An autonomous driving path optimization control system fusing environment perception and prediction

Abstract

The application discloses a kind of automatic driving path optimization control systems of fusion environment perception and prediction, and the application relates to the field of automatic driving technology system.The automatic driving path optimization control system of fusion environment perception and prediction, including environment perception and data acquisition module, safety distance and lane change space analysis module, brake condition monitoring and risk prediction module, safety area determination and lane change trajectory planning module, lane change process monitoring and dynamic adjustment module, abnormal close-in vehicle avoidance module;Brake condition is judged by image processing, can predict the deceleration risk of preceding vehicle in advance, enhance safety, brake frequency is counted and brake probability is predicted in combination with multiple factors, so that lane change risk assessment is more scientific and reasonable;According to the planning of risk assessment result, and calculate specific control parameters, so that lane change operation is more standardized and controllable;Model predictive control algorithm is used to optimize input, and the stability of lane change is improved.Abnormal close-in vehicle avoidance module can respond to emergency situation in time, and ensure driving safety.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology systems, specifically to an autonomous driving path optimization control system that integrates environmental perception and prediction. Background Technology

[0002] Autonomous driving technology is a core technology system that integrates sensors (such as LiDAR, cameras, and millimeter-wave radar), high-precision maps, artificial intelligence algorithms (such as deep learning and computer vision), and complex decision-making systems to enable vehicles to achieve environmental perception, path planning, dynamic decision-making, and safety control without active human intervention. Its core objective is to simulate human driving behavior and exceed its capabilities by analyzing road information (such as obstacles, traffic signs, and pedestrian dynamics) and vehicle status in real time (such as all-time, all-domain monitoring and millisecond-level response), ultimately achieving a graded evolution from assisted driving (such as adaptive cruise control and automatic parking) to fully autonomous driving (Level 5).

[0003] Publication No. CN119568197A discloses an autonomous driving path optimization control method and device that integrates environmental perception and prediction. The method includes: acquiring environmental perception data of the autonomous vehicle through radar and cameras; using the initial posture, target posture, and environmental perception data of the autonomous vehicle as road constraints, and planning an initial driving path of the autonomous vehicle from the current position to the target position through a path planning algorithm; estimating and predicting the state of surrounding vehicles and obstacles through a multi-object tracking algorithm; optimizing the initial path to obtain an optimized driving path based on the future motion trajectories of surrounding vehicles and obstacles, road constraints, and physical limitations of the autonomous vehicle; converting the optimized driving path into control commands for the autonomous vehicle, and driving the autonomous vehicle smoothly along the optimized driving path through the control commands.

[0004] The above technologies represent conventional solutions for existing autonomous driving technologies, which plan paths by perceiving surrounding objects. However, such autonomous driving technologies can only achieve relatively stable and passive driving. In situations requiring lane changes or overtaking, especially under complex conditions with heavy traffic, lane changes face many uncontrollable factors, such as sudden braking by the vehicle in front, acceleration by the vehicle behind, or road factors causing the vehicle to skid during a circumferential lane change. Existing autonomous driving control technologies are insufficient to cope with these uncontrollable external factors, thus requiring further research. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an autonomous driving path optimization control system that integrates environmental perception and prediction, thus solving the problems described above.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an autonomous driving path optimization control system integrating environmental perception and prediction, comprising:

[0007] Environmental perception and data acquisition module: It acquires the position, speed and acceleration information of the vehicle and surrounding vehicles in real time through on-board sensors, identifies vehicles using image recognition algorithms, tracks motion state using Kalman filtering, and monitors the status of vehicle taillights through image processing to determine braking status;

[0008] Safe Distance and Lane Change Space Analysis Module: Sets a safe distance threshold, establishes a three-dimensional spatial model to analyze lane change space, calculates critical distances and predicts the risk of deceleration of the vehicle in front, ensuring sufficient safe distance;

[0009] Braking situation monitoring and risk prediction module: Counts the number of braking times of surrounding vehicles, calculates the braking frequency, combines multiple factors, uses a logistic regression model to predict the probability of braking, and comprehensively assesses the risk level of lane changing.

[0010] Safe zone determination and lane change trajectory planning module: Determines safe lane change areas based on risk assessment results, plans lane change trajectories, and then calculates steering wheel parameters and gear shift parameters for lane change control.

[0011] Lane change process monitoring and dynamic adjustment module: continuously collects vehicle parameters during the lane change process, combines real-time environmental information to predict lane change risks, and immediately stops the lane change or adjusts the lane change path if the risk exceeds the threshold. The control input is optimized by using model predictive control algorithm.

[0012] Abnormal Approach Vehicle Avoidance Module: Real-time identification of abnormally approaching vehicles, planning of a safe range that can be moved to the other side, and using a fast random search tree algorithm to plan the avoidance trajectory.

[0013] Preferably, let the road length direction be the X direction, the road width direction be the Y direction, let the vehicle be a0, the vehicle in front of vehicle a0 in the same lane be a1, the vehicle in front of vehicle a2 in the lane where vehicle a0 is waiting to change lanes be a2, the vehicle behind vehicle a2 be a3, the vehicle in front of vehicle a2 be a4, and the vehicle close to vehicle a0 in the other lane on the other side be a5.

[0014] Preferably, the environmental perception and data acquisition module specifically includes the following functions:

[0015] Data Acquisition: Real-time position coordinates (x, y) of vehicle a0 and surrounding vehicles a1, a2, a3, a4, and a5 are acquired using sensors such as onboard cameras or radar. i y i ), velocity v i acceleration a i Direction angle θ iInformation such as i = (0, 1, 2, 3, 4, 5); at the same time, obtain the dimensions of the vehicle a0, vehicle speed v0, steering wheel angle δ0, and gear shifting parameters, where the dimensions include length L0 and width W0, and the gear shifting state is represented by acceleration A0;

[0016] Vehicle status recognition: Using image recognition algorithms (such as YOLO algorithm) combined with sensor data, each vehicle is accurately identified and numbered. The Kalman filter algorithm is used to continuously track its motion state and estimate the vehicle's position and speed in the future.

[0017] The taillight status of vehicles a1 and a2 is monitored in real time using image processing technology. If the taillight brightness exceeds a set threshold I... th If both taillights are lit simultaneously, it indicates that the vehicle is in a braking state.

[0018] Preferably, the specific functions of the safety distance and lane change space analysis module include:

[0019] Establish a spatial model: Based on the dimensions and speed of vehicle a0, set the safety distance threshold required for lane changing; let the safety distance D be... safe The relationship with vehicle speed v0 is D safe =k·v0+b, where k and b are constants obtained by fitting experimental data, for example, k=0.5, b=5 (unit: m);

[0020] Using vehicle a0 as the center and combining the position information of vehicles a1, a2, and a3, a three-dimensional spatial model is established. On the XY plane, with the center of vehicle a0 as the origin, the positions of vehicles a1, a2, and a3 constitute spatial points. It is then determined whether vehicle a0 meets the safety distance requirements with surrounding vehicles in the X and Y directions, i.e., |x0−x i |≥D safe and (i=1, 2, 3);

[0021] Calculating critical distances and predicting deceleration risks: Real-time calculation of the distance d between vehicle a0 and vehicle a2 in the X direction. 02X =∣x0−x2∣, and the X-direction distance d between vehicle a2 and the vehicle a4 in front of it. 24X = |x2−x4|;

[0022] Monitoring d 24X rate of change Let the velocity threshold be v th For example, 2m / s, if If so, it is predicted that vehicle a2 is at risk of deceleration.

[0023] Preferably, the braking condition monitoring and risk prediction module specifically includes the following functions:

[0024] Braking frequency calculation: Count the number of times vehicles a1 and a2 brake within time window T, N. a1 and N a2 Calculate the braking frequencies of vehicles a1 and a2. and ;

[0025] Braking probability prediction: combined with braking frequency f ai Vehicle speed v i Acceleration A i In addition to the current traffic flow factor F, a logistic regression model is used to predict the probability P of vehicles a1 and a2 braking during the lane change of vehicle a0. a1 and P a2 The formula for the logistic regression model is:

[0026] ;

[0027] Where β0, β1, β2, β3, and β4 are model parameters, obtained through training with historical data;

[0028] Lane change risk assessment: Based on the braking probability, the deceleration risk of vehicle a2, and the lane change space analysis results, the lane change risk level R is comprehensively assessed; let the risk assessment function be:

[0029] ;

[0030] Among them, w1, w2, w3, and w4 are weighting coefficients, and I(·) is an indicator function, which takes the value of 1 when the condition is met and 0 otherwise. Based on the value of R, the lane change risk is divided into three levels: low (R<0.3), medium (0.3≤R<0.7), and high (R≥0.7).

[0031] Preferably, the specific functions of the safe zone determination and lane change trajectory planning module include:

[0032] Safe lane change zone determination: When the lane change risk assessment is low or medium, a safe lane change zone is determined; the safe lane change zone should meet the requirements of vehicle a. Maintain a sufficient safe distance from surrounding vehicles (a1, a2, a3) in both the X and Y directions, i.e. ;

[0033] Lane change trajectory planning: A fifth-order polynomial spline curve interpolation algorithm is used to plan the lane change trajectory; let the displacements of the lane change trajectory in the X and Y directions be x(t) and y(t), respectively, and their expressions are:

[0034] x(t) = c0 + c1t + c2t 2 +c3t 3 +c4t 4 +c5t 5;

[0035] y(t) = b0 + b1t + b2t 2 +b3t 3 +b4t 4 +b5t 5 ;

[0036] Solve for the polynomial coefficients c based on the initial and objective conditions. j and b j (j=0, 1, ..., 5), initial conditions include current position, velocity, and acceleration, target conditions include target lane position, velocity, and acceleration;

[0037] Parameter calculation: Based on the planned lane change trajectory, calculate the steering wheel parameters and gear shift parameters. The steering wheel parameters include the angle δ(t) and the steering speed. Where L is the vehicle wheelbase; the steering speed is calculated based on the rate of change of the curvature of the trajectory; the gear shift parameters are adjusted according to the slope change of the trajectory and the target speed requirement.

[0038] Preferably, the specific functions of the lane change process monitoring and dynamic adjustment module include:

[0039] Vehicle parameter collection: During lane change, continuously collect the vehicle's steering wheel angle δ0(t), vehicle speed v0(t), acceleration A0(t), and vehicle attitude parameters; vehicle attitude parameters include vibration amplitude d(t) and drift distance D. drift (t), the shaking amplitude is obtained by measuring the acceleration sensor, and the drift distance is calculated by the deviation between the vehicle position and the planned trajectory;

[0040] Risk Prediction and Adjustment: Using collected vehicle parameters and real-time environmental information, predict the risk of vehicle lane changes; let the risk prediction function be:

[0041] R dynamic =w5A0(t)+w6D drift (t)+w7∣v0(t)−v target (t)∣;

[0042] Where w5, w6, and w7 are weighting coefficients, v target (t) represents the target vehicle speed; the dynamic risk threshold is set to R. th If R dynamic >R th If the lane change is not possible, immediately stop changing lanes or adjust the lane change path.

[0043] When adjusting the lane change path, a new safety distance analysis, trajectory planning, and parameter calculation are performed. The Model Predictive Control (MPC) algorithm is used to optimize the control input for a future period of time, including steering wheel angle and acceleration, based on the current vehicle status and environmental information in each control cycle, so that the vehicle can complete the lane change safely and smoothly.

[0044] Preferably, the specific functions of the abnormal approaching vehicle avoidance module include:

[0045] Spacing recognition: Real-time recognition of the X and Y distances between vehicles a3 and a5 and this vehicle a0.

[0046] ;

[0047] Where, d 03X (t) represents the distance between vehicle a3 and this vehicle a0 in the X direction, d 05X (t) represents the distance between vehicle a5 and this vehicle a0 in the X direction, d 03Y (t) represents the distance between vehicle a3 and this vehicle a0 in the Y direction, d 05Y (t) represents the distance between vehicle a5 and this vehicle a0 in the Y direction;

[0048] Anomaly detection and avoidance planning: When vehicle a3 or a5 abnormally approaches vehicle a0, i.e., d 03X (t)<0 or d 05X (t)<0, and d 03Y (t) <d th or d 05Y (t) <d th d th For a given spacing threshold, such as 1m, plan the safe range within which movement to the other side is permitted; the safe range should meet the safe distance requirements from surrounding vehicles.

[0049] Based on the safe range, the Fast Random Search Tree (RRT) algorithm is used to plan the avoidance trajectory and control the vehicle to move safely to the other side to avoid a collision.

[0050] Preferably, the safety distance D safe During the calculation process, the influence of different road conditions on vehicle braking distance is considered. The k and b values ​​for each vehicle type are corrected, and a road condition influence coefficient table is established. According to different road conditions, the k and b values ​​are adjusted by corresponding multiplication or addition.

[0051] This invention provides an autonomous driving path optimization control system that integrates environmental perception and prediction. Compared with existing technologies, it has the following advantages:

[0052] 1. This autonomous driving path optimization control system, integrating environmental perception and prediction, comprehensively utilizes multiple methods in environmental perception. It not only acquires basic vehicle motion information but also uses image processing to determine braking conditions, resulting in more comprehensive and accurate perception. The safe distance and lane change space analysis module establishes a 3D model, enabling more intuitive and accurate analysis of lane change conditions and early prediction of the risk of deceleration by the vehicle ahead, enhancing safety. The braking condition monitoring and risk prediction module statistically analyzes braking frequency and combines multiple factors to predict braking probability, making lane change risk assessment more scientific and reasonable. The safe zone determination and lane change trajectory planning module plans based on risk assessment results and calculates specific control parameters, making lane change operations more standardized and controllable. The lane change process monitoring and dynamic adjustment module provides real-time feedback and adjustments, using model predictive control algorithms to optimize input and improve lane change stability. The abnormal approach vehicle avoidance module can promptly respond to emergencies, ensuring driving safety and comprehensively improving the reliability and safety of autonomous driving.

[0053] 2. This autonomous driving path optimization control system, which integrates environmental perception and prediction, identifies, detects, and marks the vehicle itself and multiple key vehicles in front and behind, providing data support for subsequent calculations. Considering the impact of different road conditions on braking distance, it adjusts the safety distance threshold through correction coefficients, making the judgment of safe distance for lane changes more accurate. By monitoring braking conditions and predicting risks, it can avoid accidents caused by risks such as the braking of the vehicle in front during lane changes. Based on the risk assessment results, a safe zone is determined, and a fifth-order polynomial spline curve interpolation algorithm is used to plan the trajectory and calculate steering wheel and transmission parameters, making the lane change process smoother and safer. Overall, these improvements enhance the system's real-time performance, accuracy, and adaptability, enabling it to better cope with complex and changing road environments and ensuring the safety and comfort of autonomous driving lane changes.

[0054] 3. This autonomous driving path optimization control system, which integrates environmental perception and prediction, continuously collects key vehicle parameters and vehicle attitude parameters through its lane change process monitoring and dynamic adjustment module, enabling a comprehensive understanding of the vehicle's real-time status during lane changes. By combining these parameters with real-time environmental information for risk prediction, the system can more accurately assess lane change risks. When the risk exceeds a threshold, it can promptly stop the lane change or adjust the path, effectively preventing accidents. During the lane change path adjustment, a series of analyses and plans are re-performed, and model predictive control algorithms are used to optimize the control input, allowing the vehicle to flexibly adjust according to actual conditions, ensuring a safe and smooth lane change and improving the safety and reliability of autonomous driving lane changes.

[0055] 4. This autonomous driving path optimization control system, which integrates environmental perception and prediction, can accurately and in real-time acquire the distances between vehicles in the left and right lanes in the X and Y directions. X-direction distance identification determines whether vehicles on the left and right overlap with the vehicle in the width direction. Further analysis in the Y direction and Y direction quickly identifies any abnormal approaching behavior by vehicles on the left and right and plans a safe distance accordingly. Employing the Fast Random Search Tree (RRT) algorithm to plan avoidance trajectories, it efficiently finds safe movement paths and controls the vehicle to move safely to the other side, effectively avoiding collisions. This module enhances the autonomous driving system's ability to respond to sudden dangerous situations, improves driving safety, protects the safety of occupants and the vehicle, and enhances the reliability and practicality of autonomous driving. Attached Figure Description

[0056] Figure 1 This is a schematic diagram of the main process of the present invention;

[0057] Figure 2 This is a schematic diagram of the vehicle markings and road coordinates of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] See Figures 1-2 The present invention provides the following four technical solutions:

[0060] First implementation: An autonomous driving path optimization control system integrating environmental perception and prediction, comprising:

[0061] Environmental perception and data acquisition module: It acquires the position, speed and acceleration information of the vehicle and surrounding vehicles in real time through on-board sensors, identifies vehicles using image recognition algorithms, tracks motion state using Kalman filtering, and monitors the status of vehicle taillights through image processing to determine braking status;

[0062] Safe Distance and Lane Change Space Analysis Module: Sets a safe distance threshold, establishes a three-dimensional spatial model to analyze lane change space, calculates critical distances and predicts the risk of deceleration of the vehicle in front, ensuring sufficient safe distance;

[0063] Braking situation monitoring and risk prediction module: Counts the number of braking times of surrounding vehicles, calculates the braking frequency, combines multiple factors, uses a logistic regression model to predict the probability of braking, and comprehensively assesses the risk level of lane changing.

[0064] Safe zone determination and lane change trajectory planning module: Determines safe lane change areas based on risk assessment results, plans lane change trajectories, and then calculates steering wheel parameters and gear shift parameters for lane change control.

[0065] Lane change process monitoring and dynamic adjustment module: continuously collects vehicle parameters during the lane change process, combines real-time environmental information to predict lane change risks, and immediately stops the lane change or adjusts the lane change path if the risk exceeds the threshold. The control input is optimized by using model predictive control algorithm.

[0066] Abnormal Approach Vehicle Avoidance Module: Real-time identification of abnormally approaching vehicles, planning of a safe range that can be moved to the other side, and using a fast random search tree algorithm to plan the avoidance trajectory.

[0067] In terms of environmental perception, multiple methods are comprehensively utilized to not only acquire basic vehicle motion information but also assess braking conditions through image processing, resulting in more comprehensive and accurate perception. The safe distance and lane change space analysis module establishes a 3D model, enabling more intuitive and accurate analysis of lane change conditions and early prediction of the risk of deceleration by the vehicle ahead, thus enhancing safety. The braking condition monitoring and risk prediction module statistically analyzes braking frequency and combines multiple factors to predict braking probability, making lane change risk assessment more scientific and reasonable. The safe zone determination and lane change trajectory planning module plans based on risk assessment results and calculates specific control parameters, making lane change operations more standardized and controllable. The lane change process monitoring and dynamic adjustment module provides real-time feedback and adjustments, using model predictive control algorithms to optimize input and improve lane change stability. The abnormal approach vehicle avoidance module can promptly respond to emergencies, ensuring driving safety and comprehensively improving the reliability and safety of autonomous driving.

[0068] The second implementation method differs from the first implementation method in that: the road length direction is X direction, the road width direction is Y direction, the vehicle is a0, the vehicle in front of vehicle a0 in the same lane is a1, the vehicle in front of vehicle a0 in the lane where the vehicle is about to change lanes is a2, the vehicle behind vehicle a2 is a3, the vehicle in front of vehicle a2 is a4, and the vehicle close to vehicle a0 in the other lane is a5.

[0069] The specific functions of the environmental perception and data acquisition module include:

[0070] Data Acquisition: Real-time position coordinates (x, y) of vehicle a0 and surrounding vehicles a1, a2, a3, a4, and a5 are acquired using sensors such as onboard cameras or radar. i y i ), velocity v i acceleration a i Direction angle θ iInformation such as i = (0, 1, 2, 3, 4, 5); at the same time, obtain the dimensions of the vehicle a0, vehicle speed v0, steering wheel angle δ0, and gear shifting parameters, where the dimensions include length L0 and width W0, and the gear shifting state is represented by acceleration A0;

[0071] Vehicle status recognition: Using image recognition algorithms (such as YOLO algorithm) combined with sensor data, each vehicle is accurately identified and numbered. The Kalman filter algorithm is used to continuously track its motion state and estimate the vehicle's position and speed in the future.

[0072] The taillight status of vehicles a1 and a2 is monitored in real time using image processing technology. If the taillight brightness exceeds a set threshold I... th If both taillights are lit simultaneously, it indicates that the vehicle is in a braking state.

[0073] The specific functions of the safety distance and lane change space analysis module include:

[0074] Establish a spatial model: Based on the dimensions and speed of vehicle a0, set the safety distance threshold required for lane changing; let the safety distance D be... safe The relationship with vehicle speed v0 is D safe =k·v0+b, where k and b are constants obtained by fitting experimental data, for example, k=0.5, b=5 (unit: m);

[0075] The safety distance D safe During the calculation, the impact of different road conditions on vehicle braking distance is considered, and the values ​​of k and b are adjusted for each vehicle type. For example, on wet and slippery roads, the braking distance of the vehicle will increase, so k and b need to be appropriately increased. Assuming k=0.3 and b=10 on dry roads, on wet and slippery roads, k is adjusted based on experimental data. wet =0.4, b wet =12. Establish a road condition influence coefficient table. Adjust the values ​​of k and b accordingly using multiplication or addition based on different road conditions. For example, for icy or snowy roads, the influence coefficient might be 1.5. Then, the k value for a car... ice =0.3×1.5=0.45, b ice =10 × 1.5 = 15.

[0076] Using vehicle a0 as the center and combining the position information of vehicles a1, a2, and a3, a three-dimensional spatial model is established. On the XY plane, with the center of vehicle a0 as the origin, the positions of vehicles a1, a2, and a3 constitute spatial points. It is then determined whether vehicle a0 meets the safety distance requirements with surrounding vehicles in the X and Y directions, i.e., |x0−x i |≥D safe and (i=1, 2, 3)

[0077] Calculating critical distances and predicting deceleration risks: Real-time calculation of the distance d between vehicle a0 and vehicle a2 in the X direction. 02X =∣x0−x2∣, and the X-direction distance d between vehicle a2 and the vehicle a4 in front of it. 24X = |x2−x4|;

[0078] Monitoring d 24X rate of change Let the velocity threshold be v th For example, 2m / s, if If so, it is predicted that vehicle a2 is at risk of deceleration.

[0079] The specific functions of the braking condition monitoring and risk prediction module include:

[0080] Braking frequency calculation: Count the number of times vehicles a1 and a2 brake within time window T, N. a1 and N a2 Calculate the braking frequencies of vehicles a1 and a2. and ;

[0081] Braking probability prediction: combined with braking frequency f ai Vehicle speed v i Acceleration A i In addition to the current traffic flow factor F, a logistic regression model is used to predict the probability P of vehicles a1 and a2 braking during the lane change of vehicle a0. a1 and P a2 The formula for the logistic regression model is:

[0082] ;

[0083] Where β0, β1, β2, β3, and β4 are model parameters, obtained through training with historical data;

[0084] Lane change risk assessment: Based on the braking probability, the deceleration risk of vehicle a2, and the lane change space analysis results, the lane change risk level R is comprehensively assessed; let the risk assessment function be:

[0085] ;

[0086] Among them, w1, w2, w3, and w4 are weighting coefficients, and I(·) is an indicator function, which takes the value of 1 when the condition is met and 0 otherwise. Based on the value of R, the lane change risk is divided into three levels: low (R<0.3), medium (0.3≤R<0.7), and high (R≥0.7).

[0087] The specific functions of the safe zone determination and lane change trajectory planning module include:

[0088] Safe lane change zone determination: When the lane change risk assessment is low or medium, a safe lane change zone is determined; the safe lane change zone should meet the requirements of vehicle a. Maintain a sufficient safe distance from surrounding vehicles (a1, a2, a3) in both the X and Y directions, i.e. ;

[0089] Lane change trajectory planning: A fifth-order polynomial spline curve interpolation algorithm is used to plan the lane change trajectory; let the displacements of the lane change trajectory in the X and Y directions be x(t) and y(t), respectively, and their expressions are:

[0090] x(t) = c0 + c1t + c2t 2 +c3t 3 +c4t 4 +c5t 5 ;

[0091] y(t) = b0 + b1t + b2t 2 +b3t 3 +b4t 4 +b5t 5 ;

[0092] Solve for the polynomial coefficients c based on the initial and objective conditions. j and b j (j=0, 1, ..., 5), initial conditions include current position, velocity, and acceleration, target conditions include target lane position, velocity, and acceleration;

[0093] Parameter calculation: Based on the planned lane change trajectory, calculate the steering wheel parameters and gear shift parameters. The steering wheel parameters include the angle δ(t) and the steering speed. The transmission parameter is acceleration A(t); the relationship between the steering wheel angle and the vehicle's turning radius r is... Where L is the vehicle wheelbase; the steering speed is calculated based on the rate of change of the curvature of the trajectory; the gear shift parameters are adjusted according to the slope change of the trajectory and the target speed requirement.

[0094] c j (j=0, 1, ..., 5): These are the polynomial coefficients of the displacement function x(t) in the X direction, which determine the change of the vehicle's position in the X direction with time t. Specifically, c0 is the initial position offset in the X direction, c1 is related to the initial velocity in the X direction, c2 is related to the initial acceleration in the X direction, while c3, c4, and c5 affect more complex motion characteristics in the X direction, such as jerk.

[0095] b j(j=0, 1, ..., 5): These are the polynomial coefficients of the displacement function y(t) in the Y direction, which determine the change of the vehicle's position in the Y direction with time t; b0 is the initial position offset in the Y direction, b1 is related to the initial velocity in the Y direction, b2 is related to the initial acceleration in the Y direction, and b3, b4, and b5 also affect more complex motion characteristics in the Y direction.

[0096] How to obtain:

[0097] (1) Solve for the polynomial coefficients c based on the initial conditions (current position, velocity, acceleration) and the target conditions (target lane position, velocity, acceleration). j and b j (j=0, 1, ..., 5); the specific steps are as follows:

[0098] (2) Determine the boundary conditions, including the initial conditions and the objective conditions:

[0099] Initial conditions:

[0100] At t=0, for x(t), we have x(0)=c0, which represents the initial position of the vehicle in the X direction; x′(0)=c1, where x′(t) is the first derivative of x(t) with respect to t, representing the initial velocity in the X direction; x′′(0)=2c2, where x′′(t) is the second derivative of x(t) with respect to t, representing the initial acceleration in the X direction;

[0101] Similarly for y(t), y(0) = b0, y′(0) = b1, y′′(0) = 2b2;

[0102] Target conditions: Let the total time of the lane change process be T. At time t=T, x(T) should be equal to the position of the target lane in the X direction, x′(T) should be equal to the velocity of the target lane in the X direction, and x′′(T) should be equal to the acceleration of the target lane in the X direction; for y(t), y(T) should be equal to the position of the target lane in the Y direction, y′(T) should be equal to the velocity of the target lane in the Y direction, and y′′(T) should be equal to the acceleration of the target lane in the Y direction.

[0103] (3) Establish a system of equations:

[0104] For x(t), substituting the above boundary conditions into the expressions for x(t) and its first and second derivatives, we can obtain the expression for c. j Six equations (j=0, 1, ..., 5);

[0105] For y(t), similarly substituting the boundary conditions into the expressions for y(t) and its first and second derivatives, we obtain the expression for b. j Six equations (j=0, 1, ..., 5);

[0106] (4) Solving the system of equations: By solving these linear equations, the polynomial coefficients c in the X and Y directions can be obtained. j and b j (j=0, 1, ..., 5).

[0107] By identifying, detecting, and marking the vehicle itself and multiple key vehicles ahead and behind, data support is provided for subsequent calculations. Considering the impact of different road conditions on braking distance, a correction coefficient is used to adjust the safe distance threshold, making the judgment of safe distance for lane changes more accurate. By monitoring braking conditions and predicting risks, accidents caused by risks such as the braking of the vehicle in front can be avoided during lane changes. Based on the risk assessment results, a safe zone is determined, and a fifth-order polynomial spline curve interpolation algorithm is used to plan the trajectory, and steering wheel and transmission parameters are calculated, making the lane change process smoother and safer. Overall, these improvements enhance the system's real-time performance, accuracy, and adaptability, enabling it to better cope with complex and changing road environments and ensuring the safety and comfort of autonomous driving lane changes.

[0108] The third implementation method differs from the first implementation method in that the specific working content of the lane change process monitoring and dynamic adjustment module includes:

[0109] Vehicle parameter collection: During lane change, continuously collect the vehicle's steering wheel angle δ0(t), vehicle speed v0(t), acceleration A0(t), and vehicle attitude parameters; vehicle attitude parameters include vibration amplitude d(t) and drift distance D. drift (t), the shaking amplitude is obtained by measuring the acceleration sensor, and the drift distance is calculated by the deviation between the vehicle position and the planned trajectory;

[0110] Risk Prediction and Adjustment: Using collected vehicle parameters and real-time environmental information, predict the risk of vehicle lane changes; let the risk prediction function be:

[0111] R dynamic =w5A0(t)+w6D drift (t)+w7∣v0(t)−v target (t)∣;

[0112] Where w5, w6, and w7 are weighting coefficients, v target (t) represents the target vehicle speed; the dynamic risk threshold is set to R. th If R dynamic >R th If the lane change is not possible, immediately stop changing lanes or adjust the lane change path.

[0113] When adjusting the lane change path, a new safety distance analysis, trajectory planning, and parameter calculation are performed. The Model Predictive Control (MPC) algorithm is used to optimize the control input for a future period of time, including steering wheel angle and acceleration, based on the current vehicle status and environmental information in each control cycle, so that the vehicle can complete the lane change safely and smoothly.

[0114] The lane change monitoring and dynamic adjustment module continuously collects key vehicle parameters and vehicle attitude parameters, enabling a comprehensive understanding of the vehicle's real-time status during lane changes. By combining these parameters with real-time environmental information for risk prediction, lane change risks can be assessed more accurately. When the risk exceeds a threshold, the module can promptly stop the lane change or adjust the path, effectively preventing accidents. When adjusting the lane change path, a series of analyses and plans are re-performed, and model predictive control algorithms are used to optimize the control input, allowing the vehicle to flexibly adjust according to actual conditions, ensuring a safe and smooth lane change and improving the safety and reliability of autonomous driving lane changes.

[0115] The fourth implementation method differs from the first implementation method in that the specific working content of the abnormal approaching vehicle avoidance module includes:

[0116] Spacing recognition: Real-time recognition of the X and Y distances between vehicles a3 and a5 and this vehicle a0.

[0117] ;

[0118] Where, d 03X (t) represents the distance between vehicle a3 and this vehicle a0 in the X direction, d 05X (t) represents the distance between vehicle a5 and this vehicle a0 in the X direction, d 03Y (t) represents the distance between vehicle a3 and this vehicle a0 in the Y direction, d 05Y (t) represents the distance between vehicle a5 and this vehicle a0 in the Y direction;

[0119] Anomaly detection and avoidance planning: When vehicle a3 or a5 abnormally approaches vehicle a0, i.e., d 03X (t)<0 or d 05X (t)<0, and d 03Y (t) <d th or d 05Y (t) <d th d th For a given spacing threshold, such as 1m, plan the safe range within which movement to the other side is permitted; the safe range should meet the safe distance requirements from surrounding vehicles.

[0120] Based on the safe range, the Fast Random Search Tree (RRT) algorithm is used to plan the avoidance trajectory and control the vehicle to move safely to the other side to avoid a collision.

[0121] The spacing recognition function can accurately and in real-time obtain the distance between vehicles in the left and right lanes of the vehicle in the X and Y directions. X-space recognition can determine whether the vehicles on the left and right overlap with the vehicle in the width direction. Further analysis in the Y and Y directions can quickly identify whether there is any abnormal approaching behavior from vehicles on the left and right, and plan a safe range in a timely manner. The Fast Random Search Tree (RRT) algorithm is used to plan the avoidance trajectory, which can efficiently find a safe movement path and control the vehicle to move safely to the other side, effectively avoiding collisions. This module enhances the autonomous driving system's ability to cope with sudden dangerous situations, improves driving safety, protects the safety of occupants and the vehicle, and enhances the reliability and practicality of autonomous driving.

[0122] Furthermore, any content not described in detail in this specification is existing technology known to those skilled in the art.

[0123] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0124] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An autonomous driving path optimization control system integrating environmental perception and prediction, characterized in that, Let the road length direction be the X direction and the road width direction be the Y direction. Let the vehicle be a0, the vehicle in front of vehicle a0 in the same lane be a1, the vehicle in front of vehicle a0 in the lane where vehicle a0 is waiting to change lanes be a2, the vehicle behind vehicle a2 be a3, the vehicle in front of vehicle a2 be a4, and the vehicle approaching vehicle a0 in the other lane be a5. The system includes: Environmental perception and data acquisition module: This module acquires real-time position, speed, and acceleration information of the vehicle and surrounding vehicles using onboard sensors. It identifies vehicles using image recognition algorithms, tracks motion using Kalman filtering, and monitors taillight status through image processing to determine braking conditions. Data acquisition includes: real-time acquisition of the position coordinates (x, y, y) of the vehicle (a0) and surrounding vehicles (a1, a2, a3, a4, a5) using onboard cameras or radar sensors. i y i ), velocity v i Acceleration A i Direction angle θ i Information, where i = (0, 1, 2, 3, 4, 5); at the same time, obtain the dimensions of this vehicle a0, vehicle speed v0, steering wheel angle δ0, and gear shifting parameters, where the dimensions include length L0 and width W0, and the gear shifting state is represented by acceleration A0; Safe Distance and Lane Change Space Analysis Module: Set safe distance thresholds, build a three-dimensional spatial model to analyze lane change space, calculate critical distances and predict the risk of deceleration of the vehicle in front; Braking Monitoring and Risk Prediction Module: This module counts the number of times surrounding vehicles brake, calculates braking frequency, combines multiple factors, and uses a logistic regression model to predict braking probability, comprehensively assessing the risk level of lane changes; specifically including: Braking frequency calculation: Count the number of times vehicles a1 and a2 brake within time window T, N. a1 and N a2 Calculate the braking frequencies of vehicles a1 and a2. and ; Braking probability prediction: combined with braking frequency f ai Vehicle speed v i Acceleration A i In addition to the current traffic flow factor F, a logistic regression model is used to predict the probability P of vehicles a1 and a2 braking during the lane change of vehicle a0. a1 and P a2 ; Lane change risk assessment: Based on the braking probability, the deceleration risk of vehicle a2, and the lane change space analysis results, the lane change risk level R is comprehensively assessed; let the risk assessment function be: ； Where w1, w2, w3, and w4 are weighting coefficients, and I(·) is an indicator function that takes the value 1 when the condition is met and 0 otherwise; d 24X Let a2 be the distance in the X direction between vehicle a2 and the vehicle a4 in front of it. For d 24X rate of change, v th As the velocity threshold, D safe For safe distance, lane change risk is divided into three levels: low, medium, and high based on the value of R; Safe zone determination and lane change trajectory planning module: Determines safe lane change areas based on risk assessment results, plans lane change trajectories, and then calculates steering wheel parameters and gear shift parameters for lane change control. Lane change process monitoring and dynamic adjustment module: continuously collects vehicle parameters during the lane change process, combines real-time environmental information to predict lane change risks, and immediately stops the lane change or adjusts the lane change path if the risk exceeds the threshold. The control input is optimized by using model predictive control algorithm. Abnormal Approach Vehicle Avoidance Module: Real-time identification of abnormally approaching vehicles, planning of a safe range that can be moved to the other side, and using a fast random search tree algorithm to plan the avoidance trajectory.

2. The autonomous driving path optimization control system integrating environmental perception and prediction according to claim 1, characterized in that: The vehicle status recognition of the environmental perception and data acquisition module includes: using image recognition algorithms combined with sensor data to accurately identify each vehicle, marking the vehicles with numbers, and using Kalman filtering algorithms to continuously track their motion state and estimate the position and speed of the vehicles in the future. The taillight status of vehicles a1 and a2 is monitored in real time using image processing technology. If the taillight brightness exceeds a set threshold I... th If both taillights are lit simultaneously, it indicates that the vehicle is in a braking state.

3. The autonomous driving path optimization control system integrating environmental perception and prediction according to claim 1, characterized in that: The specific functions of the safety distance and lane change space analysis module include: Establish a spatial model: Based on the dimensions and speed of vehicle a0, set the safety distance threshold required for lane changing; let the safety distance D be... safe The relationship with vehicle speed v0 is D safe =k·v0+b, where k and b are constants, obtained by fitting experimental data; Using vehicle a0 as the center and combining the position information of vehicles a1, a2, and a3, a three-dimensional spatial model is established. On the XY plane, with the center of vehicle a0 as the origin, the positions of vehicles a1, a2, and a3 constitute spatial points. It is then determined whether vehicle a0 meets the safety distance requirements with surrounding vehicles in the X and Y directions. ； Calculating critical distances and predicting deceleration risks: Real-time calculation of the distance d between vehicle a0 and vehicle a2 in the X direction. 02X =|x0-x2|, and the X-direction distance d between vehicle a2 and the vehicle a4 in front of it. 24X = |x2-x4|; Monitoring d 24X rate of change Let the velocity threshold be v th ,like If so, it is predicted that vehicle a2 is at risk of deceleration.

4. The autonomous driving path optimization control system integrating environmental perception and prediction according to claim 3, characterized in that: The specific functions of the safe zone determination and lane change trajectory planning module include: Safe zone determination: When the lane change risk assessment is low or medium, a safe lane change zone is determined; Lane change trajectory planning: A fifth-order polynomial spline curve interpolation algorithm is used to plan the lane change trajectory; let the displacements of the lane change trajectory in the X and Y directions be x(t) and y(t), respectively, and their expressions are: x(t)=c0+c1t+c2t 2 +c3t 3 +c4t 4 +c5t 5 ; y(t)=b0+b1t+b2t 2 +b3t 3 +b4t 4 +b5t 5 ; Solve for the polynomial coefficients c based on the initial and objective conditions. j and b j (j=0, 1, ..., 5), initial conditions include current position, velocity, and acceleration, target conditions include target lane position, velocity, and acceleration; Parameter calculation: Based on the planned lane change trajectory, calculate the steering wheel parameters and gear shift parameters. The steering wheel parameters include the angle δ(t) and the steering speed. The transmission parameter is acceleration A(t); the relationship between the steering wheel angle and the vehicle's turning radius r is... Where L is the vehicle wheelbase; the steering speed is calculated based on the rate of change of the curvature of the trajectory; the gear shift parameters are adjusted according to the slope change of the trajectory and the target speed requirement.

5. The autonomous driving path optimization control system integrating environmental perception and prediction according to claim 1, characterized in that: The specific functions of the lane change process monitoring and dynamic adjustment module include: Vehicle parameter collection: During lane change, continuously collect the vehicle's steering wheel angle δ0(t), vehicle speed v0(t), acceleration A0(t), and vehicle attitude parameters; vehicle attitude parameters include vibration amplitude d(t) and drift distance D. drift (t), the shaking amplitude is obtained by measuring the acceleration sensor, and the drift distance is calculated by the deviation between the vehicle position and the planned trajectory; Risk Prediction and Adjustment: Using collected vehicle parameters and real-time environmental information, predict the risk of vehicle lane changes; let the risk prediction function be: ； Where w5, w6, and w7 are weighting coefficients, v target (t) represents the target vehicle speed; the dynamic risk threshold is set to R. th If R dynamic >R th If the lane change is not possible, immediately stop changing lanes or adjust the lane change path. When adjusting the lane change path, a new safety distance analysis, trajectory planning, and parameter calculation are performed. A model predictive control algorithm is used to optimize the control input for a future period of time, including steering wheel angle and acceleration, based on the current vehicle status and environmental information in each control cycle, so that the vehicle can complete the lane change safely and smoothly.

6. The autonomous driving path optimization control system integrating environmental perception and prediction according to claim 3, characterized in that: The safety distance D safe During the calculation process, the influence of different road conditions on vehicle braking distance is considered. The k and b values ​​for each vehicle type are corrected, and a road condition influence coefficient table is established. According to different road conditions, the k and b values ​​are adjusted by corresponding multiplication or addition.

Images