Heavy truck automatic emergency braking method and system based on multi-sensor fusion

Abstract

The invention relates to a heavy truck automatic emergency braking method and system based on multi-sensor fusion, and the method comprises the steps: M1, enabling a vehicle to run on a road, obtaining the image data information of the road in real time based on a vehicle-mounted camera, obtaining the data information of the state parameters of an obstacle in real time based on a vehicle-mounted millimeter wave radar, data information of state parameters of a driver is obtained in real time based on a built-in DMS camera, data information of state parameters of a vehicle and data information of weather condition parameters are collected, and data of all sensors are fused through a multi-Bayesian estimation algorithm based on integrated random inertia weight factors. And data information of the fused vehicle driving decision parameters is obtained. The defect that a white light camera is low in recognition precision at night or in severe weather and even cannot recognize pedestrians and vehicles is overcome, the vehicles have the all-weather danger early warning capacity, and the heavy truck has the all-weather danger early warning recognition capacity through fusion of multiple sensors.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-sensor fusion, and in particular to a heavy truck automatic emergency braking method and system based on multi-sensor fusion. Background Art

[0002] Traditional AEB systems mainly detect obstacles ahead through white light cameras and millimeter wave radars. However, at night or in bad weather (such as heavy fog, haze, rain, snow, sand and dust and other bad weather), the white light camera will have low recognition accuracy or even be unable to recognize the conditions of pedestrians and vehicles. As a production tool, heavy trucks operate all day long, including in various bad weather conditions, which undoubtedly increases the driving risk. Summary of the Invention

[0003] In view of the above problems, the present invention provides a heavy truck automatic emergency braking method and system based on multi-sensor fusion, which not only solves the defect that the white light camera has low recognition accuracy or even cannot recognize pedestrians and vehicles at night or in bad weather (such as heavy fog, haze, rain, snow, sand and dust and other bad weather), enables the vehicle to have all-weather danger warning capabilities, but also the fusion of multi-sensors enables the heavy truck to have all-weather danger warning and recognition capabilities.

[0004] In order to achieve the above object and other related objects, the technical solutions provided by the present invention are as follows: A heavy truck automatic emergency braking method based on multi-sensor fusion, the method comprising: M1. When the vehicle is driving on the road, real-time image data information of the road is obtained based on an in-vehicle camera, real-time data information of the state parameters of obstacles is obtained based on an in-vehicle millimeter wave radar, real-time data information of the state parameters of the driver is obtained based on a built-in DMS camera, and real-time data information of the state parameters of the vehicle and data information of weather condition parameters are collected. M2. Based on the real-time image data information of the road, the real-time data information of the state parameters of obstacles, the real-time data information of the state parameters of the driver, the real-time data information of the state parameters of the vehicle, and the data information of weather condition parameters, a multi-Bayesian estimation algorithm based on an integrated random inertia weight factor is used to fuse the data of each sensor to obtain data information of the vehicle driving decision parameters after fusion. M3. Based on the data information of the vehicle driving decision parameters after fusion, a regression prediction algorithm of a BARNN model based on beluga optimization is used to predict the braking state parameters of the vehicle to obtain data information of the braking state parameters of the vehicle after prediction. M4. Based on the data information of the braking state parameters of the vehicle after prediction, a braking smoothness evaluation function R of the vehicle is constructed to evaluate the smoothness of vehicle braking and obtain data information of the evaluation value of the smoothness of vehicle braking.

[0005] Further, the method further includes: M5. Based on the data information of the evaluation value of the smoothness of the vehicle braking, set a preset threshold. If the evaluation value of the smoothness of the vehicle braking is less than the preset threshold, it does not meet the requirements, and return to step M3. If the evaluation value of the smoothness of the vehicle braking is greater than the preset threshold, it meets the requirements, and the vehicle brakes according to the braking state parameters of the vehicle.

[0006] Further, the braking smoothness evaluation function R of the vehicle is , where x is the data information of the braking state parameters of the predicted vehicle, and α1, α2, and α3 are the weight factors for evaluating the braking performance of the vehicle.

[0007] Further, the weight factors α1, α2, and α3 for evaluating the braking performance of the vehicle are , , , where x is the data information of the braking state parameters of the predicted vehicle.

[0008] Further, in step M2, the fusion of the data of each sensor by using the multi-Bayesian estimation algorithm based on the integrated random inertia weight factor includes:[[]] M21. Based on the data information of the road image, the state parameter data information of the obstacle, the state parameter data information of the driver, the state parameter data information of the vehicle, and the weather condition parameter data information, respectively construct their respective associated probability distribution functions G 道路 , G 障碍物 , G 驾驶员 , G 车辆 and G 天气 , , , , , , Among them, y1 is the image data information of the road, y2 is the data information of the state parameters of the obstacle, y3 is the data information of the state parameters of the driver, y4 is the data information of the state parameters of the vehicle, and y5 is the data information of the weather condition parameters, which characterize the joint probability distribution of the images of the obstacle, driver, vehicle, road and weather condition, and obtain the data information of the joint probability distribution of the images of the obstacle, driver, vehicle, road and weather condition; M22. Based on the data information of the joint probability distribution of the images of the obstacle, driver, vehicle, road and weather condition, establish a joint posterior probability distribution function H, , wherein, g1 is the data information of the associated probability distribution of the obstacle, g2 is the data information of the associated probability distribution of the driver, g3 is the data information of the associated probability distribution of the vehicle, g4 is the data information of the associated probability distribution of the road image, g5 is the data information of the associated probability distribution of the weather condition, and η1, η2 and η3 are integrated random inertia weight factors; M23. Based on the joint posterior probability distribution function H, fuse the data of each sensor to obtain the data information of the fused vehicle driving decision parameters.

[0009] Further, the integrated random inertia weight factors η1, η2 and η3 are , , , wherein, g1 is the data information of the associated probability distribution of the obstacle, g2 is the data information of the associated probability distribution of the driver, g3 is the data information of the associated probability distribution of the vehicle, g4 is the data information of the associated probability distribution of the road image, g5 is the data information of the associated probability distribution of the weather condition.

[0010] Further, in step M3, the regression prediction algorithm using the BARNN model based on beluga optimization to predict the braking state parameters of the vehicle includes: M31. Input the data information of the fused vehicle driving decision parameters into the BARNN regression prediction model for training and learning, and initialize the parameters of the model to obtain the data information of the initialized model parameters; M32. Based on the data information of the initialized model parameters, initialize the beluga population, determine the population parameters and the maximum number of iterations, and obtain the data information of the initialized beluga population; M33. Based on the data information of the initialized beluga population, establish a fitness function S for the population individuals, , Among them, h is the data information of the initialized beluga whale population, and the fitness values of the individuals in the population are calculated to obtain the data information of the fitness values of the individuals in the beluga whale population; M34. Based on the data information of the fitness values of the beluga whale population individuals, establish the objective function F, , Among them, q is the data information of the fitness value of the beluga whale population individuals. The parameters of the model are optimized to obtain the optimized BARNN regression prediction model.

[0011] Furthermore, the prediction of the braking state parameters of the vehicle using the regression prediction algorithm based on the BARNN model optimized by Beluga also includes: M35. Based on the optimized BARNN regression prediction model, the fused data information of the vehicle driving decision parameters is input, the braking state parameters of the vehicle are predicted, and the data information of the predicted braking state parameters of the vehicle is obtained.

[0012] In order to achieve the above-mentioned object and other related objects, the present invention also provides a system for implementing any one of the above-mentioned methods for automatic emergency braking of heavy trucks based on multi-sensor fusion, the system comprising: a front-view camera connected to the vehicle-mounted controller and used for transmitting a captured image to the vehicle-mounted controller; The front millimeter-wave radar is installed on the front grille of the vehicle and is used to process the received signal and transmit it to the on-board controller. It determines the position, speed and other characteristics of the target object by comparing the difference between the transmitted signal and the received signal; Night vision system, which receives images from night vision cameras, processes them through algorithms, and outputs the target object, distance, and speed, and transmits the output information to the on-board controller; The DMS system receives the driver status information output by the DMS camera, processes it through an algorithm, and outputs the driver status and alarm signal, which is then transmitted to the vehicle sensors and central control instrument; Central control instrument, used to receive information from the vehicle controller and display DMS system alarm and forward collision warning information; The on-board controller is installed under the vehicle's auxiliary instrument panel. It is used to receive information from various sensors and systems, and transmit the information to the wire-controlled chassis and instruments through the bus.

[0013] Further, the front millimeter-wave radar transmits the perception result signal to the vehicle-mounted controller via the CAN bus. The front-view camera and the night vision system transmit the signals to the vehicle-mounted controller in the LVDS signal format. Meanwhile, the night vision system transmits the obstacle position and distance information to the vehicle-mounted controller via the CAN bus. The vehicle-mounted controller outputs the final obstacle information after fusion processing.

[0014] The present invention has the following positive effects: 1. The present invention fuses the data of each sensor by adopting a multi-Bayesian estimation algorithm based on an integrated random inertia weight factor, and combines a regression prediction algorithm based on a Beluga optimization-based BARNN model to predict the braking state parameters of the vehicle. It not only solves the defect that the white light camera has low recognition accuracy or even cannot recognize pedestrians and vehicles at night or in bad weather (such as heavy fog, haze, rain, snow, sand and dust and other bad weather), enabling the vehicle to have all-weather danger warning capabilities, but also the fusion of multiple sensors enables the heavy truck to have all-weather danger warning and recognition capabilities.

[0015] 2. The present invention constructs a vehicle braking smoothness evaluation function R to evaluate the smoothness of vehicle braking and sets a preset threshold. If the evaluation value of the vehicle braking smoothness is less than the preset threshold, it does not meet the requirements; if the evaluation value of the vehicle braking smoothness is greater than the preset threshold, it meets the requirements. The vehicle brakes according to the vehicle braking state parameters, which can not only dynamically adjust the entire braking process of the vehicle, improve driving safety, but also reduce the driving discomfort caused by emergency braking and increase driving comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic diagram of the system framework of the present invention; Figure 2 is a schematic diagram of the method flow of the present invention; Figure 3 is a schematic diagram of the flow of the multi-Bayesian estimation algorithm based on an integrated random inertia weight factor of the present invention; Figure 4 is a schematic diagram of the regression prediction algorithm of the Beluga optimization-based BARNN model of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The following describes exemplary embodiments of the present disclosure, including various details of the embodiments of the present disclosure to facilitate understanding. It should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted below.

[0018] Embodiment 1: AsFigure 2 As shown in the figure, a heavy truck automatic emergency braking method based on multi-sensor fusion, the method comprising: M1. When the vehicle is traveling on the road, the image data information of the road is obtained in real time based on the in-vehicle camera, the data information of the state parameters of the obstacle is obtained in real time based on the in-vehicle millimeter-wave radar, the data information of the state parameters of the driver is obtained in real time based on the built-in DMS camera, and the data information of the state parameters of the vehicle and the data information of the weather condition parameters are collected; M2. Based on the image data information of the road, the data information of the state parameters of the obstacle, the data information of the state parameters of the driver, the data information of the state parameters of the vehicle, and the data information of the weather condition parameters, the multi-Bayesian estimation algorithm based on the integrated random inertia weight factor is used to fuse the data of each sensor to obtain the data information of the vehicle driving decision parameters after fusion; M3. Based on the data information of the vehicle driving decision parameters after fusion, the regression prediction algorithm of the BARNN model based on beluga optimization is used to predict the braking state parameters of the vehicle to obtain the data information of the braking state parameters of the vehicle after prediction; M4. Based on the data information of the braking state parameters of the vehicle after prediction, a braking smoothness evaluation function R of the vehicle is constructed to evaluate the smoothness of vehicle braking, and the data information of the evaluation value of the smoothness of vehicle braking is obtained.

[0019] In this embodiment, as Figure 3 shown, in step M2, the use of the multi-Bayesian estimation algorithm based on the integrated random inertia weight factor to fuse the data of each sensor includes: M21. Based on the image data information of the road, the data information of the state parameters of the obstacle, the data information of the state parameters of the driver, the data information of the state parameters of the vehicle, and the data information of the weather condition parameters, the respective associated probability distribution functions G 道路 、G 障碍物 、G 驾驶员 、G 车辆 and G 天气 are constructed respectively, , , , , , Among them, y1 is the image data information of the road, y2 is the data information of the state parameters of the obstacle, y3 is the data information of the state parameters of the driver, y4 is the data information of the state parameters of the vehicle, y5 is the data information of the weather condition parameters, and the associated probability distribution of the images of the obstacle, driver, vehicle, road and weather condition is characterized to obtain the data information of the associated probability distribution of the images of the obstacle, driver, vehicle, road and weather condition; M22. Based on the data information of the associated probability distribution of the images of the obstacle, driver, vehicle, road and weather condition, a joint posterior probability distribution function H is established, , where g1 is the data information of the associated probability distribution of the obstacle, g2 is the data information of the associated probability distribution of the driver, g3 is the data information of the associated probability distribution of the vehicle, g4 is the data information of the associated probability distribution of the road image, g5 is the data information of the associated probability distribution of the weather condition, and η1, η2 and η3 are integrated random inertia weight factors; M23. Based on the joint posterior probability distribution function H, the data of each sensor are fused to obtain the data information of the fused vehicle driving decision parameters.

[0020] In this embodiment, the integrated random inertia weight factors η1, η2 and η3 are, , , , where g1 is the data information of the associated probability distribution of the obstacle, g2 is the data information of the associated probability distribution of the driver, g3 is the data information of the associated probability distribution of the vehicle, g4 is the data information of the associated probability distribution of the road image, g5 is the data information of the associated probability distribution of the weather condition.

[0021] In this embodiment, as Figure 4 shown, in step M3, the regression prediction algorithm using the BARNN model based on beluga optimization to predict the braking state parameters of the vehicle includes: M31. Input the data information of the fused vehicle driving decision parameters into the BARNN regression prediction model for training and learning, and initialize the parameters of the model to obtain the data information of the initialized model parameters; M32. Based on the data information of the initialized model parameters, initialize the beluga population, determine the population parameters and the maximum number of iterations to obtain the data information of the initialized beluga population; M33. Based on the data information of the initialized beluga whale population, establish the fitness function S of the population individuals, , where h is the data information of the initialized beluga whale population, and the fitness values of the population individuals are deduced to obtain the data information of the fitness values of the beluga whale population individuals; M34. Based on the data information of the fitness values of the beluga whale population individuals, establish the objective function F, , where q is the data information of the fitness values of the beluga whale population individuals, and the parameters of the model are optimized to obtain the optimized BARNN regression prediction model.

[0022] In this embodiment, the regression prediction algorithm using the BARNN model optimized by beluga whale optimization to predict the braking state parameters of the vehicle further includes: M35. Based on the optimized BARNN regression prediction model, input the data information of the fused vehicle driving decision parameters, and predict the braking state parameters of the vehicle to obtain the data information of the predicted braking state parameters of the vehicle.

[0023] Embodiment 2: On the basis of a heavy truck automatic emergency braking method based on multi-sensor fusion in Embodiment 1, the present invention will be further described and explained below.

[0024] As Figure 2 shown, a heavy truck automatic emergency braking method based on multi-sensor fusion, the method includes: M1. When the vehicle is driving on the road, based on the on-vehicle camera, the image data information of the road is obtained in real time, based on the on-vehicle millimeter-wave radar, the state parameter data information of the obstacle is obtained in real time, based on the built-in DMS camera, the state parameter data information of the driver is obtained in real time, and the state parameter data information of the vehicle and the weather condition parameter data information are collected; M2. Based on the image data information of the road, the state parameter data information of the obstacle, the state parameter data information of the driver, the state parameter data information of the vehicle, and the weather condition parameter data information, use the multi-Bayesian estimation algorithm based on the integrated random inertia weight factor to fuse the data of each sensor to obtain the data information of the fused vehicle driving decision parameters; M3. Based on the data information of the fused vehicle driving decision parameters, use the regression prediction algorithm of the BARNN model optimized by beluga whale optimization to predict the braking state parameters of the vehicle to obtain the data information of the predicted braking state parameters of the vehicle; M4. Based on the data information of the predicted braking state parameters of the vehicle, construct an evaluation function R for the braking smoothness of the vehicle to evaluate the braking smoothness of the vehicle and obtain the data information of the evaluation value of the braking smoothness of the vehicle.

[0025] In this embodiment, the method further includes: M5. Based on the data information of the evaluation value of the braking smoothness of the vehicle, set a preset threshold. If the evaluation value of the braking smoothness of the vehicle is less than the preset threshold, it does not meet the requirements, and return to step M3. If the evaluation value of the braking smoothness of the vehicle is greater than the preset threshold, it meets the requirements, and the vehicle brakes according to the braking state parameters of the vehicle.

[0026] In this embodiment, the evaluation function R for the braking smoothness of the vehicle is , where x is the data information of the predicted braking state parameters of the vehicle, and α1, α2, and α3 are the weight factors for evaluating the braking performance of the vehicle.

[0027] In this embodiment, the weight factors α1, α2, and α3 for evaluating the braking performance of the vehicle are , , , where x is the data information of the predicted braking state parameters of the vehicle.

[0028] In this embodiment, as Figure 1 shown, a heavy truck automatic emergency braking system based on multi-sensor fusion mainly includes a front-view camera, a front millimeter-wave radar, a night vision system, a DMS system, a vehicle-mounted controller, and a central control instrument.

[0029] The front-view camera is connected to the vehicle-mounted controller and mainly transmits the captured image to the vehicle-mounted controller; The front millimeter-wave radar is installed at the front grille of the vehicle and mainly processes the received signal and transmits it to the vehicle-mounted controller, and determines the position, speed, and other characteristics of the target object by comparing the differences between the transmitted signal and the received signal; The night vision system receives the image output from the night vision camera, processes it through an algorithm, and outputs the target object, the distance and speed of the target object, and transmits the output information to the vehicle-mounted controller; The DMS system receives the driver state information output from the DMS camera, processes it through an algorithm, and outputs the driver state and an alarm signal, and transmits it to the vehicle-mounted sensor and the central control instrument; The central control instrument mainly receives the information of the vehicle-mounted controller and realizes the display of the DMS system alarm and the forward collision warning information; The vehicle-mounted controller is installed under the vehicle's passenger instrument panel. It mainly receives information from various sensors and systems and transmits the information to the by-wire chassis and instrument panel via a bus.

[0030] In this embodiment, the front-view camera is connected to the vehicle-mounted controller via an LVDS video harness. The camera is a wide-angle camera with an angle of 100°. In this embodiment, the front millimeter-wave radar is connected to the vehicle-mounted controller via a CAN bus and outputs a perception result signal to the vehicle-mounted controller. The model of the millimeter-wave radar is Continental ARS408-21. In this embodiment, the night vision system mainly includes a night vision camera and a night vision processor. The night vision camera is installed in the vehicle's front grille and outputs a night vision image to the night vision processor. In this embodiment, the night vision processor is mainly responsible for processing the night vision image and outputs the target object's distance, speed, and image to the vehicle-mounted controller after algorithm processing. In this embodiment, the DMS system mainly includes a DMS camera and a DMS processor. The DMS camera is installed above the left A-pillar of the vehicle and outputs a driver image to the DMS processor. In this embodiment, the DMS processor is mainly responsible for processing the driver image, outputs the driver's status information after deep learning algorithm processing, and transmits the information to the vehicle-mounted controller and the central control instrument. In this embodiment, the vehicle-mounted controller is mainly connected to the front millimeter-wave radar, the front-view camera, the night vision system, the DMS system, the central control instrument, and the by-wire chassis.

[0031] In this embodiment, the front millimeter-wave radar transmits the perception result signal to the vehicle-mounted controller via a CAN bus. The front-view camera and the night vision system transmit signals to the vehicle-mounted controller in the LVDS signal format. At the same time, the night vision system transmits obstacle position, distance information, etc. to the vehicle-mounted controller via a CAN bus. The vehicle-mounted controller outputs the final obstacle information after fusion processing. The DMS system transmits the driver's status information to the vehicle-mounted controller via a CAN bus. The vehicle-mounted controller decides whether to extend the TTC by judging the driver's status. The central control instrument is connected to the vehicle-mounted controller via a CAN bus and mainly displays the collision alarm signal output by the vehicle-mounted controller. The by-wire chassis is connected to the vehicle-mounted controller via a CAN bus and mainly receives the braking signal from the vehicle-mounted controller.

[0032] In this embodiment, after the system is started, the vehicle-mounted controller will determine whether it is night or bad weather based on the vehicle light sensor signal and the signal provided by the white light camera. Thus, the vehicle-mounted controller will select the image signal with high confidence for signal fusion with the millimeter wave radar; after algorithm processing, information such as the position and speed of the final target obstacle will be output; On the other hand, the vehicle-mounted controller simultaneously receives the driver status signal output from the DMS system. When an obstacle in front of the vehicle triggers a forward alarm and the vehicle-mounted controller simultaneously detects that the driver is fatigued and has not taken any measures, the vehicle-mounted controller will increase the TTC and at the same time take the brakes in advance to ensure driving safety and reduce the discomfort caused by emergency braking.

[0033] In summary, the present invention not only solves the defect that the white light camera has low recognition accuracy or even cannot recognize pedestrians and vehicles at night or in bad weather (such as heavy fog, haze, rain, snow, sand and dust, etc.), enabling the vehicle to have all-weather danger warning capabilities, but also the fusion of multiple sensors enables the heavy truck to have all-weather danger warning and recognition capabilities.

[0034] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A heavy truck automatic emergency braking method based on multi-sensor fusion, characterized in that, The method includes: M1. When the vehicle is driving on the road, the image data information of the road is obtained in real time based on the in-vehicle camera, the data information of the state parameters of the obstacle is obtained in real time based on the in-vehicle millimeter-wave radar, the data information of the state parameters of the driver is obtained in real time based on the built-in DMS camera, and the data information of the state parameters of the vehicle and the data information of the weather condition parameters are collected. M2. Based on the image data information of the road, the data information of the state parameters of the obstacle, the data information of the state parameters of the driver, the data information of the state parameters of the vehicle, and the data information of the weather condition parameters, the data of each sensor is fused by using the multi-Bayesian estimation algorithm based on the integrated random inertia weight factor to obtain the data information of the vehicle driving decision parameters after fusion. M3. Based on the data information of the vehicle driving decision parameters after fusion, the regression prediction algorithm of the BARNN model based on beluga optimization is used to predict the braking state parameters of the vehicle to obtain the data information of the braking state parameters of the vehicle after prediction. M4. Based on the data information of the braking state parameters of the vehicle after prediction, a braking smoothness evaluation function R of the vehicle is constructed to evaluate the smoothness of the vehicle braking and obtain the data information of the evaluation value of the smoothness of the vehicle braking.

2. The method for automatically emergency braking of heavy trucks based on multi-sensor fusion according to claim 1, characterized in that The method further includes: M5. Based on the data information of the evaluation value of the smoothness of the vehicle braking, a preset threshold is set. If the evaluation value of the smoothness of the vehicle braking is less than the preset threshold, the requirement is not met, and step M3 is returned. If the evaluation value of the smoothness of the vehicle braking is greater than the preset threshold, the requirement is met, and the vehicle brakes according to the braking state parameters of the vehicle.

3. The heavy truck automatic emergency braking method based on multi-sensor fusion according to claim 1, wherein: The braking smoothness evaluation function R of the vehicle is ， where x is the data information of the braking state parameters of the vehicle after prediction, and α1, α2, and α3 are the weight factors for evaluating the braking performance of the vehicle.

4. The method for heavy truck automatic emergency braking based on multi-sensor fusion according to claim 3, characterized in that: The weight factors α1, α2, and α3 for evaluating the braking performance of the vehicle are ， ， ， where x is the data information of the braking state parameters of the vehicle after prediction.

5. The heavy truck automatic emergency braking method based on multi-sensor fusion according to claim 1, characterized in that, In step M2, the fusion of the data of each sensor by using the multi-Bayesian estimation algorithm based on the integrated random inertia weight factor includes: M21. Based on the data information of the road image data, the data information of the obstacle state parameters, the data information of the driver state parameters, the data information of the vehicle state parameters, and the data information of the weather condition parameters, respectively construct their respective associated probability distribution functions G 道路 、G 障碍物 、G 驾驶员 、G 车辆 and G 天气 , ， ， ， ， ， where y1 is the image data information of the road, y2 is the data information of the state parameters of the obstacle, y3 is the data information of the state parameters of the driver, y4 is the data information of the state parameters of the vehicle, y5 is the data information of the weather condition parameters, and the associated probability distribution of the obstacle, the driver, the vehicle, the image of the road, and the weather condition is characterized to obtain the data information of the associated probability distribution of the obstacle, the driver, the vehicle, the image of the road, and the weather condition. M22. Based on the data information of the associated probability distribution of the obstacle, the driver, the vehicle, the image of the road, and the weather condition, a joint posterior probability distribution function H is established. ， Among them, g1 is the data information of the associated probability distribution of obstacles, g2 is the data information of the associated probability distribution of the driver, g3 is the data information of the associated probability distribution of the vehicle, g4 is the data information of the associated probability distribution of the road image, g5 is the data information of the associated probability distribution of the weather condition, and η1, η2, and η3 are integrated random inertia weight factors; M23. Based on the joint posterior probability distribution function H, fuse the data of each sensor to obtain the data information of the vehicle driving decision parameters after fusion.

6. The method for automatically emergency braking of heavy trucks based on multi-sensor fusion according to claim 5, characterized in that: The integrated random inertia weight factors η1, η2, and η3 are ， ， ， Among them, g1 is the data information of the associated probability distribution of obstacles, g2 is the data information of the associated probability distribution of the driver, g3 is the data information of the associated probability distribution of the vehicle, g4 is the data information of the associated probability distribution of the road image, g5 is the data information of the associated probability distribution of the weather condition.

7. The method for automatically emergency braking of heavy trucks based on multi-sensor fusion according to claim 1, characterized in that, In step M3, the regression prediction algorithm using the BARNN model based on beluga optimization to predict the braking state parameters of the vehicle includes: M31. Input the data information of the vehicle driving decision parameters after fusion into the BARNN regression prediction model for training and learning, initialize the parameters of the model, and obtain the data information of the initialized model parameters; M32. Based on the data information of the initialized model parameters, initialize the beluga population, determine the population parameters and the maximum number of iterations, and obtain the data information of the initialized beluga population; M33. Based on the data information of the initialized beluga population, establish the fitness function S of the population individuals ， Among them, h is the data information of the initialized beluga population, calculate the fitness values of the population individuals, and obtain the data information of the fitness values of the beluga population individuals; M34. Based on the data information of the fitness values of the beluga population individuals, establish the objective function F ， Among them, q is the data information of the fitness values of the beluga population individuals, optimize the parameters of the model, and obtain the optimized BARNN regression prediction model.

8. The method for automatically emergency braking of heavy trucks based on multi-sensor fusion according to claim 7, characterized in that, The regression prediction algorithm using the BARNN model based on beluga optimization to predict the braking state parameters of the vehicle further includes: M35. Based on the optimized BARNN regression prediction model, input the data information of the vehicle driving decision parameters after fusion, predict the braking state parameters of the vehicle, and obtain the data information of the predicted braking state parameters of the vehicle.

9. A system for implementing the heavy truck automatic emergency braking method based on multi-sensor fusion according to any one of claims 1-8, characterized in that, The system includes: A front-view camera, connected to the vehicle-mounted controller, for transmitting the captured image to the vehicle-mounted controller; A front millimeter-wave radar, installed at the front grille of the vehicle, for processing the received signal and transmitting it to the vehicle-mounted controller, and determining the position, speed, and other characteristics of the target object by comparing the differences between the transmitted signal and the received signal; A night vision system, receiving the image output from the night vision camera, processing it through an algorithm, outputting the target object and the distance and speed of the target object, and transmitting the output information to the vehicle-mounted controller; The DMS system receives the driver status information output by the DMS camera, processes it through algorithms, and then outputs the driver status and alarm signals, which are transmitted to in-vehicle sensors and the central control instrument panel. The central control instrument panel is used to receive the in-vehicle controller information and display the DMS system alarm and forward collision warning information. The in-vehicle controller is installed under the vehicle's passenger instrument panel and is used to receive the information from various sensors and systems, and transmit the information to the drive-by-wire chassis and instrument panel through the bus.

10. The system according to claim 9, wherein: The front millimeter-wave radar transmits the sensed result signal to the in-vehicle controller through the CAN bus. The front-view camera and the night vision system transmit the signals to the in-vehicle controller in the LVDS signal format. At the same time, the night vision system transmits the obstacle position and distance information to the in-vehicle controller through the CAN bus. The in-vehicle controller outputs the final obstacle information after fusion processing.

Images