A smart car anti-aiming and tracking method that takes driving modes into account

Abstract

This invention relates to the field of intelligent vehicle technology and provides an intelligent vehicle anti-tracking method that considers driving modes. The method includes two steps: offline design and online anti-tracking. In the offline design phase, driving templates are designed, optimization problems are constructed, value function sub-modules are designed, and traffic rules and maneuverability constraint modules are designed. In the online anti-tracking phase, driving mode weight allocation, driving task acquisition and driving template determination, motion anti-tracking, and re-anti-tracking judgment are performed. This method fully considers four different driving modes: comfort, energy efficiency, sportiness, and accuracy. It also designs an evaluation value function that can be used for optimization, allowing driving modes to switch autonomously or allocate weights autonomously. It also considers traffic rule constraints and vehicle maneuverability, ensuring both safety and drivability in the anti-tracking results. A finite multi-segment description of the driving task is used, resulting in low-frequency changes in speed and turning intensity.

Description

Technical Field

[0001] This invention belongs to the field of intelligent vehicle technology, and particularly relates to an intelligent vehicle anti-targeting and tracking method that takes into account driving modes. Background Technology

[0002] With the development of intelligent vehicles, safety-oriented driver assistance technologies are now widely used in mass-produced models, while fully autonomous vehicles are still in the demonstration phase. Currently, public acceptance of autonomous vehicles is low because the driving systems focus solely on safety, failing to adequately consider the driver's driving style, mode, skill level, and other driving habits. This results in a driving experience that doesn't match the driver's accustomed driving habits, thus reducing driver trust and acceptance. Motion planning, as a key aspect of intelligent vehicle operation, is a crucial factor in determining whether an intelligent vehicle truly embodies human-like characteristics.

[0003] Current research on motion planning, while ensuring safety, mostly focuses on simply reducing acceleration to improve comfort, increasing desired speed to improve traffic efficiency, or using fixed weights to achieve universal motion planning, without comprehensively considering driving modes. Drivers may adopt different driving attitudes in different driving scenarios. Classifying driving modes allows for free switching or autonomous weight allocation, thereby enabling motion planning for different driving modes and helping to increase driver trust and acceptance of the driving system's motion planning.

[0004] Therefore, this invention addresses the motion planning problem by employing a motion pre-aiming method, which fully considers driving modes while ensuring safety. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent vehicle anti-targeting and tracking method that takes into account driving modes, in order to solve the problems mentioned in the background art.

[0006] The present invention is implemented as follows: an intelligent vehicle anti-targeting and tracking method considering driving modes includes the following steps:

[0007] Step 1: Offline design;

[0008] Step 1.1: Design of the driving template;

[0009] Break down driving tasks into Each curvature motion primitive and A velocity motion primitive is used to generate a pose driving template through a curvature motion primitive, and a vehicle speed driving template is generated through a velocity motion primitive.

[0010] The pose driving template fits the desired driving trajectory using an arc or a straight line, with the goal of the fitted trajectory deviating from the desired trajectory by less than a threshold. The initial curvature is obtained using the bisection method. and initial arc length ;

[0011] The vehicle speed driving template fits the desired vehicle speed through uniform acceleration motion, with the goal of the deviation between the fitted vehicle speed and the desired vehicle speed being less than a threshold. The initial acceleration is obtained using the bisection method. and initial segment length .

[0012] Step 1.2: Optimize the problem formulation;

[0013] With curvature increment Acceleration increment Increment of arc length of curvature segment and the increment of the arc length of the acceleration segment To optimize the parameters, an optimization problem is constructed that includes pose and velocity schemes;

[0014] The pose scheme is composed of multiple segments of circular arcs or straight lines, each segment having the same curvature; the velocity scheme is composed of multiple segments of uniformly accelerated or uniformly accelerated motion, each segment having the same acceleration.

[0015] Step 1.3: Design of the value function submodule;

[0016] The design should include at least the following four value function sub-modules:

[0017] Comfort value function submodule: Based on the weighted root mean square of the driver's head acceleration, calculate the sum of squares of lateral and longitudinal accelerations during cornering and gear shifting.

[0018] Energy efficiency value function submodule: The total specific work squared of the vehicle's driving is used as the evaluation index, taking into account acceleration resistance, rolling resistance, wind resistance and slope resistance;

[0019] The motion performance value function submodule uses the sum of squared driving times as the evaluation index;

[0020] Accuracy Value Function Submodule: The lateral deviation and speed deviation between the pre-aiming result and the expected trajectory are used as evaluation indicators;

[0021] Step 1.4: Design of traffic rules and mobility constraint module;

[0022] Set constraints, including: traffic rule constraints, vehicle maneuverability constraints, and driving task target point constraints;

[0023] Traffic rule constraints include the upper and lower speed limits for vehicles, as well as left and right boundaries; vehicle maneuverability constraints include the upper limit of acceleration that varies with vehicle speed. Upper limit of lateral acceleration The driving task target point constraints include position and speed requirements.

[0024] Step 2: Online pre-aiming;

[0025] Step 2.1: Driving mode weight allocation;

[0026] Predefined driving modes; select a driving mode according to the driving mode command, that is, weight allocation to precision, comfort, energy saving and sport mode, or manually allocate driving mode weight according to the driver's preference;

[0027] Step 2.2: Obtain driving tasks and determine driving templates;

[0028] The desired driving trajectory in the driving task is used as the reference curve of the Frenet coordinate system. The left and right boundaries are transformed into the Frenet coordinate system to form a deviation curve that varies with mileage. At the same time, the upper and lower limits of road speed are transformed into speed curves that vary with mileage. If a new driving task exists, the current state of the vehicle is obtained and transferred to the Frenet coordinate system. The driving task is determined according to the established template mentioned in step 1, thereby calculating a driving template with initial optimization parameters.

[0029] Step 2.3: Pre-aiming at the target;

[0030] Based on the motion pre-aiming optimization problem designed in step 1, and taking the driving task as input, the SQP optimization method is used to solve for the optimal curvature increment. Optimal acceleration increment And the optimal arc length increment of the curvature segment and the optimal acceleration segment arc length increment This allows for the calculation of the pre-aiming pose scheme, the pre-aiming speed scheme, and the pre-aiming turning intensity. and the intensity of the pre-aimed shifting ;

[0031] Step 2.4: Re-aiming and judgment;

[0032] If the pose deviation of the pre-aiming result from the desired pose exceeds the tolerance, or the vehicle speed deviation exceeds the tolerance, then repeat steps 2.1 and 2.2 to perform motion pre-aiming again. If there is no deviation, then determine if there is a new driving task. If there is a new driving task, repeat steps 2.1 and 2.2; otherwise, end the motion pre-aiming.

[0033] A further technical solution is that, in step 1.1, the driving task is divided into segments including driving along the road, turning left, turning right, changing lanes to the left, changing lanes to the right, overtaking to the left, overtaking to the right, and making a U-turn.

[0034] The curvature motion unit consists of two motion forms: straight-line and turning, while the velocity motion unit consists of two motion forms: steady speed and variable speed.

[0035] In a further technical solution, step 1.2 includes the following specific steps:

[0036] Based on the driving template, the pose scheme is selected. The arc or straight line follows the desired trajectory of the car; the arc is determined by two variables: curvature and the length of the curvature segment; speed scheme selection. The fitting of uniform and uniformly accelerated motion segments is determined by two variables: acceleration and the length of the acceleration segment; curvature increment is used. Acceleration increment Increment of arc length of curvature segment and the increment of the arc length of the acceleration segment The number of optimization parameters is: The feasible domain of the optimized parameters consists of the vehicle's maneuverability and driving template, as shown in equation (1):

[0037] (1);

[0038] In the formula, This represents the threshold value for the turning segment. This indicates the threshold value for the gear shift. Indicates the total length of the driving mission. The minimum segment length; the pose curve is composed of... It is composed of segments of circular arcs or straight lines, with each segment having the same curvature; the pose estimation equation adopts the circular arc estimation formula (2) and the straight line estimation formula (3):

[0039] (2);

[0040] (3);

[0041] In the above formula, Indicates the radius of the arc. Indicates the angle turned. The initial heading angle of the car. This represents the current longitudinal position of the car. This represents the initial longitudinal position of the car. This indicates the car's current lateral position. The initial lateral position of the car is determined; after a segment of the pose curve is predicted, the end point of that segment is taken as the starting point of the next segment, and the prediction is repeated, in a loop. Similarly, the speed curve is derived from... It consists of segments of uniformly accelerated or uniform motion, with each segment having the same acceleration and speed. The inference equation adopts the inference formula for uniformly accelerated motion:

[0042] (4);

[0043] in, The initial speed of the car. The car's current acceleration, The length of the speed change segment is given; the inference equation relies on the input of optimization parameters to output the desired pose scheme, desired speed scheme, turning strength scheme, and speed change strength scheme.

[0044] In a further technical solution, the specific design of the comfort value function submodule in step 1.3 is as follows:

[0045] Comfort is defined as whether the anticipated vehicle acceleration makes the driver and passengers feel comfortable; the root mean square acceleration in the time domain is used as the comfort evaluation index.

[0046] The vehicle acceleration and centripetal acceleration obtained from the spatial discrete sampling preview results are converted into driver head acceleration, and the definite integral is approximately calculated using the discrete summation method; the acceleration conversion formula is shown in equation (5):

[0047] (5);

[0048] Where, This indicates the acceleration of the driver's head. This represents the centripetal acceleration of the driver's head. Indicates the acceleration of a car. Indicates the centripetal acceleration of a car. This represents the driver's head position vector. Indicates the roll rate. Indicates the pitch angular velocity. Indicates yaw rate;

[0049] The weighted root mean square formula is shown in equation (6):

[0050] (6);

[0051] Where, This represents the weighted root mean square. Indicates the first The sampling point and the first The interval between sampling points Indicates the first The acceleration of each sampling point Indicates the number of sampling points;

[0052] For comfort evaluation indicators, only turning and shifting sections are considered, while road sections and steady-speed sections are not considered; that is, the comfort value function is taken as... The sum of the square root mean square lateral accelerations of each turning segment and The root mean square sum of the longitudinal accelerations for each speed range; comfort value function As shown in formula (7):

[0053] (7);

[0054] In the formula, Indicates the first The centripetal acceleration of the driver's head, Indicates the first The acceleration of the driver's head.

[0055] In a further technical solution, the specific design of the energy-saving value function submodule in step 1.3 is as follows:

[0056] Assuming the car's mass remains constant during its journey, the work done is divided by the mass to obtain the specific work. The sum of the squares of the total specific work consumed by the car in each segment is used as the energy-saving value function.

[0057] (8);

[0058] In the formula, For the energy-saving value function, Indicates the first The ratio of the force exerted by the car in the segment is equal to the net force. Indicates the first The sampling point and the first The interval between sampling points is specifically represented by equal distances; during vehicle movement, the resultant force of the vehicle's movement considers overcoming acceleration resistance and rolling resistance generated by tire rolling. Wind resistance generated by interaction with air And overcoming slope resistance of road gradient Therefore, the combined force The calculation formula is as follows:

[0059] (9).

[0060] A further technical solution, in step 1.3, is as follows: The specific design of the motion value function submodule is as follows:

[0061] The sum of the squares of the time consumed in each segment is used as the value function of the motion:

[0062] (10);

[0063] In the formula, The value function of motion, Indicates the first The travel time for a segment is calculated using the following formula:

[0064] (11).

[0065] A further technical solution, in step 1.3, is as follows: The specific design of the accuracy value function submodule is as follows:

[0066] Establish the lateral deviation between the pre-aiming result and the expectation as the accuracy value function.

[0067] (12);

[0068] Where, For the accuracy value function, Indicates the first The maximum offset of the segment in the Frenet coordinate system with the desired driving trajectory as the reference line. This represents the maximum speed deviation in the segment.

[0069] In a further technical solution, step 1.4 includes the following specific steps:

[0070] The constraints include traffic rules and the vehicle's maneuverability, as well as the target point requirements of the driving task. The constraints also include traffic rule constraints for the vehicle in different spatial locations, such as the vehicle's speed limit, speed limit, left boundary, and right boundary.

[0071] (13);

[0072] Where, for, and This represents the offset of the left and right boundaries in the Frenet coordinate system, with the desired driving trajectory as the reference line. and The minimum and maximum speeds permitted by traffic rules;

[0073] Vehicle maneuverability constraints include the upper limit of vehicle acceleration as speed changes. Upper limit of lateral acceleration ;

[0074] (14);

[0075] Where, This represents the current lateral acceleration of the vehicle.

[0076] The driving mission objective requirements include position and speed requirements, using the expected destination mileage. And the estimated destination mileage deviation Expected endpoint deviation and the predicted endpoint deviation deviation Expected endpoint angle and the predicted endpoint deviation deviation Expected final speed and the estimated final speed deviation Each must satisfy its own tolerance as a constraint:

[0077] (15);

[0078] In summary, the objective function is composed of the sum of four driving modes. Assigning different weights to each value function represents the degree of preference for different driving modes, resulting in the optimization problem as follows:

[0079] (16);

[0080] Where, To optimize the objective, The weights for accuracy, comfort, energy efficiency, and athleticism are respectively assigned.

[0081] A further technical solution involves, in step 2.3, pre-aiming at the turning intensity. and the intensity of the pre-aimed shifting The calculation formula is as follows:

[0082] (17);

[0083] in, It is the acceleration due to gravity. The curvature of the car.

[0084] Another objective of this invention is to provide an intelligent vehicle anti-targeting and tracking system that considers driving modes, based on the aforementioned intelligent vehicle anti-targeting and tracking method, including a driving mode performance evaluation module, a traffic rules and maneuverability constraint module, and a motion anti-targeting module.

[0085] The driving mode performance evaluation module includes a comfort value function submodule, an energy consumption value function submodule, a sportiness value function submodule, and a precision value function submodule;

[0086] The driving mode performance evaluation module and the traffic rules and maneuverability constraint module are used for optimization solutions in the motion preview module. The motion preview module is used to generate motion schemes and optimized desired driving trajectories based on the desired driving task. The motion scheme refers to the shift intensity and turning intensity required to complete the driving task.

[0087] This invention provides an intelligent vehicle pre-aiming and tracking method that considers driving modes. This method fully considers four different driving modes: comfort, energy efficiency, sportiness, and accuracy. It also designs an evaluation value function for optimization, enabling autonomous switching or weight allocation of driving modes, which is beneficial for achieving personalized motion pre-aiming in the future. Furthermore, it considers traffic rule constraints and the vehicle's maneuverability, ensuring both safety and drivability in the pre-aiming results. By employing a finite multi-segment description of the driving task, the frequency of gear shifting and turning intensities is low, which is more in line with human driving habits. Attached Figure Description

[0088] Figure 1 Design the basic architecture for the system;

[0089] Figure 2 A flowchart for online pre-aiming;

[0090] Figure 3 Example diagram of a pose scheme for overtaking. Detailed Implementation

[0091] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0092] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0093] An embodiment of the present invention provides an intelligent vehicle anti-aiming and tracking method that takes into account driving modes. The intelligent vehicle anti-aiming and tracking system is based on a designed intelligent vehicle anti-aiming and tracking system, which includes a driving mode performance evaluation module, a traffic rule and maneuverability constraint module, and a motion anti-aiming module.

[0094] The driving mode performance evaluation module includes a comfort value function submodule, an energy consumption value function submodule, a sportiness value function submodule, and a precision value function submodule.

[0095] The driving mode performance evaluation module and the traffic rules and maneuverability constraint module are mainly used for optimization solutions in the motion preview module. The motion preview module is responsible for generating motion schemes and optimized desired driving trajectories based on the desired driving task. The motion scheme refers to the shifting intensity and turning intensity required to complete the driving task. The offline design includes steps 1-7, and the online preview includes steps 8-11.

[0096] Step 1: Design of the driving template;

[0097] The driving task is segmented, and each driving task is broken down and combined into a fixed driving template that conforms to human driving habits.

[0098] The desired driving trajectory of a driving task can be divided into driving tasks such as driving along the road, turning left, turning right, changing lanes left and right, overtaking on the left, overtaking on the right, and making a U-turn. Decomposing the driving task into two types of curvature motion units—straight driving and turning—results in... A posture driving template composed of curvature motion primitives.

[0099] For the desired vehicle speed in a driving task, based on the two basic speed motion elements of steady speed and variable speed, it can be divided into acceleration segment, deceleration segment, and constant speed segment. From this, we can derive... A vehicle speed driving template composed of speed motion primitives.

[0100] For the driving template, a circular arc (or straight line) is used to fit the desired driving trajectory, with the goal of the fitted trajectory deviating from the desired trajectory by less than a threshold. The initial curvature and initial arc length are calculated using the bisection method. Similarly, uniform acceleration motion is used, with the goal of the fitted vehicle speed deviating from the desired vehicle speed by less than a threshold. The initial acceleration and initial segment length are calculated using the bisection method. Therefore, in this step, the number of segments of the pose driving template is determined based on the input driving task. Initial curvature and initial arc length The number of segments in the speed driving template Initial acceleration and initial segment length .

[0101] Step 2: Optimize the problem formulation;

[0102] Based on the driving template, the pose scheme is selected. The arc (or straight line) follows the desired trajectory of the car, and the arc (or straight line) is determined by two variables: curvature and the length of the curvature segment. Speed ​​scheme selection. Fitting of uniform and uniformly accelerated motion segments is determined by two variables: acceleration and the length of the acceleration segment. Curvature increments are used. Acceleration increment and the arc length increment of the curvature segment and the increment of the arc length of the acceleration segment The number of optimization parameters is: The feasible domain of the optimization parameters consists of the vehicle's maneuverability and driving template, as shown in equation (1):

[0103] (1);

[0104] In the formula, This represents the threshold value for the turning segment. This indicates the threshold value for the gear shift. Indicates the total length of the driving mission. The minimum segment length is given. The pose curve is composed of... It is composed of segments of circular arcs (or straight lines), each with the same curvature. The pose estimation equation uses the circular arc estimation formula (2) and the straight line estimation formula (3):

[0105] (2);

[0106] (3);

[0107] In the above formula, Indicates the radius of the arc. Indicates the angle turned. The initial heading angle of the car. This represents the current longitudinal position of the car. This represents the initial longitudinal position of the car. This indicates the car's current lateral position. This represents the initial lateral position of the vehicle. After a pose curve segment is predicted, the end point of that segment is considered the starting point of the next segment, and the prediction is repeated, in a loop. Next. Similarly, the speed curve is from It consists of segments of uniformly accelerated or uniform motion, with consistent acceleration in each segment. The equation for predicting the vehicle speed uses the formula for predicting uniformly accelerated motion.

[0108] (4);

[0109] in, The initial speed of the car. The car's current acceleration, Let be the length of the speed change segment. The prediction equations rely on the optimized parameter inputs to output the desired pose scheme, desired velocity scheme, turning strength scheme, and speed change strength scheme.

[0110] Step 3: Design of the comfort value function submodule;

[0111] Comfort refers to whether the anticipated vehicle acceleration makes the driver and passengers feel comfortable. ISO 2631 defines it as random or transient mechanical vibrations generated during vehicle operation that can interfere with driver and passenger comfort and activity. If the vibration frequency and amplitude are very low, velocity can be measured and converted into acceleration. ISO 2631 calculates the weighted root mean square (RMS) of the acceleration based on different frequencies, using the RMS as the evaluation metric. This method anticipates ultra-low frequency changes in acceleration; therefore, based on requirements and optimized real-time performance, the time-domain RMS is used as the comfort evaluation metric.

[0112] The vehicle acceleration and centripetal acceleration obtained from the discrete sampling and aiming results in the same space are converted into the driver's head acceleration, and the definite integral is approximately calculated using the discrete summation method. The acceleration conversion formula is shown in equation (5):

[0113] (5);

[0114] In the formula, This indicates the acceleration of the driver's head. This represents the centripetal acceleration of the driver's head. Indicates the acceleration of a car. Indicates the centripetal acceleration of a car. This represents the driver's head position vector. Indicates the roll rate. Indicates the pitch angular velocity. This represents the yaw rate.

[0115] The weighted root mean square formula is shown in equation (6):

[0116] (6);

[0117] In the formula, This represents the weighted root mean square. Indicates the first The sampling point and the first The interval between sampling points Indicates the first The acceleration of each sampling point This indicates the number of sampling points.

[0118] For comfort evaluation indicators, only turning and shifting sections are considered, while road sections and steady-speed sections are not considered; that is, the comfort value function is taken as... The sum of the square root mean square lateral accelerations of each turning segment and The root mean square sum of the longitudinal accelerations for each speed range. Comfort value function. As shown in formula (7):

[0119] (7);

[0120] In the formula, Indicates the first The centripetal acceleration of the driver's head, Indicates the first The acceleration of the driver's head.

[0121] Step 4: Design of the energy-saving value function submodule;

[0122] Energy efficiency refers to the total energy consumed by a car during operation. Good energy efficiency is specifically reflected in low fuel consumption and low electricity consumption. Assuming the car's mass remains constant during operation, the work done is divided by the mass to obtain the specific work. The sum of the squares of the total specific work consumed by the car in each segment is used as the energy efficiency value function.

[0123] (8);

[0124] Where, For the energy-saving value function, Indicates the first The ratio of the force exerted by the car in the segment is equal to the net force. Indicates the first The sampling point and the first The interval between sampling points is specifically represented by equal distances. During vehicle movement, the net force acting on the vehicle must overcome acceleration resistance and rolling resistance generated by tire rolling. Wind resistance generated by interaction with air And overcoming the slope resistance of the road gradient Therefore, the formula for calculating the resultant force is as follows:

[0125] (9);

[0126] Step 5: Design of the Motion Value Function Submodule;

[0127] Sportiness refers to the total time it takes for a car to complete a driving task. Good sportiness is mainly characterized by high speed and the ability to take shortcuts. The sum of the squares of the time consumed in each segment is used as the value function of sportiness.

[0128] (10);

[0129] Where, The value function of motion, Indicates the first The travel time for a segment is calculated using the following formula:

[0130] (11);

[0131] Step 6: Design of the accuracy value function submodule;

[0132] Accuracy refers to whether the pre-aimed pose scheme is as consistent as possible with the desired driving trajectory, and whether the speed scheme is as consistent as possible with the desired speed scheme. The lateral deviation between the pre-aimed result and the desired trajectory is established as the accuracy value function.

[0133] (12);

[0134] Where, For the accuracy value function, Indicates the first The maximum offset of the segment in the Frenet coordinate system with the desired driving trajectory as the reference line. For the The largest speed deviation in the segment.

[0135] Step 7: Design of traffic rules and mobility constraint module;

[0136] The constraints include traffic rules and the vehicle's maneuverability, as well as the target point requirements of the driving task. Traffic rule constraints are considered for the vehicle's location in different spatial positions. These constraints include the vehicle's maximum speed limit, minimum speed limit, left boundary, and right boundary.

[0137] (13);

[0138] Where, for, and This represents the offset of the left and right boundaries in the Frenet coordinate system, with the desired driving trajectory as the reference line. and The minimum and maximum speeds permitted by traffic rules.

[0139] Vehicle maneuverability constraints include the upper limit of vehicle acceleration as speed changes. Upper limit of lateral acceleration ;

[0140] (14);

[0141] Where, This represents the current lateral acceleration of the vehicle.

[0142] The driving mission objective requirements include position and speed requirements, using the expected destination mileage. And the estimated destination mileage deviation Expected endpoint deviation and the predicted endpoint deviation deviation Expected endpoint angle and the predicted endpoint deviation deviation Expected final speed and the estimated final speed deviation Each must satisfy its own tolerance as a constraint:

[0143] (15);

[0144] In summary, the objective function is composed of the sum of four driving modes, and different weights are assigned to each value function to represent the degree of preference for different driving modes. The optimization problem can be derived as follows:

[0145] (16);

[0146] Where, To optimize the objective, The weights for accuracy, comfort, energy efficiency, and athleticism are respectively assigned.

[0147] Step 8: Driving mode weight allocation;

[0148] Predefined driving modes allow drivers to select a driving mode based on driving mode instructions. The driving mode weights can be assigned to Precision, Comfort, Eco, and Sport modes, or drivers can manually assign driving mode weights according to their preferences.

[0149] Step 9: Obtain driving tasks and confirm driving templates;

[0150] The desired driving trajectory in the driving task is used as the reference curve in the Frenet coordinate system. The left and right boundaries are transformed into the Frenet coordinate system to form a deviation curve that varies with mileage. At the same time, the upper and lower limits of road speed are transformed into speed curves that vary with mileage. If a new driving task exists, the current state of the vehicle is obtained and transferred to the Frenet coordinate system. Simultaneously, the driving task is determined according to the predetermined template mentioned in step 1, thereby calculating a driving template with initial optimization parameters.

[0151] Step 10: Pre-aiming during movement;

[0152] Based on the motion pre-aiming optimization problem designed in steps 1-6, and taking the driving task as input, the SQP optimization method is used to solve for the optimal curvature increment. Optimal acceleration increment And the optimal arc length increment of the curvature segment and the optimal acceleration segment arc length increment This allows for the calculation of the pre-aiming pose scheme, the pre-aiming speed scheme, and the pre-aiming turning intensity. and the intensity of the pre-aimed shifting .

[0153] (17);

[0154] in, It is the acceleration due to gravity. The curvature of the car.

[0155] Step 11: Re-aiming and judgment;

[0156] If the deviation between the pre-aiming pose and the desired pose exceeds the tolerance, or if the vehicle speed deviation exceeds the tolerance, repeat steps 8 and 9 to perform motion pre-aiming again. If there is no deviation, determine if there is a new driving task. If there is a new driving task, repeat steps 8 and 9; otherwise, end the motion pre-aiming.

[0157] Taking the left-hand overtaking scenario as an example, based on the two curvature motion elements of the driving template, namely straight and turning, the left-hand overtaking is divided into nine segments: straight, left turn, straight, right turn, straight, right turn, straight, left turn, and straight. According to formulas (2) and (3) in step 2, the pose scheme is predicted using arcs and straight lines. Figure 3 As shown, it can be seen that a pose can be determined by two variables: curvature and arc length of curvature segment. A driving task can be achieved by a finite number of pose segments, which demonstrates the low-frequency advantage of the present invention.

[0158] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart car anti-aiming and tracking method considering driving modes, characterized in that, Includes the following steps: Step 1: Offline design; Step 1.1: Design the driving template; Break down driving tasks into Each curvature motion primitive and A velocity motion primitive is used to generate a pose driving template and a vehicle speed driving template. Step 1.2: Construct the optimization problem; With curvature increment Acceleration increment Increment of arc length of curvature segment and the increment of the arc length of the acceleration segment To optimize the parameters, an optimization problem is constructed that includes pose and velocity schemes; Step 1.3: Design the value function sub-modules, which specifically include a comfort value function sub-module, an energy-saving value function sub-module, a motion performance value function sub-module, and a precision value function sub-module; Step 1.4: Design of traffic rules and maneuverability constraint module; setting constraints including traffic rule constraints, vehicle maneuverability constraints, and driving task target point constraints; Step 2: Online pre-aiming; Step 2.1: Driving mode weight allocation; Based on driving mode instructions or driver preferences, select the driving mode, which is weighted as Precision, Comfort, Eco, and Sport modes; Step 2.2: Obtain driving tasks and determine driving templates; The desired driving trajectory is used as the reference curve of the Frenet coordinate system, and deviation curves and speed curves are generated based on the upper and lower speed limits of the left and right boundary roads; a driving template with initial optimization parameters is generated by combining the vehicle status. Step 2.3: Pre-aiming at the target; Input a driving task, solve the optimization problem using the SQP optimization method, and output the aiming pose scheme, aiming speed scheme, and aiming turning intensity. and the intensity of the pre-aimed shifting ; Step 2.4: Re-aiming and judgment; If the pose or speed deviation of the pre-aiming result exceeds the tolerance, repeat steps 2.1 and 2.2; otherwise, decide whether to restart pre-aiming based on the new task. Step 1.4 includes the following specific steps: Traffic rule constraints include the upper and lower speed limits for vehicles, as well as left and right boundaries; vehicle maneuverability constraints include the upper limit of acceleration that varies with vehicle speed. Upper limit of lateral acceleration The driving task target point constraints include position and speed requirements. The constraints include traffic rules and the vehicle's maneuverability, as well as the target point requirements of the driving task. The constraints also include traffic rule constraints for the vehicle in different spatial locations, such as the vehicle's speed limit, speed limit, left boundary, and right boundary. （13）； In the formula, For vehicle speed, and This represents the offset of the left and right boundaries in the Frenet coordinate system, with the desired driving trajectory as the reference line. and The minimum and maximum speeds permitted by traffic rules; Vehicle maneuverability constraints include the upper limit of vehicle acceleration as speed changes. Upper limit of lateral acceleration ; （14）； In the formula, The current lateral acceleration of the car, This represents the car's current acceleration. The driving mission objective requirements include position and speed requirements, using the expected destination mileage. and estimated destination mileage deviation Expected endpoint deviation and the predicted endpoint deviation deviation Expected endpoint angle and the predicted endpoint angle deviation Expected final speed and the estimated final speed deviation Each must satisfy its own tolerance as a constraint: （15）； In summary, the objective function is composed of the sum of four driving modes. Assigning different weights to each value function represents the degree of preference for different driving modes, resulting in the optimization problem as follows: （16）； In the formula, To optimize the objective, The weights for accuracy, comfort, energy efficiency, and athleticism are respectively assigned. For the accuracy value function, For comfort value function, For the energy-saving value function, It is the value function of motion.

2. The intelligent vehicle anti-aiming and tracking method considering driving modes according to claim 1, characterized in that, In step 1.1, the driving task is divided into segments including driving along the road, turning left, turning right, changing lanes to the left, changing lanes to the right, overtaking on the left, overtaking on the right, and making a U-turn. The curvature motion unit consists of two motion forms: straight-line and turning motion; the velocity motion unit consists of two motion forms: constant speed and variable speed. The pose driving template fits the desired driving trajectory using an arc or a straight line, with the goal of the fitted trajectory deviating from the desired trajectory by less than a threshold. The initial curvature is obtained using the bisection method. and initial arc length ; The vehicle speed driving template fits the desired vehicle speed through uniform acceleration motion, with the goal of the deviation between the fitted vehicle speed and the desired vehicle speed being less than a threshold. The initial acceleration is obtained using the bisection method. and initial segment length .

3. The intelligent vehicle anti-aiming and tracking method considering driving modes according to claim 1, characterized in that, Step 1.2 includes the following specific steps: Based on the driving template, the posture scheme is selected. The arc or straight line follows the desired trajectory of the car; the arc is determined by two variables: curvature and the length of the curvature segment; speed scheme selection. The fitting of uniform and uniformly accelerated motion segments is determined by two variables: acceleration and the length of the acceleration segment; curvature increment is used. Acceleration increment Increment of arc length of curvature segment and the increment of the arc length of the acceleration segment The number of optimization parameters is: The feasible domain of the optimized parameters consists of the vehicle's maneuverability and driving template, as shown in equation (1): （1）； In the formula, This represents the threshold value for the turning segment. This indicates the threshold value for the gear shift. Indicates the total length of the driving mission. The minimum segment length; the pose curve is composed of... It is composed of segments of circular arcs or straight lines, with each segment having the same curvature; the pose estimation equation adopts the circular arc estimation formula (2) and the straight line estimation formula (3): （2）； （3）； In the above formula, Indicates the radius of the arc. Indicates the angle turned. The initial heading angle of the car. This represents the current longitudinal position of the car. This represents the initial longitudinal position of the car. This indicates the car's current lateral position. The initial lateral position of the car is determined; after one segment of the pose curve is predicted, the end point of that segment is taken as the starting point of the next segment, and the prediction is repeated. Similarly, the speed curve is derived from... It consists of segments of uniformly accelerated or uniform motion, with consistent acceleration in each segment. The equation for the vehicle speed V is derived using the formula for uniformly accelerated motion. （4）； in, The initial speed of the car. The car's current acceleration, The length of the speed change segment is given; the inference equation relies on the input of optimization parameters to output the desired pose scheme, desired speed scheme, turning strength scheme, and speed change strength scheme.

4. The intelligent vehicle pre-aiming and tracking method considering driving modes according to claim 3, characterized in that, In step 1.3, the specific design of the comfort value function submodule is as follows: Comfort is defined as whether the anticipated vehicle acceleration makes the driver and passengers feel comfortable; the root mean square acceleration in the time domain is used as the comfort evaluation index. The vehicle acceleration and centripetal acceleration obtained from the spatial discrete sampling preview results are converted into driver head acceleration, and the definite integral is approximately calculated using the discrete summation method; the acceleration conversion formula is shown in equation (5): （5）； In the formula, This indicates the acceleration of the driver's head. This represents the centripetal acceleration of the driver's head. Indicates the acceleration of a car. Indicates the centripetal acceleration of a car. This represents the driver's head position vector. Indicates the roll rate. Indicates the pitch angular velocity. Indicates yaw rate; The weighted root mean square formula is shown in equation (6): （6）； In the formula, This represents the weighted root mean square. Indicates the first The sampling point and the first The interval between sampling points Indicates the first The acceleration of each sampling point Indicates the number of sampling points; For comfort evaluation indicators, only turning and shifting sections are considered, while road sections and steady-speed sections are not considered; that is, the comfort value function is taken as... The sum of the square root mean square lateral accelerations of each turning segment and The root mean square sum of the longitudinal accelerations for each speed range; comfort value function As shown in formula (7): （7）； In the formula, Indicates the first The centripetal acceleration of the driver's head, Indicates the first The acceleration of the driver's head.

5. The intelligent vehicle anti-aiming and tracking method considering driving modes according to claim 4, characterized in that, In step 1.3, the specific design of the energy-saving value function submodule is as follows: Assuming the car's mass remains constant during its journey, the work done is divided by the mass to obtain the specific work. The sum of the squares of the total specific work consumed by the car in each segment is used as the energy-saving value function. （8）； In the formula, For the energy-saving value function, Indicates the first The ratio of the force exerted by the car in the segment is equal to the net force. Indicates the first The sampling point and the first The interval between sampling points is specifically represented by equal distances; during vehicle movement, the resultant force of the vehicle's movement considers overcoming acceleration resistance and rolling resistance generated by tire rolling. Wind resistance generated by interaction with air And overcoming slope resistance of road gradient Therefore, the formula for calculating the resultant force f is as follows: （9）。 6. The intelligent vehicle pre-aiming and tracking method considering driving modes according to claim 5, characterized in that, In step 1.3, the specific design of the motion value function submodule is as follows: The sum of the squares of the time consumed in each segment is used as the value function of the motion: （10）； In the formula, The value function of motion, Indicates the first The travel time for a segment is calculated using the following formula: （11）。 7. The intelligent vehicle anti-aiming and tracking method considering driving modes according to claim 6, characterized in that, In step 1.3, the specific design of the accuracy value function submodule is as follows: Establish the lateral deviation between the pre-aiming result and the expected result as the accuracy value function: （12）； In the formula, For the accuracy value function, Indicates the first The maximum offset of the segment in the Frenet coordinate system with the desired driving trajectory as the reference line. To indicate the first The largest speed deviation in the segment.

8. The intelligent vehicle anti-aiming and tracking method considering driving modes according to claim 1, characterized in that, In step 2.3, the pre-aiming turning intensity and the intensity of the pre-aimed shifting The calculation formula is as follows: （17）； in, It is the acceleration due to gravity. The curvature of the car.

Images