Intelligent vehicle planning control system and method in uncertain environment

Abstract

The application relates to an intelligent automobile planning control system and method in an uncertain environment, which comprises a local trajectory planning module and a tracking control module connected therewith; the local trajectory planning module constructs a quantitative expression of a blind area in a FOV according to obstacle information, constructs a multi-target optimization problem by considering a perception result factor, adjusts an initial reference trajectory, outputs a new trajectory, and transmits the new trajectory to the tracking control module; the tracking control module outputs corresponding control signals to a vehicle actuator according to the new trajectory, and controls a driving trajectory of the vehicle. Compared with the prior art, the application introduces information entropy in automatic driving trajectory planning, uses blind area information to construct an information entropy optimization item in real time, and adds an optimization target of a trajectory planning layer to realize active perception. By making the perception result act on a lower layer planning control, the application reduces the perception uncertainty in the blind area, reduces the collision risk, avoids a conservative obstacle avoidance strategy, and effectively improves the driving safety of the vehicle.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent vehicle planning control, in particular to an intelligent vehicle planning control system and method in uncertain environment. BACKGROUND

[0002] In recent years, self-driving vehicles have become a research hotspot because they can solve social problems such as road congestion and traffic accidents. The typical self-driving vehicle system architecture is mainly divided into three parts: an environment perception system, a decision planning system, and a motion control system. The decision planning system outputs a safe and collision-free optimal expected driving trajectory based on perception and positioning information, and the motion control system takes the expected trajectory as input for trajectory tracking to obtain the control amount acting on the bottom layer to complete automatic driving. The trajectory planning of self-driving vehicles is mostly derived from the research in the field of mobile robots and unmanned aerial vehicles, which is defined as generating a path connecting the initial position of the vehicle to the target position, and the corresponding speed at each point on the path, while requiring the vehicle to meet the kinematic or dynamic constraints, collision constraints, and other time and space constraints derived from the system itself or the external environment during the movement along the path containing speed information.

[0003] In the traditional self-driving vehicle architecture system, perception, decision planning, and control belong to three different modules. With the in-depth study of self-driving technology, researchers have found the advantages and necessity of combining trajectory planning and tracking control, and a unified control method of trajectory planning and tracking has emerged. Considering the characteristics of trajectory planning and trajectory tracking, MPC (Model Predictive Control) has become the most suitable algorithm for designing a planning control combined framework due to its great advantages in using prediction information and considering multiple constraints.

[0004] There are objective, extensive, and real uncertainties in the traffic environment in which self-driving vehicles travel. It can be said that the existence of uncertainties from all aspects is inevitable. Uncertainties will affect the risk assessment, decision-making, and trajectory planning of self-driving vehicles, thereby bringing great challenges to the safety, reliability, and comfort of self-driving vehicles. Traditional trajectory planning algorithms are difficult to effectively cope with, so recently, decision planning algorithms for self-driving uncertainties have been proposed. In summary, existing research mainly targets perception uncertainty and prediction uncertainty. The former includes uncertainty in the position of other vehicles, and the latter involves the interaction between the ego vehicle and other vehicles. However, most of the above research can be considered as passive avoidance of risks caused by external uncertainties, i.e., it belongs to a conservative trajectory planning control method. There are few studies aimed at actively reducing uncertainties and improving front-end perception, making it difficult to ensure the scientificity, rationality, and comfort of path planning. SUMMARY

[0005] The purpose of the present application is to overcome the defects of the prior art, and provide an intelligent vehicle planning control system and method in an uncertain environment, which can actively explore the blocked area, obtain relevant environmental information required for vehicle path planning, reduce risks, and make more scientific, reasonable and comfortable path planning.

[0006] The purpose of the present application can be achieved by the following technical solutions: an intelligent vehicle planning control system in an uncertain environment, comprising a local trajectory planning module and a tracking control module connected thereto, the local trajectory planning module being connected to a vehicle perception module, constructing a quantitative expression of a blind area within a FOV (Field of View) according to obstacle information output by the vehicle perception module, and constructing a multi-objective optimization problem considering the perception result factor to adjust the initial reference trajectory, outputting a new trajectory and transmitting it to the tracking control module; the tracking control module outputs corresponding control signals to the vehicle actuator according to the new trajectory to control the driving trajectory of the vehicle.

[0007] An intelligent vehicle planning control method in an uncertain environment, comprising the following steps:

[0008] S1, obtaining environmental and obstacle information, and constructing an information entropy factor;

[0009] S2, combining the information entropy factor and the current speed, heading angle and position of the vehicle to construct a trajectory planning objective function, and solving to obtain a new trajectory;

[0010] S3, tracking control of the new trajectory, and outputting corresponding control signals;

[0011] S4, controlling the driving trajectory of the vehicle according to the control signals.

[0012] Further, the information entropy factor includes: FOV blocked area; unperceived area in the fixed area behind the obstacle; and unperceived area in each actual blind area.

[0013] Further, the step S2 specifically uses a vehicle kinematics model, and considers vehicle dynamics constraints and lateral acceleration constraints to construct the trajectory planning objective function.

[0014] Further, the vehicle kinematics model is specifically:

[0015]

[0016] wherein, are the lateral and longitudinal accelerations of the vehicle in the vehicle coordinate system, the vehicle longitudinal acceleration is zero, is the heading angle change rate, and respectively are the longitudinal and lateral velocities of the vehicle in the inertial coordinate system;

[0017] The trajectory planning objective function is specifically:

[0018]

[0019]

[0020] U min ≤U t ≤U max

[0021] |U t |=|a t |<μg

[0022] wherein J obs,i is the obstacle avoidance function at sampling time i, J IE,i is the information entropy factor at sampling time i, U max and U min are the upper and lower limits of the control variable U t , μ is the adhesion coefficient, g is the gravitational acceleration, Q and R are respectively the semi-positive definite weight matrices of the system state vector and the control input vector.

[0023] Further, when the area of the FOV blocked region is taken as the optimization item, the information entropy factor is specifically:

[0024] J IE,i =S IE S obscured,i

[0025] wherein S IE is the weight coefficient, S obscured,i is the area of the FOV blocked region caused by the obstacle at the current position of the ego vehicle.

[0026] Further, when the area of the obstacle behind the fixed region that is not perceived is taken as the optimization item, the information entropy factor is specifically:

[0027]

[0028] wherein S IE is the weight coefficient, S inttal is the area of the fixed region that is not perceived after the ego vehicle actually executes one step of the control instruction based on the last step, and S obscured,i is the area of the fixed region that is not perceived in each step of the prediction step.

[0029] Further, when the area of the actual blind area in each step that is not perceived is taken as the optimization item, the information entropy factor is specifically:

[0030]

[0031] wherein S IE is a weight coefficient, S inital is the initial blind area of the ego vehicle after each actual execution of a control instruction, S obscured,i is the area of the initial blind area that is not yet perceived in each step of the prediction step.

[0032] Further, the step S3 specifically uses a vehicle dynamics nonlinear model based on a smaller front wheel side slip angle and a linear tire model assumption, and takes the lateral and longitudinal speeds, the heading angle and the heading angle change rate of the vehicle in the ego vehicle coordinate system, and the lateral and longitudinal coordinates of the vehicle in the earth coordinate system as state variables, and takes the front wheel steering angle as a control variable, to construct a trajectory tracking optimization objective function, and then solve the control signal corresponding to the new trajectory.

[0033] Further, the vehicle dynamics nonlinear model is specifically:

[0034]

[0035] wherein C lf and C lr are the longitudinal side stiffness of the front and rear wheels of the vehicle, respectively, is a state variable, are the lateral and longitudinal speeds of the vehicle in the ego vehicle coordinate system, Y and X are the lateral and longitudinal coordinates of the vehicle in the inertial coordinate system, and the control variable is u dyn = δ f ;

[0036] The trajectory tracking optimization objective function is specifically:

[0037]

[0038] s.t.ξ dyn,k+1 = A dyn,k+1 ξ dyn,k + B dyn,k u dyn,k

[0039] ΔU dyn,min ≤ ΔU dyn,t ≤ ΔU dyn,max

[0040] U dyn,min ≤ AΔU dyn,t + U dyn,t ≤ U dyn,max

[0041] y hc,min ≤ y hc ≤ y hc,max

[0042] y sc,min -ε≤y sc ≤y sc,max +ε

[0043] ε>0

[0044] wherein, is the local reference trajectory, i.e., the new trajectory, ε is the relaxation factor, ΔU dyn,t is the control variable sequence, AΔU dyn,t +U dyn,t is the control variable, constraints are imposed on the control variable and the control variable sequence, y hc is the hard constraint output, y sc is the soft constraint output.

[0045] Compared with the prior art, the present application constructs a local trajectory planning module and a tracking control module connected thereto, uses the local trajectory planning module to construct a quantitative expression of the blind area within the FOV according to the obstacle information output by the vehicle perception module, constructs a multi-objective optimization problem considering the perception result factor to adjust the initial reference trajectory, and outputs a new trajectory and transmits it to the tracking control module; the tracking control module uses the new trajectory to output a corresponding control signal to the vehicle actuator, thereby controlling the driving trajectory of the vehicle. Thus, the entire automatic driving process is linked by perception decision and planning control, so that the upper layer perception affects the lower layer path planning, and acts on the perception, can actively explore the blocked area, obtain the relevant environmental information required for the ego path planning, reduce the risk, and ensure the scientificity, rationality and comfort of the path planning.

[0046] In order to describe the environmental uncertainty information in the blind area, the present application introduces the concept of information entropy in trajectory planning, constructs an information entropy optimization term according to the blind area information within the FOV in the actual driving condition in real time, and adds it to the optimization target of trajectory planning. Compared with the existing conservative trajectory planning control algorithm, the present application reduces the range of the blind area caused by the obstacle blocking by active perception on the basis of realizing obstacle avoidance when facing potential risks in the driving scene, reduces the environmental uncertainty and potential risks caused by the blind area, avoids passive obstacle avoidance strategy, and improves safety.

[0047] The information entropy factor of the application includes a FOV blocked area, an area not perceived in a fixed area behind an obstacle, and an area not perceived in an actual blind area of each step, can actively and effectively explore the blind area behind the obstacle, and increase the final perceived area in the initial blind area. At the same time, it can avoid too conservative driving strategy, improve time efficiency, effectively find the time of potential obstacles in the blind area, and increase the reaction distance. After the information entropy is considered, the overall perception efficiency can be greatly improved, and strong support is provided for making more positive and active decision planning subsequently. The overall effect makes the behavior of the unmanned vehicle to cope with environmental uncertainty more humanized, close to the driving observation habits of human drivers. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 It is a schematic diagram of the system structure of the application;

[0049] Figure 2 It is a schematic diagram of the method flow of the application;

[0050] Figure 3 It is a schematic diagram of the application framework in the embodiment;

[0051] Figure 4a It is a schematic diagram of the FOV blocked area;

[0052] Figure 4b It is a schematic diagram of the area not perceived in the fixed area behind the obstacle;

[0053] Figure 4c It is a schematic diagram of the area not perceived in the actual blind area of each step;

[0054] Figure 5 It is a schematic diagram of scene one in the embodiment;

[0055] Figure 6 It is a schematic diagram of scene two in the embodiment;

[0056] Figure 7 It is a schematic diagram of the visibility index in the embodiment;

[0057] Figure 8 It is a schematic diagram of the safety index in the embodiment;

[0058] Figure 9 It is a schematic diagram of the comparison of the planning trajectory before and after considering the information entropy in the case of scene one in the embodiment;

[0059] Figure 10 It is a schematic diagram of the comparison of the planning trajectory before and after considering the information entropy in the case of scene two in the embodiment;

[0060] Marking description in the figure: 1, vehicle perception module, 2, local trajectory planning module, 3, tracking control module, 4, vehicle execution mechanism. Detailed Implementation

[0061] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0062] Example

[0063] like Figure 1 As shown, an intelligent vehicle planning and control system under uncertain environment includes a local trajectory planning module 2 and a tracking control module 3 connected thereto. The local trajectory planning module 2 is connected to the vehicle perception module 1. Based on the obstacle information output by the vehicle perception module 1, it constructs a quantitative expression of the blind zone within the field of view (FOV) and considers the perception result factors to construct a multi-objective optimization problem to adjust the initial reference trajectory. The new trajectory is then output and transmitted to the tracking control module 3. The tracking control module 3 outputs corresponding control signals to the vehicle actuator 4 based on the new trajectory to control the vehicle's driving trajectory.

[0064] In this embodiment, both the upper-level local trajectory planning module and the lower-level tracking control module are constructed based on the Model Predictive Control (MPC) algorithm. In practical applications, they can also be constructed based on other optimization methods and are not limited to the MPC algorithm. This embodiment takes into account the great advantages of the MPC algorithm in terms of predictive information, rolling optimization, and consideration of multiple constraints, which can fully guarantee the timeliness and reliability of the calculation results.

[0065] Based on the above system, a planning and control method for intelligent vehicles under uncertain environments is implemented, such as... Figure 2 As shown, it includes the following steps:

[0066] S1. Obtain environmental and obstacle information and construct information entropy factor;

[0067] S2. Combining the information entropy factor with the vehicle's current speed, heading angle, and position, construct the trajectory planning objective function and solve for the new trajectory.

[0068] S3. Track and control the new trajectory, and output the corresponding control signal;

[0069] S4. Control the vehicle's trajectory according to the control signal.

[0070] This embodiment applies the above technical solution, and the specific application framework is as follows: Figure 3 As shown, step S2 is executed by the trajectory planning module, specifically using a vehicle kinematics model and considering vehicle dynamics constraints and lateral acceleration constraints to construct the trajectory planning objective function. The vehicle kinematics model is as follows:

[0071]

[0072] In the formula, respectively, are the lateral and longitudinal acceleration of the vehicle in the ego coordinate system, and the longitudinal acceleration of the vehicle is zero, is the rate of change of the heading angle, and respectively, are the longitudinal and lateral velocity of the vehicle in the inertial coordinate system;

[0073] The state variables are the vehicle speed in the y and x direction of the ego coordinate system, the vehicle heading angle, the longitudinal and lateral coordinates of the vehicle position in the geodetic coordinate system. The state variable is the lateral acceleration a y . The above formula is simplified as Considering the dynamics constraints of the vehicle, and since the longitudinal acceleration of the vehicle model is zero, the constraint condition |a y | < μg is also added. The trajectory planning objective function is constructed as:

[0074]

[0075]

[0076] U min ≤ U t ≤ U max

[0077] | U t | = |a y | < μg

[0078] In the formula, J obs,i is the obstacle avoidance function at sampling time i, J IE,i is the information entropy factor at sampling time i, U max and U min are the upper and lower limits of the control variable U t , μ is the adhesion coefficient, g is the gravitational acceleration, Q and R are the semi-positive definite weight matrices of the system state vector and the control input vector respectively.

[0079] The information entropy factor in the technical solution includes: the area of the FOV blocked region; the area not perceived in the fixed region behind the obstacle; the area not perceived in the actual blind area of each step.

[0080] 1) Take the area of the FOV blocked region as the optimization item

[0081] The goal of considering information entropy is to reduce the area of the FOV unobservable region as soon as possible, so as to reduce the risk of uncertainty caused by blocking. Therefore, the most direct idea is to take the FOV blocked area as the optimization item. The information entropy function in this form is as follows:

[0082] J IE,i = S IE S obscured,i

[0083] where S IE is a weight coefficient, S obscured,i is the area of the FOV occluded region caused by the existence of the obstacle at the current position of the ego vehicle, as shown in Figure 4a .

[0084] 2) Taking the area of the fixed region behind the obstacle that is not perceived as the optimization item

[0085] In order to make the ego vehicle plan a trajectory that focuses more on reducing the occluded region behind the obstacle, as shown in Figure 4b , once the critical point of the obstacle and the occluded region of the FOV are detected, the information entropy of the fixed rectangular space behind the obstacle corner point is calculated. The information entropy function is as follows:

[0086]

[0087] where S IE is a weight coefficient, S inital is the area of the fixed region updated on the basis of the previous step after the ego vehicle actually executes a control instruction, and S obscured,i is the area of the fixed region that has not been perceived in each step of the prediction step.

[0088] 3) Taking the area of the actual blind area in each step that is not perceived as the optimization item

[0089] This form of information entropy function can be regarded as a combination of the previous two ideas. The information entropy function expression is as follows:

[0090]

[0091] where S IE is a weight coefficient, S inital is the initial blind area of the ego vehicle after actually executing a control instruction in each step, as shown in Figure 4c , S obscured,i is the area of the initial blind area that has not been perceived in each step of the prediction step, as shown in Figure 4c .

[0092] The process of step S3 is performed by the trajectory tracking control module. Specifically, based on the assumption of a smaller front wheel side slip angle and a linear tire model, the vehicle dynamics nonlinear model is used, and the lateral and longitudinal speed, heading angle and heading angle rate of change of the vehicle in the ego vehicle coordinate system, and the lateral and longitudinal coordinates of the vehicle in the earth coordinate system are taken as state variables, and the front wheel steering angle is taken as the control variable, to construct a trajectory tracking optimization objective function, and then solve the control signal corresponding to the new trajectory.

[0093] where the vehicle dynamics nonlinear model is specifically:

[0094]

[0095] wherein C lf and C lr are the longitudinal cornering stiffness of the front and rear wheels of the vehicle respectively, is the state variable, are the lateral and longitudinal velocities of the vehicle in the body frame, Y and X are the lateral and longitudinal coordinates of the vehicle in the inertial frame, and u dyn = δ f ;

[0096] The trajectory tracking optimization objective function is specifically:

[0097]

[0098] s.t.ξ dyn,k+1 =A dyn,k+1 ξ dyn,k +B dyn,k u dyn,k

[0099] ΔU dyn,min ≤ΔU dyn,t ≤ΔU dyn,max

[0100] U dyn,min ≤AΔU dyn,t +U dyn,t ≤U dyn,max

[0101] y hc,min ≤y hc ≤y hc,amx

[0102] y sc,min -ε≤y sc ≤y sc,max +ε

[0103] ε>0

[0104] wherein, is the local reference trajectory, i.e., the new trajectory, ε is a relaxation factor, ΔU dyn,t is the control variable sequence, AΔU dyn,t +U dyn,t is the control variable, and constraints are imposed on the control variable and the control variable sequence, y hc is the hard constraint output, and y sc is the soft constraint output.

[0105] To verify the effectiveness of the technical solution, the embodiment is used in Figure 5 and Figure 6Two typical traffic scenarios are tested and compared, where scenario one is the ego vehicle overtaking a bus which is stopping at a bus stop, and there is a potential pedestrian walking out from behind the bus; scenario two is the cornering situation in an unstructured road such as a garage or a community, and there is a potential vehicle or other obstacle near the original reference trajectory in the blind area, the initial reference trajectory is shown as a dashed line in front of the vehicle in the figure.

[0106] The embodiment assumes that the FOV complete area is an isosceles triangle with a fixed view angle of 90°. The test is carried out in a CarSim and Simulink joint simulation environment, and the evaluation indexes in the test include:

[0107] 1) Visibility index. Take a rectangular evaluation area behind the obstacle as shown in Figure 7 , because in actual situations, the potential danger in this area has a greater impact on the ego vehicle. The visibility index is:

[0108]

[0109] Where S is the initial area of each evaluation area, S vis is the final perceived area in the evaluation area;

[0110] 2) Safety index. For different traffic scenarios, take a point in the blind area behind the obstacle as shown in Figure 8 , assuming it is the coordinate position of another potential obstacle, and compare the time t found when the point is perceived with the distance d found .

[0111] When the information entropy is not considered and only the obstacle avoidance function is considered, the visibility indexes p vis in the two scenarios are 72.56% and 29.43%, respectively. When the unperceived area in each step in the actual blind area is taken as the optimization item, the simulation results are shown in Tables 1 and 2. As can be seen from Tables 1 and 2, increasing S IE can make the ego vehicle trajectory away from the obstacle corner point, and introducing information entropy makes the exploration effect of the ego vehicle in the blind area more obvious.

[0112] Table 1 p vis of different weight combinations in scenario one

[0113]

[0114] Table 2 p vis of different weight combinations in scenario two

[0115]

[0116] When only the obstacle avoidance function is considered, the evaluation index t found ≈6.7s, d found≈4.24m, under scenario two t found ≈5.4s, d found ≈2.39m. After adding information entropy, the results are shown in Table 3 and Table 4, as shown in Table 3, in scenario one, although the ego vehicle cannot observe the obstacle point in advance in time, the observed position is greatly increased from the obstacle point, which is very important for the ego vehicle to make decision planning adjustment in actual situation; in Table 4, it can be seen that after the planning problem considers the information entropy, the ego vehicle can observe the obstacle point in advance in time, increase the reaction distance, and effectively avoid if the obstacle point exists real object. found And d found All have an impact on early observation time and increased reaction distance.

[0117] Table 3 S obs =100 under scenario one IE Comparison

[0118] [SA IE ]] 4000 40000 400000 t found / s]] 6.801 6.701 6.701 d found / m]] 11.6738 10.8883 10.9445

[0119] Table 4 S obs =100 under scenario two IE Comparison

[0120] [SA IE ]] 4000 40000 400000 t found / s]]> 5.201 5.101 5.101 d found / m]] 6.527 6.9381 7.1501

[0121] From Figure 9 and Figure 10 It can be seen that after considering information entropy for trajectory planning control, the vehicle can obviously achieve better active obstacle avoidance under scenario one and scenario two, thereby reducing the environmental uncertainty and potential risk brought by the blind area and effectively improving safety.

[0122] As can be seen from the above, in the technical solution, the upper trajectory planning module adjusts the initial reference trajectory according to the obstacle information and the blind area in the FOV to obtain a new trajectory sent to the tracking control layer, and outputs a control signal. In order to describe the environmental uncertainty information in the blind area, the technical solution introduces the concept of information entropy in the automatic driving trajectory planning, and adds it as an optimization item to the optimization objective of the trajectory planning layer, so as to realize guiding the trajectory planning with the perception result. Compared with the existing planning control algorithm, the technical solution can reduce the blind area range caused by the obstacle shielding through active perception on the basis of realizing obstacle avoidance when facing the driving scene with obstacles, reduce the environmental uncertainty and potential risk brought by the blind area, avoid passive obstacle avoidance strategy, and effectively improve the vehicle driving safety.

Claims

1. A planning and control method for intelligent vehicles under uncertain environments, characterized in that, Includes the following steps: S1. Obtain environmental and obstacle information and construct information entropy factor; S2. Combining the information entropy factor with the vehicle's current speed, heading angle, and position, construct the trajectory planning objective function and solve for the new trajectory. S3. Track and control the new trajectory, and output the corresponding control signal; S4. Control the vehicle's trajectory according to the control signal; The information entropy factor includes: the area of ​​the FOV-obscured region; the unperceived area within a fixed region behind the obstacle; and the unperceived area within the actual blind zone at each step. When the area of ​​the occluded region within the field of view (FOV) is used as the optimization term, the information entropy factor is specifically: in, These are the weighting coefficients. The area of ​​the FOV (Field of View) obstructed by obstacles at the vehicle's current position.

2. The intelligent vehicle planning and control method under uncertain environments according to claim 1, characterized in that, Step S2 specifically uses a vehicle motion model and considers vehicle dynamics constraints and lateral acceleration constraints to construct a trajectory local planning objective function.

3. The intelligent vehicle planning and control method under uncertain environments according to claim 2, characterized in that, The vehicle kinematics model is specifically as follows: in, , These represent the lateral and longitudinal accelerations of the vehicle in its own coordinate system, respectively, with the longitudinal acceleration being zero. The rate of change of heading angle, and These represent the vehicle's longitudinal and lateral velocities in the inertial coordinate system, respectively. The objective function for the local trajectory planning is specifically: in, Let i be the obstacle avoidance function at sampling time i. Let i be the information entropy factor at sampling time i. To control the amount The upper and lower limits, The adhesion coefficient, Let be the gravitational acceleration, and Q and R be the positive semidefinite weight matrices of the system state vector and control input vector, respectively.

4. The intelligent vehicle planning and control method under uncertain environments according to claim 3, characterized in that, When the unperceived area within a fixed region behind an obstacle is used as the optimization term, the information entropy factor is specifically: This refers to the unperceived area within a fixed region updated based on the previous step after each actual control command executed by the vehicle. This is to predict the area that has not yet been perceived in the fixed region of each step length during the prediction step.

5. The intelligent vehicle planning and control method under uncertain environments according to claim 3, characterized in that, When the area of ​​the actual blind zone that is not perceived at each step is used as the optimization term, the information entropy factor is specifically: in, These are the weighting coefficients. The initial blind spot area after each actual control command executed by the vehicle. To predict the area in the initial blind zone that has not yet been perceived for each step length in the prediction step.

6. The intelligent vehicle planning and control method under uncertain environments according to claim 3, characterized in that, Step S3 specifically uses a nonlinear vehicle dynamics model based on the assumption of a small front wheel slip angle and a linear tire model. The vehicle's lateral and longitudinal velocities in the vehicle coordinate system, heading angle and rate of change of heading angle, and the vehicle's lateral and longitudinal coordinates in the geodetic coordinate system are taken as state variables, and the front wheel steering angle is taken as the control variable. In order to construct a trajectory tracking optimization objective function, the control signal corresponding to the new trajectory is then solved.

7. The intelligent vehicle planning and control method under uncertain environments according to claim 6, characterized in that, The specific nonlinear model of vehicle dynamics is as follows: in, and These are the longitudinal lateral stiffness of the front and rear wheels of the vehicle, respectively. For state variables, , Let Y and X be the lateral and longitudinal velocities of the vehicle in the vehicle's own coordinate system, respectively, and Y and X be the lateral and longitudinal coordinates of the vehicle in the inertial coordinate system, respectively. The control variable is... ; The objective function for trajectory tracking optimization is specifically as follows: in, This is a local reference trajectory, i.e., the new trajectory. As a relaxation factor, For the control variable sequence, To control the quantity, constraints are imposed on the control quantity and its variables. For hard-constrained output, This is a soft constraint output.

8. An intelligent vehicle planning and control system for uncertain environments, used to implement the intelligent vehicle planning and control method for uncertain environments as described in any one of claims 1 to 7, characterized in that, It includes a local trajectory planning module and a tracking control module connected thereto. The local trajectory planning module is connected to the vehicle perception module. Based on the obstacle information output by the vehicle perception module, it constructs a quantitative expression of the blind zone within the field of view (FOV) and considers the perception result factors to construct a multi-objective optimization problem to adjust the initial reference trajectory, outputs a new trajectory, and transmits it to the tracking control module. The tracking control module outputs corresponding control signals to the vehicle actuators based on the new trajectory, thereby controlling the vehicle's driving trajectory.

Images