Driving torque safety auxiliary control system based on multi-sensor data fusion

Abstract

The invention relates to the technical field of driving torque control of vehicles, in particular to a driving torque safety auxiliary control system based on multi-sensor data fusion. Comprising a sensor module, a multi-sensor data fusion and processing module and a self-adaptive driving torque revision module, the sensor module is used for detecting the speed, the position, the motion information, the road surface information, the surrounding environment and the like of the vehicle. The invention provides a driving torque safety auxiliary control system based on multi-sensor data fusion, and aims to optimize driving torque control of an electric vehicle and improve dynamic response precision, safety and stability of the vehicle under various complex road conditions. By integrating data from a plurality of sensors in real time, the driving state, the external environment and the vehicle position information of the vehicle are accurately sensed, and the driving torque output of the vehicle is dynamically and auxiliarily adjusted in combination with the information, so that the optimal power performance and control stability of the vehicle under different working conditions can be ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle driving torque control, and in particular to a driving torque safety auxiliary control system with multi-sensor data fusion. Background Art

[0002] With the advancement of electric vehicle technology and the continuous improvement of people's living standards, people have an increasing demand for vehicle power. In this context, distributed drive electric vehicles have become the choice of more users. However, distributed drive electric vehicles with strong power output may introduce various new safety and comfort issues when driving on ice, snow, mud, slippery, dark ice, climbing roads, or encountering obstacles, speed bumps, water, potholes, etc. on the road.

[0003] Drive torque assist control technology plays a vital role in improving vehicle power performance, economy, safety and handling stability. Traditional electric vehicle drive torque control methods mainly rely on basic information such as driver input (such as the opening of the accelerator pedal and brake pedal) and the current state of the vehicle (such as vehicle speed and gear status). Although these information inputs can meet the basic power response requirements, they often cannot effectively adapt to the rapid changes in the external environment, road adhesion and vehicle dynamic state in complex driving environments, thereby limiting the performance of the vehicle under complex working conditions and failing to effectively guarantee the safety and power performance of the vehicle.

[0004] Most of the existing research on driving torque control strategies is based on a single factor as the control input, for example, only relying on the accelerator pedal opening and change rate to dynamically adjust the torque output. This type of control method may result in the vehicle being unable to take into account both the optimal power response and comfort stability control under complex working conditions such as the driver's failure to comprehensively consider the external environment (such as the real-time road conditions), the real-time vehicle status (such as vehicle speed, speed, etc.) and the dynamic changes in road adhesion.

[0005] On roads with frequently changing road adhesion coefficients, such as mountainous roads with changeable weather, it is possible to encounter dark ice sections, slippery sections, ice and snow sections, muddy sections, as well as water accumulation, potholes, obstacles, speed bumps, uphill and downhill road conditions. Under such complex working conditions, how to reasonably coordinate and control the driving torque of distributed drive electric vehicles is still a huge challenge. Commonly used vehicle longitudinal speed estimation methods include two methods based on kinematics and dynamics. The kinematics-based method relies on wheel speed sensor data for calculation and feedback control. Although the calculation amount is small, it is easily affected by sensor errors and environmental factors. The dynamics-based method estimates the vehicle speed by establishing an accurate vehicle model to control it, and uses state estimation and other technologies to improve the recognition accuracy. However, this method is highly dependent on the model, especially under complex working conditions, it may not be able to provide reliable data as control input. This type of method that passively collects information from sensors when driving on the road surface and judges the current vehicle state is difficult for the vehicle to take control measures in advance, which is not conducive to ensuring driving safety and improving vehicle comfort.

[0006] At present, vehicle status information and external environment information are obtained through single or multiple sensors. The output of a single sensor is easily affected by noise, error and environmental changes, and it is difficult to serve as an accurate and reliable basis for vehicle control. The current multi-sensor fusion mainly synchronizes and fuses the data from each sensor, eliminates the error and noise of a single sensor, improves the accuracy and reliability of the data, and obtains more accurate vehicle position information and positioning information, but this requires users to have a deep understanding of sensor characteristics, data processing algorithms, environmental factors and the real-time performance of the system.

[0007] The existing drive torque control method has improved the acceleration performance and handling stability of the vehicle to a certain extent, but when facing complex road conditions, low adhesion coefficients and diverse driving behaviors, there may still be problems such as insufficient response, insufficient accuracy, and failure to consider safety and comfort. Therefore, how to make full use of multi-sensor fusion technology to accurately perceive the vehicle status and external environment, and take control measures in advance based on these identification information to achieve coordination and optimization of the drive torque so that it meets the driving requirements under different road conditions has become a technical challenge that needs to be solved in the field of electric vehicles.

[0008] In summary, although the driving torque control technology of distributed drive electric vehicles has made great progress in many aspects in recent years, there are still problems in how to ensure that the vehicle can safely and comfortably adapt to different road conditions and further improve the output torque control accuracy and adaptability.

[0009] To this end, a driving torque safety auxiliary control system with multi-sensor data fusion is designed to provide another technical solution to the above technical problems. Summary of the invention

[0010] Based on this, it is necessary to provide a multi-sensor data fusion driving torque safety auxiliary control system to solve the technical problems in the above-mentioned background technology.

[0011] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0012] A multi-sensor data fusion driving torque safety auxiliary control system, comprising a sensor module, a multi-sensor data fusion and processing module and an adaptive driving torque revision module;

[0013] The sensor module is used to detect the speed, position, motion information, road surface information and surrounding environment of the vehicle;

[0014] The multi-sensor data fusion and processing module is used to process the data provided by the sensor module and combine the role of each sensor in the sensor module in different scenarios;

[0015] The adaptive driving torque revision module is used to adjust the driving torque output value in real time according to the vehicle status, external environment, and positioning information provided by the multi-sensor data fusion and processing module.

[0016] As a preferred implementation of the multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention, the sensor module includes a wheel speed sensor, an inertial measurement unit, a GPS module, a visual sensor and a laser radar;

[0017] The wheel speed sensor is used to monitor the rotation speed of each wheel in real time and provide the actual driving speed and power distribution information of the vehicle;

[0018] The inertial measurement unit is used to monitor the vehicle's acceleration, tilt angle, yaw angle and other dynamic data through an accelerometer and a gyroscope, provide the control system with vehicle motion status information and assist in positioning;

[0019] The GPS module is used to provide geographic information such as the real-time position, speed and heading angle of the vehicle, and determine the precise position and driving direction of the vehicle on the road;

[0020] The visual sensor is used to collect the road conditions in front of the vehicle in real time, identify complex road conditions, and assist in control decision making;

[0021] The laser radar is used to provide the vehicle with high-precision three-dimensional point cloud data of the surrounding environment, and cooperates with the visual sensor to accurately perceive obstacles, ramps, potholes and road conditions, and judge the driving environment.

[0022] As a preferred implementation of the multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention, the multi-sensor data fusion and processing module processes the information provided by the sensor module in the following steps:

[0023] A. The GPS module provides global positioning information of the vehicle, the inertial measurement unit provides acceleration and angular velocity information, the visual sensor provides visual image information, the laser radar provides accurate three-dimensional environmental point cloud data, and the wheel speed sensor measures the vehicle's motion state;

[0024] B. Synchronize the data provided in step A in time and perform preprocessing;

[0025] C. Extract features from the data obtained in step B;

[0026] D. Fusion of data from multiple sensors based on the extracted features;

[0027] E. Build a deep learning model and use the data obtained in step D as training data for the deep learning model;

[0028] F. Based on the scene information, the deep learning model obtained in step E dynamically adjusts the weight of the sensor.

[0029] As a preferred implementation of the multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention, in step C, feature extraction is performed on the data obtained in step B, and the steps are as follows:

[0030] Extract environmental features from the images of the visual sensor through a convolutional neural network, including at least road signs, pedestrians, other vehicles and road conditions;

[0031] Extract three-dimensional environment information from LiDAR through point cloud processing network;

[0032] Capturing motion changes in time series data from wheel speed sensors and inertial measurement units using long short-term memory networks.

[0033] As a preferred implementation of the multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention, the adaptive driving torque revision module is used to adjust the driving torque output value in real time according to the vehicle state, external environment, and positioning information provided by the multi-sensor data fusion and processing module, and the steps are as follows:

[0034] 1) According to different driving scenarios, construct a power compensation algorithm for climbing and headwind conditions, a power adaptation algorithm for low-adhesion road conditions, and a power buffer algorithm for potholes and obstacles on the road;

[0035] 2) By capturing the algorithm in 1), the torque output is flexibly adjusted according to vehicle status, road conditions, slope changes, and positioning information factors.

[0036] As a preferred implementation of the multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention, the power compensation algorithm under the climbing and headwind conditions is constructed in the following steps:

[0037] Analyze whether the current road has a slope through accurate vehicle positioning information, and use laser radar and cameras to detect in real time whether the road ahead actually has a slope;

[0038] Analyze the slope of the road ahead and compensate part of the driving torque in advance according to the slope;

[0039] The acceleration information of the current vehicle is collected through the IMU and defined as a 1 ;

[0040] According to the vehicle status information, the expected acceleration is calculated, which is defined as a 2 ;

[0041] The impact factor Q of the calculation environment is expressed as:

[0042] Q=(a 2 -a 1 ) / a 2

[0043] Determine whether to adjust and increase the output torque of the drive system according to the Q value;

[0044] If the Q value is positive, the output torque of the drive system is increased;

[0045] If the Q value is negative, the output torque of the drive system is reduced.

[0046] As a preferred implementation of the multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention, the power adaptation algorithm under low-adhesion road conditions has the following steps:

[0047] The sensor module monitors the road adhesion coefficient and vehicle acceleration information in real time, dynamically sets the slip rate threshold according to different road conditions, and optimizes the torque distribution between the two axles;

[0048] When a single-axle slip state is detected, the torque transfer speed between the two axes is adjusted according to the degree to which the slip ratio exceeds the threshold. The higher the slip ratio, the faster the transfer speed, by reducing the torque of the slipping axle and transferring it to the non-slipping axle.

[0049] As a preferred implementation of the multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention, the pothole and obstacle road dynamic buffering algorithm has the following steps:

[0050] The perception module uses multi-sensor fusion technology, combined with lidar and cameras, to identify road features through deep learning algorithms, and uses a time-series-based LSTM neural network to predict dynamically changing road features;

[0051] The decision module uses the target detection algorithm to classify obstacle types in real time, and models the geometric features of speed bumps and potholes, and combines the suspension dynamics model to predict vibration and impact intensity;

[0052] The control module adjusts the torque according to the output of the decision module;

[0053] The torque adjustment adopts a combination of feedforward control and feedback control, and the steps are as follows:

[0054] Feedforward control presets a reasonable torque output range based on road condition characteristics identified in advance;

[0055] Feedback control dynamically corrects the torque by real-time monitoring of vehicle posture sensor data to ensure smooth passage of the vehicle.

[0056] As a preferred implementation of the multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention, it also includes a safety warning visualization module;

[0057] The safety warning visualization module includes a visualization module and a safety warning module;

[0058] The visualization module is used to display the current road status and surrounding road environment information through the vehicle screen, and mark them on the display screen. The vehicle screen will also prompt that the torque is limited and the limited value, so that the driver can observe the surrounding environment and ensure safe driving;

[0059] The safety warning module is used to provide real-time suggestions and warnings through language and screen flashing operations on the vehicle screen when it detects that the road conditions are about to change or a road obstacle is about to be encountered, and the driver has no intention of slowing down the vehicle, and asks and informs the driver whether he needs to start the system to ensure driving safety and comfort.

[0060] It can be seen without a doubt that the above-mentioned technical solution of the present application can definitely solve the technical problem to be solved by the present application.

[0061] At the same time, through the above technical solutions, the present invention has at least the following beneficial effects:

[0062] 1. A multi-sensor data fusion driving torque safety auxiliary control system provided by the present invention is intended to optimize the driving torque control of electric vehicles and improve the power response accuracy, safety and stability of the vehicle under various complex road conditions; by real-time integration of data from multiple sensors (including but not limited to wheel speed sensors, inertial measurement units IMU, global positioning systems GPS, visual sensors, lidar), the vehicle's driving status, external environment, and vehicle position information are accurately sensed, and the vehicle's driving torque output is dynamically assisted by these information to ensure that the vehicle can achieve optimal power performance and handling stability under different working conditions.

[0063] 2. The present invention improves the multi-sensor data fusion algorithm and introduces a deep learning model to fully mine the vehicle dynamic information and external environmental status, optimize the drive torque control strategy to enhance the vehicle's power and safety, and thereby improve the accuracy and stability of drive control.

[0064] 3. The present invention can comprehensively perceive changes in vehicles and the environment, reduce the error impact caused by a single sensor, and improve the positioning accuracy of vehicles. The fusion positioning algorithm effectively combines the data of different sensors, uses deep learning and environmental scene data screening technology, and provides vehicles with more comprehensive and accurate vehicle status and location information.

[0065] 4. The present invention uses visual sensors and laser radars, combined with advanced algorithms, to enable the system to analyze and identify road conditions (such as slippery, waterlogged, snowy, muddy, etc.) and sudden obstacles (such as potholes, obstacles, etc.) in real time. The system can not only sense the current road conditions, but also predict the changing trend of road adhesion, provide prediction information for the torque control strategy, and make countermeasures in advance.

[0066] 5. The present invention can achieve the best acceleration and handling stability of the vehicle under various complex road conditions by intelligently adjusting the driving torque; combined with the real-time data feedback of the sensor, the system can quickly respond to emergencies (such as slippery roads, ice and snow cover, etc.), and timely adjust the driving torque output to ensure the stability and safety of the vehicle; the system has strong adaptability and can adjust the control strategy according to different driving conditions to optimize the driving performance of the vehicle. Under low adhesion coefficient or special driving conditions, the system can also dynamically adjust the torque output to maintain the stability of the vehicle and avoid unsafe phenomena such as slipping and loss of control. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0068] Figure 1 It is a schematic diagram of driving torque safety auxiliary control based on multi-sensor data fusion of the present invention;

[0069] Figure 2 Provide an environmental monitoring flow chart for the camera and lidar data fusion of the present invention;

[0070] Figure 3 This is a flow chart of data fusion and positioning of each sensor of the present invention;

[0071] Figure 4 It is a flow chart of driving torque auxiliary control and safety warning in each mode of the present invention;

[0072] Figure 5 It is a system structure diagram of the present invention. DETAILED DESCRIPTION

[0073] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0074] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings.

[0075] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features and technical solutions in the embodiments may be combined with each other.

[0076] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0077] Reference Figure 1-Figure 5 , a multi-sensor data fusion driving torque safety auxiliary control system, including the following:

[0078] 1. Sensor module

[0079] Wheel speed sensor: monitors the rotation speed of each wheel in real time, and provides the vehicle's actual driving speed and power distribution information.

[0080] Inertial Measurement Unit (IMU): Through accelerometers and gyroscopes, it monitors the vehicle's acceleration, tilt angle, yaw angle and other dynamic data, provides the control system with vehicle motion status information and assists in positioning.

[0081] GPS module: provides geographic information such as the vehicle's real-time position, speed, and heading angle to determine the vehicle's precise position on the road and driving direction.

[0082] Visual sensor (camera): collects the road conditions in front of the vehicle in real time, identifies complex road conditions (such as water, wetness, ice and snow, potholes, slopes, speed bumps, obstacles, etc.), and assists in control decision-making.

[0083] Laser Radar (L iDAR): Provides the vehicle with high-precision three-dimensional point cloud data of the surrounding environment, and cooperates with the camera to accurately perceive obstacles, ramps, potholes and other road conditions, and effectively judge the driving environment.

[0084] 2. Multi-sensor data fusion and processing module

[0085] Data fusion positioning algorithm: The positioning data fusion algorithm of the present invention uses deep learning to improve the accuracy of vehicle positioning, combining the characteristics of each sensor with its unique advantages and limitations in different scenarios. For example, GPS can provide global positioning information, but signal loss may occur in urban canyons or tunnels; IMU can provide high-precision motion estimation in a short period of time, but errors will accumulate over time; the camera effect depends on the ambient lighting; LiDAR can help build an environmental model through high-precision three-dimensional scanning, but it is easily affected by rain and snow; wheel speed sensors help to accurately measure the vehicle's motion state, especially when the GPS signal is lost, it can still provide key wheel speed information, but there may be data errors when the wheels slip.

[0086] This module first synchronizes and preprocesses the environment and positioning information collected by multiple sensors of the vehicle (GPS, IMU, camera, lidar, wheel speed sensor).

[0087] GPS provides global positioning information of the vehicle, IMU provides acceleration and angular velocity information, camera provides visual image information, lidar provides accurate three-dimensional environmental point cloud data, and wheel speed sensor measures the vehicle's motion state.

[0088] The data of these sensors need to be synchronized in time to ensure that their measured data at the same time are consistent. At the same time, the sensor data is preprocessed, including denoising, filtering and format conversion, for subsequent processing and input of deep learning models.

[0089] The driving environment of vehicles varies greatly. Different scenarios such as urban canyons, tunnels, or rainy and snowy weather will affect the performance of various sensors. Based on the current environment and scenario, the system needs to determine which sensor data is the most reliable and which can be ignored or reduced in weight. For example, GPS signals may be lost in urban canyons or tunnels. At this time, the role of IMU and wheel speed sensors will be more important. In tunnels, IMU and wheel speed sensors can provide more accurate short-term positioning, while the effects of vision and lidar can be used as auxiliary positioning. Therefore, the system needs to dynamically adjust the weights of sensors based on scene information (identified through deep learning models) and select the most appropriate sensor data for further processing.

[0090] After sensor data preprocessing, the data needs to be feature extracted. For visual data, a convolutional neural network (CNN) is used to extract environmental features from camera images, such as road signs, pedestrians, other vehicles, road conditions, etc. For lidar data, a dedicated point cloud processing network is used to extract three-dimensional environmental information to help build a more accurate environmental model. For IMU and wheel speed sensor data, a long short-term memory network (LSTM) is used to capture motion changes in time series data, thereby helping to infer motion status in a short period of time. All of these feature extraction processes will provide a basis for subsequent data fusion and positioning optimization.

[0091] Based on feature extraction, the data from multiple sensors are fused. Through the multimodal deep learning model, the information from different sensors is effectively integrated to ensure that the advantages of each sensor can be fully utilized under different environmental conditions. Convolutional neural networks (CNN) are used to process image and point cloud data, while LSTM networks process time series data. Different types of networks can capture different levels of information. At this stage, the system will dynamically adjust the weights of each sensor according to the scene conditions and real-time feedback data to achieve accurate positioning optimization.

[0092] The trained deep learning model will be deployed to the vehicle's real-time operation system. During the driving process of the vehicle, the system will continuously obtain data from various sensors and judge the current scene, and input it into the deep learning model for reasoning, and correct the weights of each sensor and the position of the vehicle in real time to ensure that the vehicle can be accurately positioned according to the current environmental conditions as a source of positioning information for subsequent regional torque limit control.

[0093] As the vehicle continues to drive, the system will continuously collect new sensor data and positioning results to generate new training data sets, and the system will subsequently optimize the deep learning model based on these data. As more driving scenarios accumulate, the robustness of the model will gradually increase, and it will be able to adapt to more complex environments, such as high-density urban blocks, tunnels, mountain roads, etc. In addition, the system can also use long-term accumulated data to continuously update the sensor fusion strategy to further improve positioning accuracy and stability. The data fusion positioning module provides a reliable source of positioning information for the regional torque control of the subsequent control module.

[0094] Environmental perception and prediction algorithms: Before data fusion, the data from LiDAR and cameras need to be preprocessed to ensure their effectiveness and accuracy in subsequent processing. LiDAR collects dense three-dimensional point cloud data, which often contains noise due to laser reflection and environmental factors. In order to eliminate these noises, common algorithms such as voxel grid filtering and statistical outlier removal are used to improve the quality of data by reducing the density of point clouds and removing irrelevant points. In addition, LiDAR and cameras need to be calibrated with external parameters in order to obtain the spatial relationship between the two, that is, the rotation matrix and displacement vector of their relative positions. Through this process, the three-dimensional coordinate system of LiDAR can be aligned with the two-dimensional image coordinate system of the camera. This is a crucial step in data fusion, because only after spatial alignment can the data of the two be effectively combined.

[0095] Once the preprocessing is completed, the next step is the feature extraction stage. LiDAR mainly provides depth information about the environment, which enables it to accurately perceive the distance, shape and structure of objects. Through the analysis of point cloud data, Euclidean clustering is used to identify objects in the scene, such as vehicles, buildings, road signs, etc. And the random sampling consistency algorithm is used to detect planes, such as the surface of the road, walls, etc., so that it can effectively distinguish between the ground and other non-ground objects. In contrast, the camera extracts visual features through image processing technology. Deep learning algorithms, such as convolutional neural networks (CNN), are used for object detection and recognition. Target detection algorithms (YOLO, Faster R-CNN) are used to identify pedestrians, vehicles, traffic signs, obstacles, slopes, etc. on the road. In addition, semantic segmentation is used to achieve pixel-level classification to identify detailed information such as road markings, road cracks, vegetation, etc.

[0096] After the LiDAR and camera have extracted useful features respectively, the next step is to align and fuse their data. Since the LiDAR provides 3D point cloud data and the camera provides 2D image data, the 3D point cloud of the LiDAR must be projected onto the 2D image plane of the camera through projection transformation. This process relies on the rotation matrix and displacement vector obtained in the external parameter calibration phase. Through these transformation matrices, the point cloud data of the LiDAR can be paired with the image data of the camera, so that the point cloud and image in the same scene can be aligned in the same coordinate system.

[0097] Since the sampling frequencies and timestamps of the lidar and camera may be different, the data from the two also need to be synchronized. The interpolation technology is used to match the data from the two so that the environmental information at the same time can be combined and used. After completing these steps, the system can obtain a fused data set that contains the depth information of the lidar and the visual information of the camera, with higher environmental perception capabilities.

[0098] The next step after data fusion is to understand the environmental information. It is particularly important to accurately perceive surrounding obstacles, traffic signs, road conditions, etc. The depth information provided by LiDAR can help the system accurately model surrounding obstacles and evaluate the distance and shape of objects. The camera plays an important role in object classification and road feature detection.

[0099] Through image recognition, the camera can detect traffic signs, traffic lights, pedestrians, other vehicles, road obstacles, etc., providing important decision-making basis for the control system. At the same time, through semantic segmentation of the image, the system can identify lane lines, road cracks, potholes and other information, and further analyze the road conditions. Combined with the depth information provided by the lidar, the actual position, size and shape of the object can be judged more accurately, avoiding the camera's misjudgment due to lighting or occlusion.

[0100] Finally, the advantages of LiDAR and camera are complemented: LiDAR has strong depth perception and can accurately obtain three-dimensional information of objects; while the camera has more advantages in detail perception and object classification, and can identify complex scenes and subtle road features. The combination of the two makes the perception system understand the environment more comprehensively and accurately, providing the subsequent torque control revision module with the function of predicting and perceiving road conditions.

[0101] 3. Adaptive drive torque revision module

[0102] Adaptive dynamic torque control algorithm: According to the vehicle status, external environment, and positioning information provided by multi-sensor data, the drive torque output value is adjusted in real time to optimize the acceleration, deceleration, and stability of the vehicle. By pre-designing decision-making control algorithms for different driving scenarios: power compensation algorithm under climbing and headwind conditions, power adaptation algorithm under low-adhesion road conditions, and power buffering algorithm for potholes and obstacles, the control system can flexibly adjust the torque output according to factors such as vehicle status, road conditions, slope changes, and positioning information to maintain the best traction and power response, and improve the vehicle's power, safety, and comfort.

[0103] Power compensation algorithm in climbing and headwind conditions: In the power compensation algorithm in complex driving conditions such as climbing and headwind, the driver's driving intention and the corresponding basic required torque are first determined by collecting information on the accelerator pedal, brake pedal and current gear position. Next, the environmental information such as the road slope is obtained through the environmental perception fusion algorithm (the vehicle positioning information is combined with the lidar and camera to analyze the slope information of the road ahead in real time), and the impact of the driving environment ahead on the vehicle power is analyzed in advance. Finally, based on the working conditions of the vehicle on a conventional (flat and low wind speed) road surface, the factors affecting the current vehicle by the environment are calculated, and this factor is used to adjust the final required torque command. Finally, the system will output a compensation torque command suitable for the current environment, based on the basic required torque command to meet the driver's demand for power in climbing and headwind conditions.

[0104] The specific process is as follows: First, the accurate vehicle positioning information is used to analyze whether the current road is sloped, and the laser radar and camera are used to detect in real time whether the road ahead is actually sloped, and the slope of the road ahead is analyzed, and part of the driving torque is compensated in advance according to the slope. Then, the acceleration information of the current vehicle is collected through the IMU, which is defined as a 1 Then, according to the vehicle status information (torque output by the drive system and current vehicle speed), an ideal expected acceleration is calculated, and this ideal expected acceleration is defined as a 2 , this expected acceleration is the expected acceleration without external environmental influence (i.e. on a flat, dry, non-slope road surface with low wind conditions). After obtaining these two acceleration values, calculate the environmental influence factor Q, the formula is:

[0105] Q=(a 2 -a 1 ) / a 2

[0106] According to the above formula, when the actual acceleration a 1 Less than the expected acceleration a 2When the Q value is positive, it means that the current output torque is less than the driver's expectation. At this time, the output torque of the drive system needs to be increased to compensate for the impact of environmental conditions on vehicle acceleration. 1 Greater than a 2 , the Q value is negative, indicating that the output torque is greater than the driver's expectations. At this time, the output torque of the drive system should be reduced to adjust the vehicle's driving state to make it more in line with the driver's needs.

[0107] Power adaptation algorithm under low adhesion road conditions: This algorithm module is a drive torque control solution suitable for distributed drive vehicles. It uses the perception and positioning module (which is a fusion of laser radar and camera) to judge the road conditions ahead (that is, the perception module that combines laser radar and camera to judge the road conditions ahead). When the road ahead is in ice, snow, wet, dry, muddy, or dark ice sections (specifically, the sensor module obtains environmental road information, and the camera and laser radar fusion module extracts road condition information to judge the road conditions ahead (ice, snow, wet, dry, muddy, or dark ice sections) in advance, providing information sources for the adaptive drive module, so as to facilitate the early adoption of auxiliary torque control), the perception module will judge the adhesion coefficient of the road ahead in advance, thereby dynamically adjusting the slip rate threshold and intelligently allocating the driving torque between the slipping axle and the non-slipping axle in advance, thereby improving the driving efficiency and vehicle stability under complex road conditions. The aforementioned sensor module is used to monitor the road adhesion coefficient and vehicle acceleration information in real time, and the slip rate threshold is dynamically set according to different road conditions to ensure that this module intervenes before the TCS (traction control system) is triggered to optimize the torque distribution between the two axles. When a single-axle slip is detected, the system adjusts the torque transfer speed between the two axles according to the degree to which the slip rate exceeds the threshold. The higher the slip rate, the faster the transfer speed, by reducing the torque of the slipping axle and transferring it to the non-slipping axle. At the same time, the transfer ratio and adjustment coefficient can be dynamically adjusted according to the real-time working conditions, thereby ensuring that the torque distribution strategy can flexibly adapt to different road conditions and driving conditions. This control module can be widely used in a variety of complex road scenarios, providing vehicles with efficient and stable driving torque control.

[0108] Pothole and obstacle road dynamic buffer algorithm: In order to reduce the driving torque of the vehicle in advance and optimize the passing experience under complex road conditions (such as speed bumps, water accumulation, potholes, etc.), the system needs to deeply coordinate the three modules of perception, decision-making and control. First, the perception module uses multi-sensor fusion technology, combined with lidar and camera through deep learning algorithms (convolutional neural network CNN) to identify road features (obstacle types, speed bumps), and can use time-series-based LSTM neural networks to predict dynamically changing road features (water accumulation expansion trend or pothole boundary changes).

[0109] The decision module first applies the target detection algorithm to classify the obstacle type in real time, and models the geometric features of speed bumps and potholes, and combines the suspension dynamics model (suspension mass-stiffness-damping system model) to predict the vibration and impact intensity. In the control module, the model predictive control (MPC) algorithm is used to optimize the vehicle's state (such as vehicle speed, pitch angle, tire impact) for a period of time in the future to generate the optimal drive torque adjustment and suspension adjustment strategy. And as the vehicle's driving and passing road history data increases, this module can also use the reinforcement learning method (based on DQN or PPO adaptive optimization model) to achieve continuous learning and optimization of torque and suspension adjustment strategies under different road conditions. The control module is responsible for implementing the strategy output by the decision module. The torque adjustment adopts a combination of feedforward control and feedback control. The feedforward control presets a reasonable torque output range based on the road condition characteristics identified in advance (such as reducing the torque before the speed bump to reduce the impact), and the feedback control dynamically corrects the torque by real-time monitoring of the vehicle posture sensor data (such as accelerometer, gyroscope) to ensure that the vehicle passes smoothly.

[0110] Regional torque limitation: This module realizes auxiliary limitation of the vehicle drive torque control system by combining the data fusion positioning module with high-precision map geographic information, helping the vehicle to realize the drive torque limitation function based on specific areas and improve the vehicle's safe driving ability. Specifically, the vehicle position is located in real time through the data fusion positioning module and the high-precision map, and it is identified whether the area where the vehicle is located belongs to the preset specific area (such as downtown roads, parking lots, mountainous roads, expressways, icy roads and other high-risk scenes). If the regional characteristics are matched successfully, the vehicle's drive torque value output is limited according to the preset maximum speed of the area. In the control logic, when the normal drive torque demand of the vehicle is lower than the limit value, it is output according to the demand. When it exceeds the limit value, it is forcibly limited to the maximum allowable value, effectively avoiding driving hazards caused by misoperation or driving without safety awareness.

[0111] 4. Security warning visualization module

[0112] Visualization module: The current road conditions (slippery, dry, icy, snowy, climbing, etc.) and surrounding road environment information (pedestrians, vehicles, obstacles, speed bumps, etc.) are displayed on the on-board screen and marked on the display screen. The on-board screen will also prompt that the torque is limited and the limited value, making it easier for the driver to observe the surrounding environment and ensure safe driving.

[0113] Safety warning module: When the vehicle perception and positioning module detects that the road conditions are about to change or that an obstacle is about to be encountered, and the driver has no intention of slowing down, the system will provide real-time suggestions and warnings through the on-board screen, with language and screen flashing operations, and ask and inform the driver whether the system needs to be started to ensure driving safety and comfort.

[0114] When this embodiment is in use, after the driver starts the driving torque assist safety system, the system will collect data from the laser radar, camera, IMU, and GPS sensors through the industrial computer and cooperate with the data processing and detection algorithm to output environmental information, positioning information, and vehicle status information, and generate the next control instruction to transmit to the VCU controller. The VCU collects wheel speed sensor data and combines it with the control instruction transmitted by the industrial computer, and cooperates with the control algorithm module to generate the revised output value of the driving torque assistance. Finally, the generated driving torque control instruction is transmitted to the MCU (motor controller) through CAN communication to realize the vehicle's driving torque revision and output. At the same time, the current surrounding environment information is displayed in real time on the vehicle screen and the features are marked. When the driving torque assist system is not started, the system will generate an alarm message in an emergency to remind the driver to pay attention to the road conditions ahead and ask whether the system needs to be started. The above systems are all powered by the on-board low-voltage battery.

[0115] 5. Technical Effects

[0116] By integrating multi-sensor data fusion technology, the power response accuracy, safety and stability of electric vehicles under complex road conditions are significantly improved. The specific technical effects are as follows: the system can accurately perceive the dynamic state of the vehicle and the external environment, including changes in road adhesion, obstacles and road conditions, and achieve rapid adaptation to slippery, icy, muddy and other low-adhesion road surfaces; by optimizing the drive torque output and dynamic adjustment strategy, the system effectively reduces the risks of slipping, loss of control, etc., and improves the vehicle's handling and traction; at the same time, the system's predictive control capabilities enable it to identify road conditions in advance and take preventive adjustment measures to avoid danger; combined with the regional torque limiting function, the system intelligently controls torque output in specific areas (such as downtown areas, parking lots, etc.) to ensure driving safety. Overall, the present invention significantly enhances the vehicle's power performance, driving safety and driving comfort, and provides reliable technical support for the application of intelligent driving and electric vehicles in complex environments. The specific technical effects are as follows:

[0117] 1. Improve vehicle dynamic response accuracy

[0118] Through multi-sensor data fusion technology, the present invention can accurately sense the dynamic state of the vehicle (such as vehicle speed, acceleration, steering angle, etc.) and the external environment (such as road adhesion and condition, obstacles, slope changes, etc.). Based on these high-precision perception information, the system adjusts the drive torque output in real time, allowing the vehicle to respond quickly and accurately to different driving conditions (such as acceleration, braking, turning, etc.), ensuring that the vehicle's power performance is always in the best state.

[0119] 2. Enhance the ability to adapt to complex road conditions

[0120] The present invention can accurately identify and predict road conditions (such as slippery, waterlogged, slushy, muddy, etc.), and dynamically adjust the drive torque output according to the real-time detected changes in road adhesion. The system has the ability to predict sudden changes in road conditions (such as waterlogging, black ice, potholes, etc.) in advance, making the adjustment of drive torque more forward-looking and avoiding unsafe phenomena such as power loss of control or vehicle skidding due to changes in road adhesion.

[0121] 3. Improve vehicle safety and stability

[0122] Through multi-sensor fusion, the system can comprehensively judge the potential risks in complex driving environments and take measures in advance to adjust the torque, thereby improving the stability and controllability of the vehicle. Especially in complex driving conditions (such as mountainous areas, rainy and snowy weather in cities, etc.), the present invention can adjust the power output in time to prevent the vehicle from losing control due to emergencies, significantly improving driving safety.

[0123] 4. Optimize the distribution and adjustment of driving torque

[0124] The dynamic torque control algorithm of the present invention can flexibly adjust the distribution ratio of driving torque according to factors such as vehicle load, road conditions, slope changes, etc., to ensure that the vehicle always maintains the best traction and power response under different road conditions. Especially under low adhesion or special road conditions (such as icy roads or muddy roads), the system can effectively avoid problems such as vehicle slippage, excess or insufficient power, and ensure a smooth driving experience.

[0125] 5. Improve the driver's operating experience

[0126] The system optimizes the vehicle's driving torque output and reduces the driver's operating burden by providing real-time feedback of vehicle status information (such as wheel speed, vehicle speed, road adhesion, etc.) and intelligent control strategies. The driver can obtain real-time vehicle status information through the human-computer interaction interface and receive necessary warnings or suggestions, thereby enhancing the driver's sense of security and confidence.

[0127] 6. Strong environmental adaptability and flexibility

[0128] The control system of the present invention has strong adaptability. Whether in urban roads, highways or complex mountainous areas, it can respond quickly according to real-time environmental information and optimize the power output and handling stability of the vehicle. The flexibility of the system enables it to be widely used in many fields such as electric vehicles, intelligent driving systems and high-performance vehicles.

[0129] 7. Reduce accident risks and power waste

[0130] The intelligent control system of the present invention can effectively reduce the risk of accidents caused by factors such as changes in road adhesion, weather changes, and improper driver operation, and avoid power waste by accurately adjusting the driving torque. By optimizing the control strategy, the stability of the vehicle in complex environments is improved, and the probability of accidents is reduced.

[0131] The present invention innovatively proposes to combine multi-sensor data fusion technology and dynamic drive torque control algorithm, which significantly improves the adaptability, safety and stability of the vehicle under complex road conditions, while optimizing the vehicle's power performance and driving experience, and has broad application prospects.

[0132] 8. The system can accurately perceive the vehicle's operating status, external environment, vehicle position and other key information in real time by integrating data from various sensors (such as wheel speed sensors, inertial measurement units IMU, GPS, visual sensors and laser radar). This information will be fully used to identify the current driving scene and the road conditions ahead, and provide information sources for the subsequent drive torque auxiliary control strategy to ensure the safety, power and comfort during driving. Specifically, the present invention aims to optimize the drive torque output value by combining data from multiple sensors, improve the road adaptability, safety and stability of the vehicle, ensure that the vehicle can efficiently and safely adjust the power under various complex road conditions (such as slippery roads, icy roads, muddy roads, potholes, roads with obstacles, speed bumps, etc.), and maximize the vehicle's power, safety and handling stability.

[0133] 9. The highlight of the present invention is that it integrates multi-sensor data to perceive the changes in the external environment in real time and combines the vehicle status and positioning information, and adopts different control strategies to revise the drive torque output value, aiming to enhance the driving experience, adapt to different road conditions, and improve vehicle driving safety.

[0134] 10. The present invention firstly integrates the data of multiple sensors (wheel speed sensor, IMU, GPS, visual sensor, lidar) to effectively overcome the problem that a single sensor is limited by environmental changes and error influences, thereby improving the accuracy and reliability of the perceived information. The system is able to accurately identify the road conditions ahead (such as wetness, water accumulation, ice and snow, potholes, slopes, speed bumps, obstacles, etc.) in advance, and adopt different drive torque auxiliary control strategies in advance in combination with dynamic environment and positioning information to help the vehicle drive safely. This feature will greatly enhance the safety and stability of the vehicle under complex driving conditions. Through multi-dimensional information fusion, the system can perceive the vehicle and environmental status more comprehensively and accurately, thereby making more scientific drive torque control decisions.

[0135] 11. Combining real-time data from multiple sensors, the system can accurately identify changes in road adhesion and vehicle dynamics, and then dynamically adjust the drive torque output according to the current driving environment and vehicle condition. This high adaptability can optimize the vehicle's acceleration performance and handling stability in a variety of complex driving environments (such as mountainous areas, slippery roads, flooded roads, etc.), ensuring that the vehicle exhibits the best power response under different working conditions. In addition, by predicting changes in road conditions and vehicle dynamics information in advance, combined with real-time feedback of wheel speed conditions, it can respond quickly in emergencies. This effective combination of active and passive control allows the system to maintain vehicle stability and quickly adjust torque output when responding to sudden road conditions (such as sudden slippery, icy and snowy roads, etc.), avoiding loss of control or insufficient power.

[0136] 12. Through precise driving torque control, the vehicle's dynamic performance is improved, and comfort and safety during driving are ensured. Especially in complex road conditions and driving environments, the present invention can significantly improve the vehicle's handling stability, reduce the driver's operating burden, and reduce the risk of accidents caused by factors such as changes in road adhesion. The system of the present invention can flexibly adapt to different load conditions, slope conditions and changing road conditions. This feature makes the present invention more adaptable to application scenarios, and it can provide excellent driving performance whether it is urban roads, highways or complex mountainous roads. Moreover, as vehicles become increasingly intelligent today, this product does not need to add sensors and actuator components to vehicles equipped with automatic driving functions, and only needs to adapt and integrate the control system before use.

[0137] 6.Technological advantages

[0138] Improve safety: Through multi-sensor data fusion, the system can fully perceive road conditions, vehicle dynamics and environmental changes, predict and take control measures in advance, and significantly improve driving safety.

[0139] Adapt to changing road conditions: The system can accurately identify and respond to different types of road surfaces (such as ice, snow, wetness, mud, etc.), ensuring the stability and controllability of the vehicle in various complex environments.

[0140] Optimizing dynamic performance: By precisely adjusting the drive torque, the system can maximize the vehicle's acceleration performance and reduce power waste or slip caused by changes in road adhesion.

[0141] Enhanced driving experience: The system optimizes torque output during driving through intelligent control strategies and real-time feedback, ensuring the vehicle's smooth ride and reducing the driver's operating burden.

[0142] Strong adaptability and flexibility: Whether in mountainous areas, urban roads, highways, or complex road environments, this system can respond flexibly and provide excellent performance.

[0143] The multi-sensor data fusion drive torque control system provided by the present invention has significant advantages in improving vehicle safety, stability and power performance due to its innovative technical solutions and high adaptability, and can meet the increasingly complex control needs of future smart cars and electric cars.

[0144] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the present invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A multi-sensor data fusion driving torque safety auxiliary control system, characterized in that: It includes a sensor module, a multi-sensor data fusion and processing module, and an adaptive drive torque revision module; The sensor module is used to detect the speed, position, motion information, road surface information and surrounding environment of the vehicle; The multi-sensor data fusion and processing module is used to process the data provided by the sensor module and combine the role of each sensor in the sensor module in different scenarios; The adaptive driving torque revision module is used to adjust the driving torque output value in real time according to the vehicle status, external environment, and positioning information provided by the multi-sensor data fusion and processing module.

2. The multi-sensor data fusion driving torque safety auxiliary control system according to claim 1 is characterized in that: The sensor module includes a wheel speed sensor, an inertial measurement unit, a GPS module, a visual sensor and a laser radar; The wheel speed sensor is used to monitor the rotation speed of each wheel in real time and provide the actual driving speed and power distribution information of the vehicle; The inertial measurement unit is used to monitor the vehicle's acceleration, tilt angle, yaw angle and other dynamic data through an accelerometer and a gyroscope, provide the control system with vehicle motion status information and assist in positioning; The GPS module is used to provide geographic information such as the real-time position, speed and heading angle of the vehicle, and determine the precise position and driving direction of the vehicle on the road; The visual sensor is used to collect the road conditions in front of the vehicle in real time, identify complex road conditions, and assist in control decision making; The laser radar is used to provide the vehicle with high-precision three-dimensional point cloud data of the surrounding environment, and cooperates with the visual sensor to accurately perceive obstacles, ramps, potholes and road conditions, and judge the driving environment.

3. The multi-sensor data fusion driving torque safety auxiliary control system according to claim 2 is characterized in that: The multi-sensor data fusion and processing module processes the information provided by the sensor module in the following steps: A. The GPS module provides global positioning information of the vehicle, the inertial measurement unit provides acceleration and angular velocity information, the visual sensor provides visual image information, the laser radar provides accurate three-dimensional environmental point cloud data, and the wheel speed sensor measures the vehicle's motion state; B. Synchronize the data provided in step A in time and perform preprocessing; C. Extract features from the data obtained in step B; D. Fusion of data from multiple sensors based on the extracted features; E. Build a deep learning model and use the data obtained in step D as training data for the deep learning model; F. Based on the scene information, the deep learning model obtained in step E dynamically adjusts the weight of the sensor.

4. The multi-sensor data fusion driving torque safety auxiliary control system according to claim 3 is characterized in that: In step C, feature extraction is performed on the data obtained in step B. The steps are as follows: Extract environmental features from the images of the visual sensor through a convolutional neural network, including at least road signs, pedestrians, other vehicles and road conditions; Extract three-dimensional environment information from LiDAR through point cloud processing network; Capturing motion changes in time series data from wheel speed sensors and inertial measurement units using long short-term memory networks.

5. The multi-sensor data fusion driving torque safety auxiliary control system according to claim 1, characterized in that: The adaptive driving torque revision module is used to adjust the driving torque output value in real time according to the vehicle status, external environment, and positioning information provided by the multi-sensor data fusion and processing module. The steps are as follows: 1) According to different driving scenarios, construct a power compensation algorithm for climbing and headwind conditions, a power adaptation algorithm for low-adhesion road conditions, and a power buffer algorithm for potholes and obstacles on the road; 2) By capturing the algorithm in 1), the torque output is flexibly adjusted according to vehicle status, road conditions, slope changes, and positioning information factors.

6. The multi-sensor data fusion driving torque safety auxiliary control system according to claim 5, characterized in that: The power compensation algorithm under the climbing and headwind conditions is constructed in the following steps: Analyze whether the current road has a slope through accurate vehicle positioning information, and use laser radar and cameras to detect in real time whether the road ahead actually has a slope; Analyze the slope of the road ahead and compensate part of the driving torque in advance according to the slope; The acceleration information of the current vehicle is collected through IMU, which is defined as a1; According to the vehicle status information, the expected acceleration is calculated, which is defined as a2; The impact factor Q of the calculation environment is expressed as: Q=(a2-a1) / a2 Determine whether to adjust and increase the output torque of the drive system according to the Q value; If the Q value is positive, the output torque of the drive system is increased; If the Q value is negative, the output torque of the drive system is reduced.

7. The multi-sensor data fusion driving torque safety auxiliary control system according to claim 5, characterized in that: The power adaptation algorithm under low adhesion road conditions has the following steps: The sensor module monitors the road adhesion coefficient and vehicle acceleration information in real time, dynamically sets the slip rate threshold according to different road conditions, and optimizes the torque distribution between the two axles; When a single-axle slip state is detected, the torque transfer speed between the two axes is adjusted according to the degree to which the slip ratio exceeds the threshold. The higher the slip ratio, the faster the transfer speed, by reducing the torque of the slipping axle and transferring it to the non-slipping axle.

8. The multi-sensor data fusion driving torque safety auxiliary control system according to claim 5, characterized in that: The pothole and obstacle road dynamic buffer algorithm has the following steps: The perception module uses multi-sensor fusion technology, combined with lidar and cameras, to identify road features through deep learning algorithms, and uses a time-series-based LSTM neural network to predict dynamically changing road features; The decision module uses the target detection algorithm to classify obstacle types in real time, and models the geometric features of speed bumps and potholes, and combines the suspension dynamics model to predict vibration and impact intensity; The control module adjusts the torque according to the output of the decision module; The torque adjustment adopts a combination of feedforward control and feedback control, and the steps are as follows: Feedforward control presets a reasonable torque output range based on road condition characteristics identified in advance; Feedback control dynamically corrects the torque by real-time monitoring of vehicle posture sensor data to ensure smooth passage of the vehicle.

9. The multi-sensor data fusion driving torque safety auxiliary control system according to claim 1, characterized in that: It also includes a security warning visualization module; The safety warning visualization module includes a visualization module and a safety warning module; The visualization module is used to display the current road status and surrounding road environment information through the vehicle screen, and mark them on the display screen. The vehicle screen will also prompt that the torque is limited and the limited value, so that the driver can observe the surrounding environment and ensure safe driving; The safety warning module is used to provide real-time suggestions and warnings through language and screen flashing operations on the vehicle screen when it detects that the road conditions are about to change or a road obstacle is about to be encountered, and the driver has no intention of slowing down the vehicle, and asks and informs the driver whether he needs to start the system to ensure driving safety and comfort.

Images