对于乘员的心电，肌电，脑电信号进行了综合分析处理，提取相应的心电舒适性指标，肌电舒适性指标，脑电舒适性指标，最后综合三类指标对于乘员乘坐舒适性进行客观评价，相对于通过单一生理信号对乘员舒适性进行辨别，提高了乘员舒适性客观评价的准确性，可靠性。
2、本发明综合了乘员主观评价、基于生理信息的客观评价以及基于车辆动力学特征的预测，包含主观、生理、车辆三种维度，并且纳入了对当前时刻、过去时刻的考量以及对未来时刻的预测，在乘员舒适性评价方面更加***和完善。
3、本发明的舒适性评价方法，不是针对某一典型工况的舒适性评价方法，而是面向智能驾驶全工况下的乘员舒适性识别方法，当车辆加速，制动，转向，换道时，都可通过本发明的舒适性评价方法进行评价。
因而，本发明可以广泛应用于智能车辆乘员舒适性评价领域。
附图说明
图1为本发明的一种智能驾驶汽车乘员舒适性评价***的***框图；
图2为本发明的一种智能驾驶汽车乘员舒适性评价***的结构图；
图3为本发明的一种智能驾驶汽车乘员状态检测***的肌电信号采集电极设置位 置图；
图4为本发明的一种智能驾驶汽车乘员状态检测***的打分软件界面；
图5为本发明的一种智能驾驶汽车乘员舒适性评价方法的流程图；
图6为本发明的一种智能驾驶汽车乘员舒适性评价方法的体征指标；
图7为本发明的智能车乘员舒适性综合评价指标的构成示意图；
图8为本发明的基于车辆动力学的成员舒适性预测模型的结构示意图；
图中各标记如下：1、智能车控制器；2、智能设备；3、脑电采集电极；4、脑电仪；5、心电仪；6、肌电仪；7、计算机；8、智能驾驶汽车；9、座椅；10、肌电采集电极；11、心电采集电极。
具体实施方式
下面结合附图和实施例对本发明进行详细的描述。
实施例一
如图1，图2所示，本实施例提供一种智能驾驶汽车乘员舒适性评价***，可以安装到智能驾驶汽车内部，并与智能驾驶汽车的智能驾驶***连接。本实施例提供的评价***既能利用智能车自身传感器获取智能驾驶过程中的车辆状态信息，又可以测量智能驾驶过程中乘员体征信息，从而对智能车辆中的乘员状态进行检测分析，为提高智能车乘员舒适性提供理论依据。具体的，其包括乘员体征检测***、乘员舒适性评价反馈***、车辆状态采集***以及计算机。其中，乘员体征检测***用于测量智能驾驶汽车行驶过程中乘员的体征信息，包括乘员的肌电、脑电和心电信息；乘员舒适性评价反馈***用于收集智能驾驶汽车行驶过程中乘员的主观舒适性评价得分；车辆状态采集***用于收集测量智能驾驶过程中智能驾驶汽车的动力学信息，包括智能驾驶汽车的三轴速度以及加速度等；计算机用于根据接收到的乘员体征信息、主观舒适性评价得分以及车辆动力学信息，对乘员舒适性进行预测，并反馈于智能驾驶车辆的智能驾驶***，为智能驾驶车辆的自动驾驶提供参考。
上述实施例中，如图1所示，计算机7中设置有通讯模块、数据接收与记录模块、数据处理分析模块以及交互与显示模块。其中，通讯模块用于实现计算机与乘员体征检测***、乘员舒适性评价反馈***、车辆状态采集***以及智能驾驶汽车的智能驾驶***之间的通信；数据接收与记录模块用于实时、同步采集并存储乘员体征信息、乘员主观舒适性评价信息以及智能驾驶汽车的动力学信息；数据处理分析模块用于对获取的乘员体征信息、乘员主观舒适性评价信息以及车辆动力学信息进行处理，结合预设的评价模型，识别当前时刻的乘员舒适性状态并预测未来预设时刻内的乘员状态，同时结合乘员的主观舒适性评价信息，对评价模型的识别正确率进行分析；交互与显示模块用于实现本评价***与操作员的交互，包括两部分：一是实时显示所有采集信 息以及数据处理与分析过程中的关键变量和结果；二是提供***控制按钮以及参数调整接口，操作员可根据实时显示结果控制评价***的运行或调整评价模型参数。
上述各实施例中，如图1所示，数据处理分析模块包括预处理模块、特征提取模块、模型适用模块、车辆状态预测模块、正确率分析及模型修正模块。其中，预处理模块采用重采样与综合滤波对采集的乘员体征信息与车辆动力学信息进行预处理，以提高信噪比，降低干扰；特征提取模块用于从预处理后的乘员体征信息以及车辆动力学信息中提取生理指标和车辆动力学指标；模型适用模块用于将提取到的生理指标、车辆动力学指标以及乘员主观舒适性评价结果输入预设的三种评价模型，并最终得到乘员舒适性综合评价指标；车辆状态预测模块包含车辆动力学模型，用于预测未来时刻的车辆状态，以进一步地对未来的乘员舒适性进行预测；正确率分析及模型修正模块用于将上述三种评价模型的结果与乘员主观舒适性评价结果相对比，分析预设时间窗以及全局范围内的乘员舒适性识别正确率，并通过模型参数修改接口，实现操作员自行修改模型参数或根据程序自动调整模型参数。
上述各实施例中，特征提取模块提取生理指标和车辆动力学指标时，对于乘员体征信息则依据不同生理信息特有的特征提取方法进行提取，如提取心电信号的R-R间期标准差、脑电信号的功率谱中的α频段能量与β频段能量比值等，作为生理指标；对于车辆动力学信息，则提取相应时间窗内的纵向加速度以及加速度均方值等，作为车辆动力学指标。
上述各实施例中，模型适用模块中预存有三种评价模型，分别为：
1.基于体征信息的乘员舒适性客观评价模型，其输入为当前时刻提取的乘员生理指标，输出为下一时刻乘员舒适性客观评价指标；
2.基于车辆动力学的乘员舒适性预测模型，其包含两部分，第一部分输入为车辆动力学指标以及控制信息，输出为下一时刻的车辆动力学指标；第二部分输入为下一时刻的车辆动力学指标与上一时刻乘员舒适性综合评价指标，输出为下一时刻的乘员舒适性预测结果；
3.乘员舒适性综合评价模型，其输入为下一时刻乘员舒适性客观评价指标、下一时刻的乘员舒适性预测结果、当前时刻乘员主观舒适性评价结果，输出为下一时刻乘员舒适性综合评价指标。
上述各实施例中，如图2所示，乘员体征检测***包括人体肌电信号检测***、人体心电信号检测***和人体脑电信号检测***；人体肌电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体肌电信号；人体心电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体心电信号；人体脑电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体脑电信号；最后通过计算机端口发送给计算机 7。
上述各实施例中，如图3所示，人体肌电信号检测***包括肌电采集电极9和肌电仪6；其中，肌电采集电极安放在人体的相应肌群位置,包括胸锁乳突肌(SCM)，斜方肌(TRAP)，腹直肌(RA)和背阔肌(LD)，以测量人体肌电信号，并通过无线传输设备与肌电仪相连；肌电仪6通过其微机接口与设置在智能车辆中的计算机设备相连，从而将采集到的肌电信号传给计算机设备。
上述各实施例中，人体心电信号检测***包括心电采集电极10和心电仪5。心电采集电极安放在人体胸口的相应位置(安放位置为本领域技术人员公知技术，本发明在此不再赘述)以测量人体心电信号，并通过无线传输设备与心电仪相连；心电仪通过其微机接口与计算机设备相连，从而将采集到的心电信号传给计算机设备。
上述各实施例中，人体脑电信号检测***包括脑电采集电极3和脑电仪4；脑电采集电极安放在人体头部的相应位置(安放位置为本领域技术人员公知技术，本发明在此不再赘述)以测量人体脑电信号，并通过无线传输设备与脑电仪相连，脑电仪通过其微机接口与计算机设备相连，从而将采集到的脑电信号传给计算机设备。
上述各实施例中，如图4所示，乘员舒适性评价反馈***采用安装有打分App的智能设备(如智能手机)，该智能设备通过WiFi与计算机7相连，在智能驾驶汽车行驶过程中，乘员手持智能设备，按照预设周期将每一次打分情况发送给计算机7。其中，乘员进行打分时也可以根据实时的强烈刺激进行实时打分。
上述各实施例中，车辆状态采集***包括六轴加速度传感器以及车辆状态同步模块，六轴加速度传感器设置在智能驾驶汽车的质心附近，用于对智能驾驶汽车的横纵向加速度数据进行采集；车辆状态同步模块与车辆OBD接口相连，用于获取智能驾驶汽车CAN总线上的车辆行驶状态数据，比如车速、方向盘转角、气节门开度、制动主缸压力等车辆动力学信息和控制信息。
实施例二
如图5～图8所示，本发明提供了一种智能驾驶汽车乘员舒适性评价方法，形成了主观-生理-车辆一体化的乘员舒适性评价，主要包括以下四部分，1、乘员主观评价；2、基于乘员体征信息的客观评价；3、基于车辆动力学的乘员舒适性预测；4、动态权重函数。具体的，包括以下步骤：
1)获取实验数据，包括各被测乘员在静止状态下的体征信息、智能驾驶汽车行驶状态下的体征信息以及被测乘员的主观舒适性评价指标；
2)对获取的体征信息进行预处理后提取相应的生理指标，并基于该生理指标和预先建立的基于体征信息的舒适性客观评价模型，得到基于体征信号的舒适性客观评价指标；
3)对获取的智能驾驶汽车的动力学信息进行预处理后提取相应的动力学指标，并基于该动力学指标和预先建立的基于车辆动力学的乘员舒适性预测模型，得到基于车辆动力学信息的乘员舒适性预测评价指标，以表征乘员各项生理指标对车辆动力学指标的响应特征及敏感程度；
4)根据得到的乘员的主观舒适性评价指标、基于体征信号的舒适性客观评价指标、基于车辆动力学信息的乘员舒适性预测评价指标以及动态权重函数，构建乘员舒适性综合评价模型，基于所述乘员舒适性综合评价模型预测得到的乘员舒适性综合评价指标以及车辆三自由度模型，建立基于乘员舒适性的车辆动力学控制域，包含舒适域、过渡域以及不舒适域，作为智能驾驶汽车动力学控制的相应指标，以保证乘坐舒适性。
上述步骤1)中，获取实验数据的方法，包括以下步骤：
1.1)选取若干乘员，并在每一被测乘员身体的相应位置处安放体征信号采集电极，开启测试体征信号采集设备，确定各体征信号采集设备工作正常，并记录下各被测乘员处于静止状态下的体征信息。其中，体征信号采集设备包括肌电仪、心电仪和脑电仪，且肌电仪、心电仪和脑电仪分别与安放在被测乘员身上的肌电采集电极、心电采集电极和脑电采集电极相连。采集的体征信号数据包括肌电信号、心电信号和脑电信号数据。
1.2)各被测乘员用力晃动头部，耸动肩膀，使得颈部、肩部、背部肌肉进行最大自主收缩运动，确保各体征信号采集电极不发生脱落，同时记录肌电信号最大自主收缩状态下的峰值。
1.3)被测乘员在智能车辆座位上落座，智能车辆在预设场景下行驶，各体征信号采集电极采集被测乘员的体征信息，在实验进行同时，各被测乘员根据舒适性评价表进行舒适性评价打分。
1.4)在车辆质心附近安装六轴加速度传感器对智能驾驶汽车的，以及通过车辆OBD口获取车辆速度、方向盘转角、节气门开度、制动主缸压力等车辆动力学信息和控制信息。
上述步骤1.1)中，对肌电信号进行采集时，本申请中将肌电采集电极分别放置在被测乘员的胸锁乳突肌(SCM)，斜方肌(TRAP)，腹直肌(RA)和背阔肌(LD)。对心电信号和脑电信号采集时，心电采集电极和脑电采集电极的安放位置与常规进行心电信号采集和脑电信号采集时的安放位置相同，本发明不再赘述。
上述步骤1.3)中，被测乘员根据舒适性评价表打分的过程为：在智能车辆行驶过程中，每隔15s或者在被测乘员接受某个刺激后(转向，加速，制动)，通过安装有打分软件的智能设备(智能手机)对刺激所造成的自身舒适性改变进行打分，分值范围为1、2、3、4、5、6、7、8、9、10，分别表示无法接受、剧烈、非常差、差、分 界线、勉强接受、一般、好、很好、极好，15s内的乘员体征数据和舒适性评分构成一个样本，当采集到足够数量的样本之后，实验结束。本发明中的舒适性评价表如下表1所示。
表1 舒适性评价表
Figure PCTCN2021129177-appb-000007
其中，舒适性主观评分也可以根据实际需要引入不同的间隔，例如引入0.25分间隔，乘员根据实际情况对分值进行细微的调整，以获得更多更精确的舒适性评分。
上述步骤2)中，建立基于体征信息的舒适性客观评价模型的方法，包括以下步骤：
2.1)对获取的体征信息进行预处理后提取相应的生理指标作为乘员舒适性评价的客观指标；
2.2)对提取的乘员舒适性评价客观指标以及舒适性评分数据进行多元回归分析，得到舒适性评价客观指标与舒适性评分数据之间的映射关系，即基于体征信息的舒适性客观评价模型。
上述步骤2.1)中，如图6所示，对体征信息进行预处理时包括重采样与降噪滤波，提取的乘员舒适性评价的客观指标包括肌电评价指标、心电评价指标和脑电评价指标，且肌电评价指标包括平均肌肉激活程度和肌电信号波动程度指标，心电评价指标包括R-R间期标准差和R-R差值标准差指标，脑电评价指标包括α波能量比和δ波与α波能量比指标。各指标定义如下：
2.1.1)肌电评价指标
①平均肌肉激活程度指标MA mean
肌肉激活程度反映了乘员乘坐智能车辆时，在抵抗来自车身各个方向振动时，各肌肉的发力程度。肌肉激活程度MA的计算公式为：
Figure PCTCN2021129177-appb-000008
式中，RMS Test为当前所测肌电信号的RMS值；RMS MVC为肌肉最大自主收缩时肌电信号的RMS值；MA为肌肉激活程度。
其中，RMS值的计算公式为：
Figure PCTCN2021129177-appb-000009
式中，EMG(t)为肌电信号电集所测的电压值；T为时间窗的长度，取0.05s。
将每次评价周期中，各检测部位的肌肉激活程度求出，相加得到平均肌肉激活程度MA mean：
Figure PCTCN2021129177-appb-000010
其中，N表示选取的肌肉部位个数；MA i为第i个检测部位的肌肉激活程度。
②肌电信号的波动范围F：
肌电信号的波动范围F是指：在一个评价周期中，各部位肌肉均方根值最大值与平均值的比值。
首先，对评价周期内各部位的肌肉信号进行归一化处理,公式如下:
Figure PCTCN2021129177-appb-000011
式中，RMS max为该肌肉部位在一个评价周期中RMS的最大值；RMS mean为该肌肉部位在一个评价周期中RMS的均值。
其次，根据所求出的被测乘员各肌肉信号的均方根值，将不同被测乘员同一位置肌肉信号的均方根值求平均值，计算得到的各平均数值可记为胸锁乳突肌RMS 1，斜方肌RMS 2，腹直肌RMS 3和背阔肌RMS 4，则各肌肉所占肌电信息权重可记为：
Figure PCTCN2021129177-appb-000012
最后，根据总体肌电信号波动范围为：
Figure PCTCN2021129177-appb-000013
其中，E i表示各肌肉肌电信号的均方根值。
2.1.2)心电评价指标
①R-R间期标准差
R-R间期标准差用于描述被测乘员在乘坐智能车辆过程中，遇到不舒适事件刺激时，心跳速率变化的程度。R-R间期标准差的计算公式如下：
Figure PCTCN2021129177-appb-000014
式中，SDNN为R-R间期标准差；N为心搏总数；RR i为第i个R-R间期；RR mean为N个R-R间期的平均值。
②相邻R-R差值标准差RMSSD
相邻R-R差值标准差RMSSD表示相邻R-R间期差值的均方根，反应相邻R-R间期的变动，代表HRV信号的快速变动程度，其计算公式为：
Figure PCTCN2021129177-appb-000015
式中，N为心搏总数；RR i、RR i-1分别为相邻的R-R间期，i为整数且i>＝1。
2.1.3)脑电评价指标
①α波能量比
α波能量比表示α波(7～13Hz)占总频带能量比，其计算公式为：
Figure PCTCN2021129177-appb-000016
其中,R α为α波能量比；E σ为σ波(1～4Hz)能量；E θ为θ波(4～7Hz)能量；E α为α波能量；E β为β波(13～25Hz)能量。
②δ波与α波能量比
δ波与α波能量比表示δ波能量与α波能量的比值，其计算公式为：
Figure PCTCN2021129177-appb-000017
其中，K δ-α为δ波与α波能量比；E σ为σ波(1～4Hz)能量；E α为α波能量。
进一步地，上述步骤3)中，提取的智能驾驶汽车的动力学指标包括平均车速、三轴加速度均方根值、横摆角速度均方根值以及横摆角加速度均方根值。
进一步地，上述步骤4)中，如图7所示，设计一条动态权重函数，依据步骤1)得到的乘员主观舒适性评价指标、步骤2)得到的基于体征信号的舒适性客观评价指标、以及步骤3)得到的基于车辆动力学信息的乘员舒适性预测评价指标的实时变化计算得出对应的权重系数，加权计算得出乘员舒适性的综合评价指标，此指标将反馈于智能驾驶***，作为参考对动力学控制进行调控，提高车辆的乘坐舒适性。
其中，动态权重函数的基本描述如下：
[a 1(n)，a 2(n)，a 3(n)]
＝f(x1(n)，x2(n)，x3(n)，x1(n-1)，x2(n-1)，x3(n-1)，...，x1(n-m)，x2(n-m)，x3(n-m))
                                              (11) 式中，a 1，a 2，a 3分别为乘员主观评价指标权重系数、基于乘员体征信息的客观评价指标权重系数以及基于车辆动力学的乘员舒适性预测指标权重系数，x 1(n)，x 2(n)，x 3(n)为上述三种评价指标的当前时刻的计算值，x 1(n-m)，x 2(n-m)，x 3(n-m)为上述三种评价指标的前m时刻的计算值；f()为动态权重函数。
最终得到的乘员舒适性综合评价模型及其指标K的基本描述如下：
K＝a 1(n)·x 1(n)+a 2(n)·x 2(n)+a 3(n)·x 3(n)     (12)
式中，x 1(n)为乘员主观评分,时序数组；x 2(n)为基于乘员体征信息的客观评价得分，时序数组；x 3(n)为基于车辆动力学信息的乘员舒适性预测得分，时序数组；根据当前时刻三种舒适性得分，以及前一时刻舒适性得分，确定3种评分的权重占比。随时间变化，权重也发生变化。
如图8所示，根据智能驾驶汽车的自身车辆特性构建车辆三自由度模型，建立状态空间，根据智能车辆当前时刻的状态量以及控制量，对未来时刻的车辆动力学特征进行预测，并依据此控制域以及当前时刻的乘员舒适性综合评价指标对未来时刻的乘员舒适性进行预测，形成基于车辆动力学的乘员舒适性预测模型，输出结果纳入乘员舒适性评价***中。其中，车辆三自由度模型包含车辆纵向平动、侧向平动以及横摆运动，不包含垂向运动、侧倾运动以及俯仰运动；状态量包括纵向速度、侧向速度以及横摆角速度；控制量包括方向盘转角、油门开度、制动强度等。基于乘坐舒适性的控制域的含义是将会导致成员舒适或不舒适的智能车辆动力学特征的范围。车辆三自由度模型的构建为本领域技术人员公知技术，本发明在此不再赘述。
上述实施例只为说明本发明的技术构思及特点，其目的在于让熟悉此项技术的人士能够了解本发明的内容并据以实施，并不能以此限制本发明的保护范围。凡根据本发明精神实质所作的等效变化或修饰，都应涵盖在本发明的保护范围之内。

Claims (18)

  1. 一种智能驾驶汽车乘员舒适性评价方法，其特征在于，包括以下步骤：
    1)获取实验数据，包括各被测乘员在静止状态下的体征信息、智能驾驶汽车行驶状态下的体征信息以及被测乘员的主观舒适性评价指标以及车辆动力学信息；
    2)对获取的体征信息进行预处理后提取相应的生理指标，并基于该生理指标和预先建立的基于体征信息的舒适性客观评价模型，得到基于体征信号的舒适性客观评价指标；
    3)对获取的智能驾驶汽车的动力学信息进行预处理后提取相应的动力学指标，并基于该动力学指标和预先建立的基于车辆动力学的乘员舒适性预测模型，得到基于车辆动力学信息的乘员舒适性预测评价指标；
    4)根据得到的乘员的主观舒适性评价指标、基于体征信号的舒适性客观评价指标、基于车辆动力学信息的乘员舒适性预测评价指标以及动态权重函数，构建乘员舒适性综合评价模型，基于所述乘员舒适性综合评价模型预测得到的乘员舒适性综合评价指标以及车辆三自由度模型，建立基于乘员舒适性的车辆动力学控制域，作为智能驾驶汽车动力学控制的相应指标，以保证乘坐舒适性。
  2. 如权利要求1所述的一种智能驾驶汽车乘员舒适性评价方法，其特征在于：所述步骤1)中，获取实验数据的方法，包括以下步骤：
    1.1)选取若干乘员，并在每一被测乘员的相应位置处安放体征信号采集电极，开启测试体征信号采集设备，确定各采集设备工作正常，并记录下各被测乘员处于静止状态下的体征信号数据；
    1.2)被测乘员用力晃动头部，耸动肩膀，使得颈部、肩部、背部肌肉进行最大自主收缩运动，确保各体征信号采集电极不发生脱落，同时记录肌电信号最大自主收缩状态下的峰值；
    1.3)被测乘员在智能车辆座位上落座，智能车辆在规定场景下行驶，各体征信号采集电极采集被测乘员的体征信号，在实验进行同时，各被测乘员根据舒适性评价表进行实时打分。
  3. 如权利要求1所述的一种智能驾驶汽车乘员舒适性评价方法，其特征在于：所述步骤1.1)中，对肌电信号进行采集时，将肌电采集电极分别放置在被测乘员的胸锁乳突肌，斜方肌，腹直肌，腹外斜肌和背阔肌。
  4. 如权利要求1所述的一种智能驾驶汽车乘员舒适性评价方法，其特征在于：所述步骤2)中，建立基于体征信息的舒适性客观评价模型的方法，包括以下步骤：
    2.1)对获取的体征信息进行预处理后提取相应的生理指标作为乘员舒适性评价的 客观指标；
    2.2)对提取的乘员舒适性评价客观指标以及舒适性评分数据进行多元回归分析，得到舒适性评价客观指标与舒适性评分数据之间的映射关系，即基于体征信息的舒适性客观评价模型。
  5. 如权利要求1所述的一种智能驾驶汽车乘员舒适性评价方法，其特征在于：所述步骤2.1)中，提取的舒适性评价指标包括肌电评价指标、心电评价指标和脑电评价指标；所述肌电评价指标包括平均肌肉激活程度和肌电信号波动程度指标，所述心电评价指标包括R-R间期标准差和R-R差值标准差指标，所述脑电评价指标包括α波能量比和δ波与α波能量比指标。
  6. 如权利要求5所述的一种智能驾驶汽车乘员舒适性评价方法，其特征在于：所述平均肌肉激活程度指标MA mean和所述肌电信号的波动范围F的计算公式分别为：
    Figure PCTCN2021129177-appb-100001
    Figure PCTCN2021129177-appb-100002
    其中，N表示选取的肌肉部位个数，MA i表示第i个检测部位的肌肉激活程度；E i表示第i个检测部位的肌肉肌电信号的均方根值，R i表示第i个检测部位的肌肉所占肌电信息权重；
    所述R-R间期标准差和所述相邻R-R差值标准差的计算公式分别为：
    Figure PCTCN2021129177-appb-100003
    Figure PCTCN2021129177-appb-100004
    式中，SDNN为R-R间期标准差；N为心搏总数；RR i为第i个R-R间期；RR mean为N个R-R间期的平均值；RR i、RR i-1分别为相邻的R-R间期，i为整数且i>＝1；
    所述α波能量比和所述δ波与α波能量比的计算公式分别为：
    Figure PCTCN2021129177-appb-100005
    Figure PCTCN2021129177-appb-100006
    其中,R α为α波能量比；E σ为σ波能量；E θ为θ波能量；E α为α波能量；E β为β波能量；K δ-α为δ波与α波能量比；E σ为σ波能量；E α为α波能量。
  7. 如权利要求1所述的一种智能驾驶汽车乘员舒适性评价方法，其特征在于：所述步骤4)中，动态权重函数的基本描述如下：
    [a 1(n)，a 2(n)，a 3(n)]＝f(x 1(n)，x 2(n)，x 3(n)，x 1(n-1)，x 2(n-1)，x 3(n-1)，...，x 1(n-m)，x 2(n-m)，x 3(n-m))
    式中，a 1，a 2，a 3分别为乘员主观评价指标权重系数、基于乘员体征信息的客观评价指标权重系数以及基于车辆动力学的乘员舒适性预测指标权重系数，x 1(n)，x 2(n)，x 3(n)为上述三种评价指标的当前时刻的计算值，x 1(n-m)，x 2(n-m)，x 3(n-m)为上述三种评价指标的前m时刻的计算值；f()为动态权重函数。
  8. 如权利要求1所述的一种智能驾驶汽车乘员舒适性评价方法，其特征在于：所述步骤4)中，所述乘员舒适性综合评价模型的基本描述如下：
    K＝a 1(n)·x 1(n)+a 2(n)·x 2(n)+a 3(n)·x 3(n)
    式中，K为乘员舒适性综合评价指标值；a 1，a 2，a 3分别为乘员主观评价指标权重系数、基于乘员体征信息的客观评价指标权重系数以及基于车辆动力学的乘员舒适性预测指标权重系数；x 1(n)为乘员主观评分,时序数组；x 2(n)为基于乘员体征信息的客观评价得分，时序数组；x 3(n)为基于车辆动力学的乘员舒适性预测得分，时序数组。
  9. 一种智能驾驶汽车乘员舒适性评价***，其特征在于，包括：
    乘员体征检测***、乘员舒适性评价反馈***、车辆状态采集***以及计算机；
    所述乘员体征检测***用于测量智能驾驶汽车行驶过程中乘员的体征信息，包括乘员的肌电、脑电和心电信息；
    所述乘员舒适性评价反馈***用于收集智能驾驶汽车行驶过程中乘员的主观舒适性评价得分；
    所述车辆状态采集***用于收集测量智能驾驶过程中智能驾驶汽车的动力学信息，包括智能驾驶汽车的三轴速度和三轴加速度；
    所述计算机用于根据接收到的乘员体征信息、主观舒适性评价得分以及车辆动力学信息，对乘员舒适性进行预测，并反馈于智能驾驶车辆的智能驾驶***，为智能驾驶车辆的自动驾驶提供参考。
  10. 如权利要求9所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述计算机中设置有通讯模块、数据接收与记录模块、数据处理分析模块以及交互与显示模块；所述通讯模块用于实现计算机与乘员体征检测***、乘员舒适性评价反馈***、车辆状态采集***以及智能驾驶汽车的智能驾驶***之间的通信；所述数据接 收与记录模块用于实时、同步采集并存储乘员体征信息、乘员主观舒适性评价信息以及智能驾驶汽车的动力学信息；所述数据处理分析模块用于对获取的乘员体征信息、乘员主观舒适性评价信息以及车辆动力学信息进行处理，结合预设的评价模型，识别当前时刻的乘员舒适性状态并预测未来预设时刻内的乘员状态，同时结合乘员的主观舒适性评价信息，对评价模型的识别正确率进行分析；所述交互与显示模块用于实现本评价***与操作员的交互。
  11. 如权利要求10所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述数据处理分析模块包括预处理模块、特征提取模块、模型适用模块、车辆状态预测模块、正确率分析及模型修正模块；
    所述预处理模块采用重采样与综合滤波对采集的乘员体征信息与车辆动力学信息进行预处理；
    所述特征提取模块用于从预处理后的乘员体征信息以及车辆动力学信息中提取生理指标和车辆动力学指标；
    所述模型适用模块用于将提取到的生理指标、车辆动力学指标以及乘员主观舒适性评价结果输入预设的三种评价模型，并最终得到乘员舒适性综合评价指标；所述车辆状态预测模块包含车辆动力学模型，用于预测未来时刻的车辆状态；所述正确率分析及模型修正模块用于将所述三种评价模型的结果与乘员主观舒适性评价结果相对比，分析预设时间窗以及全局范围内的乘员舒适性识别正确率，并通过模型参数修改接口，实现操作员自行修改模型参数或根据程序自动调整模型参数。
  12. 如权利要求11所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述特征提取模块提取特征值时，对于乘员体征信号数据，则依据不同生理信息特有的特征提取方法，提取肌电信号的平均肌肉激活程度和肌电信号波动程度，心电信号的R-R间期标准差和R-R差值标准差，脑电信号的功率谱中的α波能量比和δ波与α波能量比，作为乘员体征信息的特征值；对于车辆状态数据，则提取相应时间窗内的纵向加速度以及加速度均方值作为车辆状态特征值。
  13. 如权利要求11所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述模型适用模块中预设的三种评价模型分别为：
    基于体征信息的乘员舒适性客观评价模型，其输入为当前时刻提取的乘员生理指标，输出为下一时刻乘员舒适性客观评价指标；
    基于车辆动力学的乘员舒适性预测模型，其包含两部分，第一部分输入为车辆动力学指标以及控制信息，输出为下一时刻的车辆动力学指标；第二部分输入为下一时刻的车辆动力学指标与上一时刻乘员舒适性综合评价指标，输出为下一时刻的乘员舒适性预测结果；
    乘员舒适性综合评价模型，其输入为下一时刻乘员舒适性客观评价指标、下一时刻的乘员舒适性预测结果、当前时刻乘员主观舒适性评价结果，输出为下一时刻乘员舒适性综合评价指标。
  14. 如权利要求9所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述乘员体征检测***包括人体肌电信号检测***、人体心电信号检测***和人体脑电信号检测***；所述人体肌电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体肌电信号；所述人体心电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体心电信号；所述人体脑电信号检测***用于测量被测乘员乘坐智能驾驶汽车时产生的人体脑电信号；最后通过计算机端口发送给所述计算机。
  15. 如权利要求14所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述人体肌电信号检测***包括肌电采集电极和肌电仪；所述肌电采集电极安放在人体的相应肌群位置,包括胸锁乳突肌，斜方肌，腹直肌和背阔肌，以测量人体肌电信号，并通过无线传输设备与所述肌电仪相连；所述肌电仪通过其微机接口与设置在智能驾驶车辆中的所述计算机相连，从而将采集到的肌电信号传给所述计算机。
  16. 如权利要求14所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述人体心电信号检测***包括心电采集电极和心电仪；所述心电采集电极安放在人体胸口的相应位置以测量人体心电信号，并通过无线传输设备与所述心电仪相连；所述心电仪通过其微机接口与所述计算机相连，从而将采集到的心电信号传给所述计算机。
  17. 如权利要求14所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述人体脑电信号检测***包括脑电采集电极和脑电仪；所述脑电采集电极安放在人体头部的相应位置以测量人体脑电信号，并通过无线传输设备与所述脑电仪相连，所述脑电仪通过其微机接口与所述计算机相连，从而将采集到的脑电信号传给所述计算机。
  18. 如权利要求9所述的一种智能驾驶汽车乘员舒适性评价***，其特征在于，所述车辆状态采集***包括六轴加速度传感器以及车辆状态同步模块，所述六轴加速度传感器设置在智能驾驶汽车的质心处，用于对智能驾驶汽车的横纵向加速度数据进行采集；所述车辆状态同步模块与车辆OBD接口相连，用于获取智能驾驶汽车CAN总线上的车辆行驶状态数据。
PCT/CN2021/129177 2020-11-09 2021-11-08 一种智能驾驶汽车乘员舒适性评价方法和*** WO2022095985A1 (zh)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN202011237853.9A CN112353392B (zh) 2020-11-09 2020-11-09 一种智能驾驶汽车乘员舒适性评价方法
CN202011237862.8A CN112353393B (zh) 2020-11-09 2020-11-09 一种智能驾驶汽车乘员状态检测***
CN202011237853.9 2020-11-09
CN202011237862.8 2020-11-09

Publications (1)

Publication Number Publication Date
WO2022095985A1 true WO2022095985A1 (zh) 2022-05-12

Family

ID=81457547

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/129177 WO2022095985A1 (zh) 2020-11-09 2021-11-08 一种智能驾驶汽车乘员舒适性评价方法和***

Country Status (1)

Country Link
WO (1) WO2022095985A1 (zh)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114996859A (zh) * 2022-07-19 2022-09-02 山东贞元汽车车轮有限公司 用于车轮钢圈智能制造的智能设备控制***
CN115406670A (zh) * 2022-08-16 2022-11-29 中国第一汽车股份有限公司 车辆性能的测试方法、装置、电子设备以及一种车辆
CN115964810A (zh) * 2023-03-16 2023-04-14 中国重汽集团济南动力有限公司 一种车辆座椅动态舒适度评价及选型方法
CN116448452A (zh) * 2023-04-13 2023-07-18 北京工业大学 一种基于zynq多传感器协同的人体振动监测***
CN117113856A (zh) * 2023-10-23 2023-11-24 中汽研（天津）汽车工程研究院有限公司 一种适用预碰撞场景的假人模型确定方法及***
CN117367788A (zh) * 2023-12-08 2024-01-09 江苏梦天机电科技有限公司 一种新能源变速箱功能测试***
CN118082636A (zh) * 2024-04-23 2024-05-28 南昌智能新能源汽车研究院 一种智能座舱内汽车座椅优化方法及***

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010058877A1 (en) * 2008-11-18 2010-05-27 Korea Railroad Research Institute Measuring system and the method of train ride comfort using bioelectrical signals
WO2015057145A1 (en) * 2013-10-16 2015-04-23 Scania Cv Ab Method and system for controlling the acceleration process of a bus
CN105628405A (zh) * 2015-12-19 2016-06-01 中车青岛四方机车车辆股份有限公司 高速列车综合舒适度试验方法及装置
CN109311478A (zh) * 2016-12-30 2019-02-05 同济大学 一种基于舒适度的自动驾驶车速控制方法
CN110843765A (zh) * 2019-11-29 2020-02-28 上海汽车集团股份有限公司 一种自动驾驶方法、装置及电子设备
CN112353393A (zh) * 2020-11-09 2021-02-12 清华大学 一种智能驾驶汽车乘员状态检测***
CN112353392A (zh) * 2020-11-09 2021-02-12 清华大学 一种智能驾驶汽车乘员舒适性评价方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010058877A1 (en) * 2008-11-18 2010-05-27 Korea Railroad Research Institute Measuring system and the method of train ride comfort using bioelectrical signals
WO2015057145A1 (en) * 2013-10-16 2015-04-23 Scania Cv Ab Method and system for controlling the acceleration process of a bus
CN105628405A (zh) * 2015-12-19 2016-06-01 中车青岛四方机车车辆股份有限公司 高速列车综合舒适度试验方法及装置
CN109311478A (zh) * 2016-12-30 2019-02-05 同济大学 一种基于舒适度的自动驾驶车速控制方法
CN110843765A (zh) * 2019-11-29 2020-02-28 上海汽车集团股份有限公司 一种自动驾驶方法、装置及电子设备
CN112353393A (zh) * 2020-11-09 2021-02-12 清华大学 一种智能驾驶汽车乘员状态检测***
CN112353392A (zh) * 2020-11-09 2021-02-12 清华大学 一种智能驾驶汽车乘员舒适性评价方法

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114996859A (zh) * 2022-07-19 2022-09-02 山东贞元汽车车轮有限公司 用于车轮钢圈智能制造的智能设备控制***
CN115406670A (zh) * 2022-08-16 2022-11-29 中国第一汽车股份有限公司 车辆性能的测试方法、装置、电子设备以及一种车辆
CN115964810A (zh) * 2023-03-16 2023-04-14 中国重汽集团济南动力有限公司 一种车辆座椅动态舒适度评价及选型方法
CN115964810B (zh) * 2023-03-16 2023-07-04 中国重汽集团济南动力有限公司 一种车辆座椅动态舒适度评价及选型方法
CN116448452A (zh) * 2023-04-13 2023-07-18 北京工业大学 一种基于zynq多传感器协同的人体振动监测***
CN116448452B (zh) * 2023-04-13 2024-05-03 北京工业大学 一种基于zynq多传感器协同的人体振动监测***
CN117113856A (zh) * 2023-10-23 2023-11-24 中汽研（天津）汽车工程研究院有限公司 一种适用预碰撞场景的假人模型确定方法及***
CN117113856B (zh) * 2023-10-23 2024-01-30 中汽研（天津）汽车工程研究院有限公司 一种适用预碰撞场景的假人模型确定方法及***
CN117367788A (zh) * 2023-12-08 2024-01-09 江苏梦天机电科技有限公司 一种新能源变速箱功能测试***
CN117367788B (zh) * 2023-12-08 2024-02-13 江苏梦天机电科技有限公司 一种新能源变速箱功能测试***
CN118082636A (zh) * 2024-04-23 2024-05-28 南昌智能新能源汽车研究院 一种智能座舱内汽车座椅优化方法及***

Similar Documents

Publication Publication Date Title
WO2022095985A1 (zh) 一种智能驾驶汽车乘员舒适性评价方法和***
CN112353392B (zh) 一种智能驾驶汽车乘员舒适性评价方法
CN112353393B (zh) 一种智能驾驶汽车乘员状态检测***
CN102138789B (zh) 一种动态心电和运动记录与分析***
CN108765876A (zh) 基于多模信号的驾驶疲劳深度分析预警***及方法
Solovey et al. Classifying driver workload using physiological and driving performance data: two field studies
CN102961126B (zh) 基于脉象诊断模式的驾驶预警方法
CN100482155C (zh) 基于脑机交互的注意力状态即时检测***及检测方法
CN106073712B (zh) 基于心生理信号的驾驶警示方向盘套装置及信号检测方法
CN102184415B (zh) 一种基于脑电信号的疲劳状态识别方法
CN105877766A (zh) 一种基于多生理信号融合的精神状态检测***及方法
CN113951903B (zh) 基于脑电数据测定的高速铁路调度员超负荷状态识别方法
CN112057087B (zh) 精神***症高风险人群自主神经功能数据处理方法及装置
CN110367975A (zh) 一种基于脑机接口的疲劳驾驶检测预警方法
CN113501005A (zh) 基于驾驶员的生理信息辅助控制车辆的方法和设备
CN114237391A (zh) 一种城市轨道交通调度虚拟训练测试***及其方法
CN114557708A (zh) 基于脑电双特征融合的体感刺激意识检测装置和方法
CN207029275U (zh) 能实时监测人体健康状况的汽车方向盘***
Zhenhai et al. The Driver’s Steering Feel Assessment Using EEG and EMG signals G
CN115331431B (zh) 信号灯倒计时条件下的行人过街心理负荷测试及评价方法
CN116884288A (zh) 抗眩晕训练平台及方法
Rahman et al. EMG signal segmentation to predict driver’s vigilance state
Fujita et al. Driver drowsiness detection using a gyroscope attached to a seatbelt
Krishnan et al. Drowsiness detection using electroencephalogram anomaly based on spectral entropy features and linear classifier
CN112656430A (zh) 基于站立位失衡诱发脑电的卒中平衡康复评估方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21888677

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21888677

Country of ref document: EP

Kind code of ref document: A1