A space-time decoupled high-speed single-lane driving method and vehicle

Abstract

This invention provides a spatiotemporally decoupled high-speed single-lane driving method and vehicle. The spatiotemporally decoupled high-speed single-lane driving method includes: selecting a decision boundary based on the vehicle's information; making a decision on target vehicles appearing within the decision boundary; and obtaining the longitudinal final state (s) of the vehicle based on the target vehicle's predicted trajectory. 纵 ,v 纵 ,a 纵 ,t sample ) and transverse final state (l sample ,v 横 ,a 横 ,s sample Based on the vehicle's position and longitudinal final state, the longitudinal trajectory equation of the vehicle is obtained; based on the vehicle's position and lateral final state, the lateral trajectory equation of the vehicle is obtained; the lateral trajectory equation and the longitudinal trajectory equation are matched to obtain a trajectory cluster; the trajectory cluster is evaluated, and the optimal trajectory is selected and output. This invention solves the technical problem of high time complexity and long solution time of traditional control algorithms.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation and autonomous driving technology, and more specifically, to a spatiotemporally decoupled high-speed single-lane driving method and a vehicle. Background Technology

[0002] Currently, intelligent driving technology, as a rising star in driver assistance, has excellent potential for development in user experience and technological depth. Furthermore, thanks to the vigorous development of its supporting disciplines, it has begun to show its potential. At present, most cars traveling on highways, elevated roads, and expressways are equipped with technologies such as adaptive cruise control, which can reduce driver fatigue. These technologies are mostly based on traditional control algorithms, capable of obtaining information from environmental feedback and updating control strategies accordingly.

[0003] However, in actual construction, there is a problem: the time complexity of traditional control algorithms is too high, and the solution takes a long time. Summary of the Invention

[0004] This invention solves the technical problem of high time complexity and long solution time of traditional control algorithms.

[0005] To address the above problems, this invention provides a spatiotemporally decoupled high-speed single-lane driving method, comprising:

[0006] Step S100: Select the decision boundary based on the vehicle information;

[0007] Step S200: Make a decision on the target vehicles appearing within the decision boundary;

[0008] Step S300: Obtain the longitudinal final state (s) of the vehicle based on the target vehicle's predicted trajectory. 纵 ,v 纵 ,a 纵 ,t sample ) and transverse final state (l sample ,v 横 ,a 横 ,s sample );

[0009] Step S400: Obtain the longitudinal trajectory equation of the vehicle based on its position and longitudinal final state;

[0010] Step S500: Based on the position and lateral final state of the vehicle, obtain the lateral trajectory equation of the vehicle.

[0011] Step S600: Match the lateral trajectory equation and the longitudinal trajectory equation to obtain the trajectory cluster;

[0012] Step S700: Evaluate the trajectory cluster, select and output the optimal trajectory;

[0013] The decision-making process includes: determining whether the target vehicle is a following vehicle or a vehicle overtaking; 纵 v represents the longitudinal distance of the vehicle at the end of the sampling point. 纵 a represents the longitudinal velocity of the vehicle at the end of the sampling time slice. 纵 t represents the final acceleration of the vehicle in the longitudinal direction. sample For the sampling time slice, l sample v represents the lateral distance between the sampling points of the vehicle's final state. 横 a represents the lateral velocity of the vehicle at the end of the sampling time slice. 横 s represents the final acceleration of the vehicle in the lateral direction. sample This represents the longitudinal distance between the sampling points of the vehicle's final state.

[0014] Compared with existing technologies, the technical effects achieved by this solution are as follows: First, the solution coordinate system is transformed from the Cartesian coordinate system to the natural coordinate system, that is, the reference line is used as the planning axis. The direct benefit of this planning method is that it can be applied to any route, even a 180° turn, with strong robustness. Steps S400 and S500 transform the trajectory calculation in the natural coordinate system (x,y,t) into a solution in two two-dimensional coordinates (s,t) and (l,s). The advantages of doing so are: effectively reducing the computational time complexity; reducing the solution space dimension; supporting multi-threaded simultaneous solution, which can be deployed under the relatively limited hardware resources of the vehicle and meets the actual planning frequency requirements; reducing the time consumption of a single trajectory solution means that more searches and calculations can be performed in the solution space, which can get closer to the optimal solution and improve the quality of the planned trajectory.

[0015] In one embodiment of the present invention, selecting a decision boundary based on vehicle information includes:

[0016] Step S110: Obtain the lane length and lane width of the lane where the vehicle is located;

[0017] Step S120: Define the length of the decision boundary region as l based on the lane length. lane ;

[0018] Step S130: Define the width of the decision boundary area based on the lane width and lateral safety distance as (d lane +d safe );

[0019] Among them, l lane d represents the furthest planned length. laneRepresents lane width, d safe This represents the lateral safety distance.

[0020] Compared with existing technologies, the technical effects achieved by this solution are as follows: when selecting the decision boundary, the lane length and lane width of the vehicle's lane need to be considered, and the nearest vehicle within the decision boundary is taken as the target vehicle. The length of the decision boundary area is the furthest planned length l. lane The width of the area is the sum of the lane width and the lateral safety distance, i.e., (d) lane +d safe By considering lateral safety distances within the area width, the safety of the vehicle during driving can be improved while increasing the decision boundary area.

[0021] In one embodiment of the present invention, the decision-making process for a target vehicle appearing within the decision boundary includes:

[0022] The vehicle closest to the vehicle within the decision boundary is identified as the target vehicle.

[0023] Step S210: If the target vehicle's predicted trajectory is a following vehicle decision, then the target vehicle is determined to be a following vehicle.

[0024] Step S220: If the target vehicle's predicted trajectory is an overtaking decision, then the target vehicle is determined to be an overtaking vehicle.

[0025] Compared with existing technologies, the technical effects achieved by this solution are as follows: The nearest vehicle within the decision boundary is designated as the target vehicle. A decision is made regarding the target vehicle, and the decision module within the vehicle receives perception information and makes a corresponding decision based on that information, which is then provided to the planning module. If the target vehicle's predicted trajectory is a following decision, the target vehicle is determined to be a following vehicle, and the vehicle's trajectory follows the following vehicle without overtaking it while maintaining a safe distance. If the target vehicle's predicted trajectory is an overtaking decision, the target vehicle is determined to be an overtaking vehicle, which is limited to vehicles in the adjacent lane of the vehicle, and the vehicle's planned path will cause it to overtake the overtaking vehicle.

[0026] In one embodiment of the present invention, the longitudinal final state (s) of the vehicle is obtained by predicting the target trajectory of the target vehicle. 纵 ,v 纵 ,a 纵 ,t sample ) and transverse final state (l sample ,v 横 ,a 横 ,s sample ),include:

[0027] Step S310: Obtain the longitudinal distance s of the target vehicle in each time slice in the natural coordinate system using the target vehicle's predicted trajectory. 纵 and speed limit v 纵 ;

[0028] Step S320: Based on the longitudinal distance and the upper limit of speed, construct a sampling range, sample within the sampling range, and obtain the longitudinal final state (s) of the vehicle. 纵 ,v 纵 ,a 纵 ,t sample );

[0029] Step S330: Obtain the transverse final state (l) based on the longitudinal final state. sample ,v 横 ,a 横 ,s sample );

[0030] Among them, s sample =s 纵 v 横 =0, a 横 =0.

[0031] Compared with existing technologies, the technical effects achieved by this solution are as follows: The position of the target vehicle is different at each moment. Therefore, for each moment, it is necessary to calculate the position and speed of the target vehicle, and then, based on the obtained target vehicle status, calculate the longitudinal distance s that the vehicle can reach at the current moment. 纵 and the corresponding final state velocity v 纵 .

[0032] In one embodiment of the present invention, the longitudinal distance s of the target vehicle in each time slice in the natural coordinate system is obtained by using the target predicted trajectory of the target vehicle. 纵 and speed limit v 纵 ,include:

[0033] Step S311: When the target vehicle is a following vehicle,

[0034] v 纵 =min(v lane_limit ,(v ego +a ego_max t sample ),v vehicle_follow );

[0035]

[0036] a 纵 =0;

[0037] Among them, v lane_limit V represents the road speed limit.ego a represents the real-time speed of the vehicle. ego_max t represents the maximum acceleration of the vehicle. sample Represents the sampling time slice, v vehicle_follow The longitudinal speed of the following vehicle is represented by s. lane_limit s represents the pre-set feasible planning distance. vehicle_follow s represents the longitudinal end point distance of the following vehicle. vehicle_follow_safe This represents the safe distance to maintain when following another vehicle.

[0038] Step S312: When the target vehicle is an overtaking vehicle,

[0039]

[0040]

[0041] a 纵 =0;

[0042] Among them, v vehicle_over The longitudinal velocity s represents the speed of the overtaking vehicle during the sampling time slice. vehicle_over The longitudinal distance of the overtaking vehicle during the sampling time slice; s vehicle_over_safe This refers to the safe distance that a vehicle must maintain from the overtaking vehicle after overtaking.

[0043] Compared with existing technologies, the technical effects achieved by this solution are as follows: If the target vehicle is a following vehicle, the longitudinal trajectory final state sampling is as follows: The position of the following vehicle is different at each moment. Therefore, for each moment, the position and speed of the following vehicle need to be calculated, and based on the acquired state of the following vehicle, the longitudinal distance and corresponding final state speed that the vehicle can reach at the current moment are calculated. The final state speed needs to meet the following requirements: 1. Lower than the current lane speed limit; 2. Lower than the speed of the following vehicle; 3. Not exceeding the maximum speed that can be reached by accelerating with maximum acceleration before the sampling time point; The final state longitudinal distance needs to meet the following requirements: 1. Planned length limit; 2. The longitudinal distance of the following vehicle, with a reserved safety distance; 3. The farthest travel distance at the sampling time point. This allows for automatic control of distance and speed based on the following vehicle, achieving adaptive adjustment of speed and distance, greatly increasing the flexibility of following capability.

[0044] If the target vehicle is making an overtaking decision, the longitudinal trajectory final state sampling is as follows: The position of the overtaking vehicle is different at each moment. Therefore, for each moment, the position and speed of the overtaking vehicle need to be calculated. Based on the acquired state of the overtaking vehicle, the longitudinal distance that the vehicle can reach at the current moment and the corresponding final state speed are then calculated. The final state speed needs to meet the following requirements: 1. Lower than the current lane speed limit; 2. Higher than the overtaking vehicle speed; 3. Not exceeding the maximum speed that can be reached by accelerating at maximum acceleration before the sampling time point. The final state longitudinal distance needs to meet the following requirements: 1. Planned length limit; 2. Longitudinal distance of the overtaking vehicle, with a reserved safety distance; 3. The farthest travel distance at the sampling time point.

[0045] In one embodiment of the present invention, obtaining the longitudinal trajectory equation of the vehicle based on the vehicle's position and final longitudinal state includes:

[0046] Step S410: Set longitudinal constraints, namely...

[0047] The longitudinal constraint of the vehicle is Endpoint longitudinal constraint is Where s(0) is the longitudinal distance of the vehicle at the planned starting point; v(0) is the longitudinal velocity of the vehicle at the planned starting point; acc(0) is the longitudinal acceleration of the vehicle at the planned starting point; s(t sample v(t) represents the longitudinal distance of the vehicle at the end of the planned time slice; sample ) represents the longitudinal velocity of the vehicle at the end of the planned time slice; acc(t) sample () represents the longitudinal acceleration of the vehicle at the end of the planned time slice;

[0048] Step S420: The longitudinal trajectory planning uses a first-to-fifth degree polynomial, denoted as s(t) = a s t 5 +b s t 4 +c s t 3 +d s t 2 +e s t+f s Where s(t) is the longitudinal distance as a function of time, a s ,b s ,c s ,d s ,e s ,f s These are the coefficients of the longitudinal trajectory equation;

[0049] Step S430: The first fifth-degree polynomial can be differentiated with respect to the independent variable to obtain the corresponding vertical objective equation, i.e. Where v(t) is the first derivative of s(t), which is the longitudinal velocity of the vehicle as a function of time; acc(t) is the second derivative of s(t), which is the longitudinal acceleration of the vehicle as a function of time.

[0050] Step S440: Substitute the longitudinal constraints into the longitudinal objective equation, and adjust the coefficients (a) of the longitudinal equation. s ,b s ,c s ,d s ,e s ,f s Solve the equation to obtain the longitudinal trajectory equation.

[0051] Compared with existing technologies, the technical effects achieved by this solution are as follows: For both following and overtaking decisions, the longitudinal trajectory generation method is consistent. Therefore, the longitudinal trajectory curve can be generated in the same way. Longitudinal constraints are divided into start-point and end-point constraints. The start-point constraints include longitudinal distance, longitudinal velocity, and longitudinal acceleration, while the end-point constraints also include longitudinal distance, longitudinal velocity, and longitudinal acceleration. Therefore, there are a total of six constraints on the longitudinal trajectory. Calculating the longitudinal trajectory requires satisfying all six constraints while simultaneously minimizing the rate of change of longitudinal acceleration. Substituting the six longitudinal constraints into the longitudinal objective equation, and based on the obtained six parametric equation constraints, the parameter (a) can be adjusted. s ,b s ,c s ,d s ,e s ,f s Solving this equation will eventually yield the corresponding longitudinal trajectory equation.

[0052] In one embodiment of the present invention, obtaining the lateral trajectory equation of the vehicle based on its position and final lateral state includes:

[0053] Step S510: Set lateral constraints, specifically the lateral constraints for the vehicle. Endpoint lateral constraint is Where, l(0) is the lateral distance of the vehicle at the planned starting point; v_lat(0) is the lateral velocity of the vehicle at the planned starting point; acc_lat(0) is the lateral acceleration of the vehicle at the planned starting point; l(t sample v_lat(s) represents the lateral distance of the vehicle at the end of the planned time slice; sample ) represents the lateral velocity of the vehicle at the end of the planned time slice; acc_lat(s sample () represents the lateral acceleration of the vehicle at the end of the planned time slice;

[0054] Step S520: The second-to-fifth degree polynomial is used in the lateral trajectory planning, denoted as l(s) = al s 5 +b l s 4 +c l s 3 +d l s 2 +e l s+f l Where l(t) is the lateral distance as a function of time, a l ,b l ,c l ,d l ,e l ,f l The coefficients of the transverse trajectory equation;

[0055] Step S530: The second-to-fifth degree polynomial can be differentiated with respect to the independent variable to obtain the corresponding horizontal objective equation, i.e. Where v_lat(s) is the first derivative of l(s), which is the lateral velocity of the vehicle as a function of time; acc_lat(s) is the second derivative of l(t), which is the lateral acceleration of the vehicle as a function of time.

[0056] Step S540: Substitute the lateral constraints into the lateral objective equation, and adjust the coefficients of the lateral equation (a) l ,b l ,c l ,d l ,e l ,f l Solve the equation to obtain the horizontal trajectory.

[0057] Compared with existing technologies, the technical effects achieved by this solution are as follows: Lateral constraints are divided into starting and ending constraints. The starting constraint includes lateral distance, lateral velocity, and lateral acceleration, while the ending constraint also includes lateral distance, lateral velocity, and lateral acceleration. Therefore, there are a total of six constraints on the lateral trajectory. Calculating the lateral trajectory requires satisfying all six constraints while simultaneously minimizing the rate of change of lateral acceleration using a functional. Substituting the six lateral constraints into the lateral objective equation, and based on the obtained six parametric equation constraints, the parameter (a) can be adjusted. l ,b l ,c l ,d l ,e l ,f l Solve the equation. The resulting equation for the transverse trajectory can then be obtained.

[0058] In one embodiment of the present invention, matching the lateral trajectory equation and the longitudinal trajectory equation to obtain a trajectory cluster includes:

[0059] The longitudinal trajectory equation and the longitudinal trajectory equation are matched and calculated by sampling time slices;

[0060] The obtained trajectory clusters are:

[0061]

[0062] Compared with existing technologies, the technical effect achieved by adopting this technical solution is: by sampling time slices, the longitudinal trajectory equations are matched and calculated to obtain trajectory clusters, which provides a basis for subsequent evaluation.

[0063] In one embodiment of the present invention, evaluating a trajectory cluster, selecting and outputting the optimal trajectory includes:

[0064] Step S710: Construct the trajectory evaluation function based on the trajectory cluster as f(trajectory) = w1·cost_objective + w2·cost_smooth + w3·cost_referenceline;

[0065] Step S720: Evaluate each trajectory in the trajectory cluster according to the trajectory evaluation function and obtain a score;

[0066] Step S730: Select the optimal trajectory based on the score and output the optimal trajectory as the final trajectory;

[0067] Where w1, w2, and w3 are weighting coefficients; cost_objective is the distance loss to each obstacle; cost_smooth is the smoothing loss of the final trajectory; and cost_referenceline is the deviation loss from the reference line.

[0068] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: based on the constructed loss function, each trajectory is evaluated, a score is obtained for each trajectory, and the optimal trajectory under this standard is selected as the final trajectory for this planning and issued. By selecting the optimal trajectory and issuing it through this step, the reliability of the solution is improved.

[0069] In another aspect, embodiments of the present invention provide a vehicle, including: a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement a high-speed single-lane driving method with spatiotemporal decoupling as described in any embodiment of the present invention.

[0070] Compared with the prior art, the technical effects achieved by adopting this technical solution are as follows: The vehicle of the present invention is used to implement the high-speed single-lane driving method with spatiotemporal decoupling as in any embodiment of the present invention, and therefore has all the beneficial effects of the high-speed single-lane driving method with spatiotemporal decoupling as in any embodiment of the present invention, which will not be repeated here.

[0071] By adopting the technical solution of the present invention, the following technical effects can be achieved:

[0072] (1) The present invention first converts the solution coordinate system from the Cartesian coordinate system to the natural coordinate system, that is, the reference line is used as the planning axis. The direct benefit of this planning method is that it can be applied to any route with one method, even a 180° turn, and has strong robustness.

[0073] (2) Steps S400 and S500 transform the trajectory calculation in the natural coordinate system (x,y,t) into a solution in two two-dimensional coordinates (s,t) and (l,s). The advantages of doing so are: effectively reducing the time complexity of the calculation; reducing the spatial dimension of the solution; supporting multi-threaded simultaneous solution, which can be deployed under the relatively limited hardware resources of the vehicle and meets the actual planning frequency requirements.

[0074] (3) Reducing the time required to solve a single trajectory means that more searches and calculations can be performed in the solution space, which can get closer to the optimal solution and improve the quality of the planned trajectory.

[0075] (4) The horizontal and vertical decoupling is solved separately, supporting multi-threaded operation, which effectively improves the solution efficiency and can be successfully solved within the corresponding planning time limit under the relatively strict hardware resource constraints of the actual vehicle. Attached Figure Description

[0076] Figure 1 This is one of the flowcharts of a high-speed single-lane driving method with spatiotemporal decoupling provided in Embodiment 1 of the present invention.

[0077] Figure 2 This is the second flowchart of a high-speed single-lane driving method with spatiotemporal decoupling provided in Embodiment 1 of the present invention.

[0078] Figure 3 This is a block diagram of a vehicle provided in Embodiment 2 of the present invention.

[0079] Explanation of reference numerals in the attached figures:

[0080] 100 - Vehicle; 110 - Memory; 111 - Computer program; 120 - Processor. Detailed Implementation

[0081] To make the above-mentioned objectives, features, and advantages of the present invention more apparent and understandable, the technical solutions in the embodiments of the present invention are clearly and completely described. 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.

[0082] Example 1

[0083] See Figures 1-2 This invention provides a spatiotemporally decoupled high-speed single-lane driving method, comprising:

[0084] Step S100: Select the decision boundary based on the vehicle information;

[0085] Step S200: Make a decision on the target vehicles appearing within the decision boundary;

[0086] Step S300: Obtain the longitudinal final state (s) of the vehicle based on the target vehicle's predicted trajectory. 纵 ,v 纵 ,a 纵 ,t sample ) and transverse final state (l sample ,v 横 ,a 横 ,s sample );

[0087] Step S400: Obtain the longitudinal trajectory equation of the vehicle based on its position and longitudinal final state;

[0088] Step S500: Based on the position and lateral final state of the vehicle, obtain the lateral trajectory equation of the vehicle.

[0089] Step S600: Match the lateral trajectory equation and the longitudinal trajectory equation to obtain the trajectory cluster;

[0090] Step S700: Evaluate the trajectory cluster, select and output the optimal trajectory;

[0091] The decision-making process includes: determining whether the target vehicle is a following vehicle or a vehicle overtaking; 纵 v represents the longitudinal distance of the vehicle at the end of the sampling point. 纵 a represents the longitudinal velocity of the vehicle at the end of the sampling time slice. 纵 t represents the final acceleration of the vehicle in the longitudinal direction. sample For the sampling time slice, l sample v represents the lateral distance between the sampling points of the vehicle's final state. 横 a represents the lateral velocity of the vehicle at the end of the sampling time slice. 横 s represents the final acceleration of the vehicle in the lateral direction. sample This represents the longitudinal distance between the sampling points of the vehicle's final state.

[0092] The solution in this embodiment is applied to the situation where the vehicle is driving on a single lane. Most vehicles driving on highways, elevated roads and expressways are equipped with technologies such as adaptive cruise control that can reduce driver fatigue. Most of these technologies are based on traditional control algorithms, which have high time complexity and take a long time to solve.

[0093] Specifically, this invention is essentially a cruise control scheme that does not involve lane changes. This invention takes the information input from upstream (sensing, prediction, and positioning), using the vehicle itself as the planning starting point, and transforms the vehicle's coordinate system into a natural coordinate system, that is, using reference lines as the horizontal axis and right-angled reference lines as the vertical axis, to perform the next horizontal and vertical planning steps. The main planning process is as follows: based on information such as the vehicle itself and its lane, the decision boundary for judging target vehicles is defined; if a target vehicle appears within the decision range, a decision is made regarding the target vehicle, mainly categorized as a following vehicle or an overtaking vehicle. Following vehicles are not limited to vehicles in the current lane; they can also be vehicles in adjacent lanes. Overtaking vehicles are limited to vehicles in adjacent lanes.

[0094] Next, through target vehicle trajectory prediction analysis, the longitudinal distance and upper speed limit of the target vehicle in each time slice in the natural coordinate system are obtained. Based on the obtained longitudinal distance and upper speed limit, a sampling range is constructed, and sampling is performed within the sampling range using a uniform sampling rule to obtain the final states in the lateral and longitudinal directions. Based on the position of the autonomous vehicle and the final states in the longitudinal and lateral directions, a fifth-order polynomial is solved through six constraints to obtain the longitudinal and lateral trajectory equations of the autonomous vehicle. In this process, the trajectory calculation in the natural coordinate system (x,y,t) is transformed and decomposed into solutions in two two-dimensional coordinates (s,t) and (l,s). The advantages of this approach are: effectively reducing the computational time complexity; reducing the solution space dimension; supporting multi-threaded simultaneous solving; enabling deployment with relatively limited hardware resources for the autonomous vehicle and meeting actual planning frequency requirements; reducing the time consumption of a single trajectory solution, which means that more searches and calculations can be performed in the solution space, resulting in a closer approach to the optimal solution and improved trajectory planning quality. The solution for lateral and longitudinal decoupling is performed separately, and multi-threaded operation is supported, which effectively improves the solution efficiency. It can successfully solve the problem within the corresponding planning time limit under the relatively strict hardware resource constraints of real vehicles.

[0095] After obtaining the longitudinal and lateral trajectory equations, each lateral and longitudinal trajectory curve is matched by time-slice sampling to obtain a trajectory cluster. For each trajectory in the trajectory cluster, the trajectory evaluation function is used for evaluation, and the trajectory with the minimum loss is selected as the output trajectory.

[0096] Furthermore, based on the vehicle information, a decision boundary is selected, including:

[0097] Step S110: Obtain the lane length and lane width of the lane where the vehicle is located;

[0098] Step S120: Define the length of the decision boundary region as l based on the lane length. lane ;

[0099] Step S130: Define the width of the decision boundary area based on the lane width and lateral safety distance as (d lane +d safe );

[0100] Among them, l lane d represents the furthest planned length. lane Represents lane width, d safe This represents the lateral safety distance.

[0101] Specifically, when selecting the decision boundary, the lane length and lane width of the lane where the vehicle is located need to be considered, and the nearest vehicle within the decision boundary is taken as the target vehicle. The length of the decision boundary area is the furthest planned length l. lane The width of the area is the sum of the lane width and the lateral safety distance, i.e., (d) lane +d safe By considering lateral safety distances within the area width, the safety of the vehicle during driving can be improved while increasing the decision boundary area.

[0102] Furthermore, decision-making is performed on target vehicles appearing within the decision boundary, including:

[0103] The vehicle closest to the vehicle within the decision boundary is identified as the target vehicle.

[0104] Step S210: If the target vehicle's predicted trajectory is a following vehicle decision, then the target vehicle is determined to be a following vehicle.

[0105] Step S220: If the target vehicle's predicted trajectory is an overtaking decision, then the target vehicle is determined to be an overtaking vehicle.

[0106] Specifically, the nearest vehicle within the decision boundary is designated as the target vehicle. A decision is made regarding this target vehicle. The decision module within the vehicle receives perception information and makes a corresponding decision based on that information, which is then fed into the planning module. If the target vehicle's predicted trajectory is a following decision, the target vehicle is determined to be a following vehicle, and the vehicle's trajectory follows the following vehicle, maintaining a safe distance without overtaking it. If the target vehicle's predicted trajectory is an overtaking decision, the target vehicle is determined to be an overtaking vehicle, limited to vehicles in the adjacent lanes of the vehicle, and the vehicle's planned path will allow it to overtake the overtaking vehicle.

[0107] Preferably, the decision module frequency can be 1 Hz, sending a result every second. During this period, the planning remains unchanged, and the decision for the target vehicle is re-marked when the next decision result is received.

[0108] Furthermore, by using the target vehicle's predicted trajectory, the longitudinal final state (s) of the vehicular vehicle is obtained. 纵 ,v 纵 ,a 纵 ,t sample ) and transverse final state (l sample ,v 横 ,a 横 ,s sample ),include:

[0109] Step S310: Obtain the longitudinal distance s of the target vehicle in each time slice in the natural coordinate system using the target vehicle's predicted trajectory. 纵 and speed limit v 纵 ;

[0110] Step S320: Based on the longitudinal distance and the upper limit of speed, construct a sampling range, sample within the sampling range, and obtain the longitudinal final state (s) of the vehicle. 纵 ,v 纵 ,a 纵 ,t sample );

[0111] Step S330: Obtain the transverse final state (l) based on the longitudinal final state. sample ,v 横 ,a 横 ,s sample );

[0112] Among them, s sample =s 纵 v 横 =0, a 横 =0.

[0113] Specifically, the target vehicle's position is different at each moment. Therefore, for each moment, it is necessary to calculate the target vehicle's position and speed, and based on the acquired target vehicle's corresponding state, calculate the longitudinal distance s that the vehicle can reach at the current moment. 纵 and the corresponding final state velocity v 纵 In the longitudinal direction, regardless of whether the target vehicle is following or overtaking, the target vehicle's previous state is the same as the current state of the vehicle itself. A sampling range is constructed based on the longitudinal distance and speed limit. Sampling is performed within this range to obtain the vehicle's final longitudinal state (s). 纵 ,v 纵 ,a 纵 ,t sample ); Obtain the transverse final state (l) based on the longitudinal final state.sample ,v 横 ,a 横 ,s sample ), through the longitudinal distance s 纵 The horizontal distance l sample and sampling time t sample Contact, and at the same time v 横 =0, a 横 =0.

[0114] Furthermore, by using the target vehicle's predicted trajectory, the longitudinal distance s of the target vehicle in each time slice in the natural coordinate system is obtained. 纵 and speed limit v 纵 ,include:

[0115] Step S311: When the target vehicle is a following vehicle,

[0116] v 纵 =min(v lane_limit ,(v ego +a ego_max t sample ),v vehicle_follow );

[0117]

[0118] a 纵 =0;

[0119] Among them, v lane_limit v represents the road speed limit. ego a represents the real-time speed of the vehicle. ego_max t represents the maximum acceleration of the vehicle. sample Represents the sampling time slice, v vehicle_follow The longitudinal speed of the following vehicle is represented by s. lane_limit s represents the pre-set feasible planning distance. vehicle_follow s represents the longitudinal end point distance of the following vehicle. vehicle_follow_safe This represents the safe distance to maintain when following another vehicle.

[0120] Step S312: When the target vehicle is an overtaking vehicle,

[0121]

[0122]

[0123] a 纵 =0;

[0124] Among them, v vehicle_over The longitudinal velocity s represents the speed of the overtaking vehicle during the sampling time slice. vehicle_overThe longitudinal distance of the overtaking vehicle during the sampling time slice; s vehicle_over_safe This refers to the safe distance that a vehicle must maintain from the overtaking vehicle after overtaking.

[0125] Specifically, if the target vehicle is a following vehicle, the longitudinal trajectory final state sampling is as follows: The position of the following vehicle is different at each moment. Therefore, for each moment, the position and speed of the following vehicle need to be calculated. Based on the acquired state of the following vehicle, the longitudinal distance that the own vehicle can reach at the current moment and the corresponding final state speed are then calculated. The final state speed needs to meet the following conditions: 1. Lower than the current lane speed limit; 2. Lower than the speed of the following vehicle; 3. Not exceeding the maximum speed that can be reached by accelerating with maximum acceleration before the sampling time point. Therefore, the final state longitudinal speed is calculated as v. 纵 =min(v lane_limit ,(v ego +a ego_max t sample ),v vehicle_follow ), v 纵 The longitudinal speed of the vehicle at the end of the sampling time slice; the longitudinal distance at the end of the sampling time slice must meet the following requirements: 1. Planning length limit; 2. Following vehicle longitudinal distance, with a safety margin; 3. The furthest travel distance at the sampling time point. Therefore, the longitudinal distance at the end of the sampling time slice is calculated as follows: s 纵 The longitudinal distance of the vehicle at the final state of the sampling point; the final acceleration a 纵 0 m / s 2 Therefore, for each sampling time, the final state of the vehicle's longitudinal trajectory is represented as (s 纵 ,v 纵 ,a 纵 ,t sample It can automatically control the distance and speed based on the vehicle it is following, achieving adaptive adjustment of speed and distance, which greatly increases the flexibility of following capability.

[0126] If the target vehicle is making an overtaking decision, the final state sampling of the longitudinal trajectory is as follows: The position of the overtaking vehicle is different at each time step. Therefore, for each time step, it is necessary to calculate the position and speed of the overtaking vehicle, and based on the acquired state of the overtaking vehicle, calculate the longitudinal distance that the vehicle can reach at the current time step and the corresponding final state speed. The final state speed needs to meet the overtaking speed conditions: 1. Lower than the current lane speed limit; 2. Higher than the speed of the overtaking vehicle; 3. Not exceeding the maximum speed that can be reached by accelerating with maximum acceleration before the sampling time point. Therefore, the final state longitudinal speed is calculated as follows: The overtaking speed condition is: the longitudinal velocity v of the vehicle at the end of the sampling time slice. 纵 The longitudinal velocity v of the overtaking vehicle in the sampling time slice is greater than that of the vehicle passing through.vehicle_over And less than or equal to the road speed limit v lane_limit When the longitudinal velocity v of the vehicle is at the end of the sampling time slice 纵 When the overtaking speed condition is met, the vehicle makes an overtaking decision; when v 纵 If the overtaking speed requirement is not met, the vehicle will not attempt to overtake and will instead follow the target vehicle.

[0127] The longitudinal distance in the final state under the overtaking decision must meet the overtaking speed and distance conditions: 1. Planned length limit; 2. Longitudinal distance of the overtaking vehicle, with a reserved safety distance; 3. The farthest travel distance at the sampling time point;

[0128] Therefore, the final longitudinal distance is calculated as follows:

[0129]

[0130] When the longitudinal distance s of the vehicle at the end of the sampling point 纵 When the overtaking distance condition is met, the vehicle makes an overtaking decision; when s 纵 If the overtaking distance requirement is not met, the vehicle will not attempt to overtake and will instead follow the target vehicle.

[0131] Simultaneously, the final acceleration a of the vehicle under the overtaking decision. 纵 0 m / s 2 Therefore, for each sampling time, the final state of the vehicle's longitudinal trajectory is represented as (s 纵 ,v 纵 ,a 纵 ,t sample ).

[0132] Furthermore, based on the vehicle's position and final longitudinal state, the longitudinal trajectory equation of the vehicle is obtained, including:

[0133] Step S410: Set longitudinal constraints, namely...

[0134] The longitudinal constraint of the vehicle is Endpoint longitudinal constraint is Where s(0) is the longitudinal distance of the vehicle at the planned starting point; v(0) is the longitudinal velocity of the vehicle at the planned starting point; acc(0) is the longitudinal acceleration of the vehicle at the planned starting point; s(t sample v(t) represents the longitudinal distance of the vehicle at the end of the planned time slice; sample ) represents the longitudinal velocity of the vehicle at the end of the planned time slice; acc(t) sample () represents the longitudinal acceleration of the vehicle at the end of the planned time slice;

[0135] Step S420: The longitudinal trajectory planning uses a first-to-fifth degree polynomial, denoted as s(t) = a s t 5 +b s t 4 +c s t 3 +d s t 2 +e s t+f s Where s(t) is the longitudinal distance as a function of time, a s ,b s ,c s ,d s ,e s ,f s These are the coefficients of the longitudinal trajectory equation;

[0136] Step S430: The first fifth-degree polynomial can be differentiated with respect to the independent variable to obtain the corresponding vertical objective equation, i.e. Where v(t) is the first derivative of s(t), which is the longitudinal velocity of the vehicle as a function of time; acc(t) is the second derivative of s(t), which is the longitudinal acceleration of the vehicle as a function of time.

[0137] Step S440: Substitute the longitudinal constraints into the longitudinal objective equation, and adjust the coefficients (a) of the longitudinal equation. s ,b s ,c s ,d s ,e s ,f s Solve the equation to obtain the longitudinal trajectory equation.

[0138] Specifically, the longitudinal trajectory generation method is consistent for both following and overtaking decisions. Therefore, the longitudinal trajectory curve can be generated in the same way. Longitudinal constraints are divided into start-point and end-point constraints. The start-point constraints include longitudinal distance, longitudinal velocity, and longitudinal acceleration, while the end-point constraints also include longitudinal distance, longitudinal velocity, and longitudinal acceleration. Therefore, there are a total of six constraints on the longitudinal trajectory. Calculating the longitudinal trajectory requires satisfying all six constraints while simultaneously minimizing the rate of change of longitudinal acceleration. The trajectory function that meets these requirements is a fifth-degree polynomial, denoted as s(t) = at. 5 +bt 4 +ct 3 +dt 2 +et+f, where: s(t) is the longitudinal position as a function of time; a,b,c,d,e,f are the coefficients of the trajectory equation.

[0139] The starting point longitudinal constraint is Endpoint longitudinal constraint is The calculation method for the corresponding vertical constraints of a fifth-degree polynomial can be obtained by differentiating the independent variable, i.e. Where v(t) is the first derivative of s(t), representing the longitudinal velocity of the vehicle as a function of time; acc(t) is the second derivative of s(t), representing the longitudinal acceleration of the vehicle as a function of time; substituting the six longitudinal constraints into the objective equation, and based on the obtained six parametric equation constraints, the parameters (a) can be adjusted. s ,b s ,c s ,d s ,e s ,f s Solve the equation. The corresponding longitudinal trajectory equation can then be obtained.

[0140] Furthermore, based on the vehicle's position and lateral final state, the lateral trajectory equation of the vehicle is obtained, including:

[0141] Step S510: Set lateral constraints, specifically the lateral constraints for the vehicle. Endpoint lateral constraint is Where, l(0) is the lateral distance of the vehicle at the planned starting point; v_lat(0) is the lateral velocity of the vehicle at the planned starting point; acc_lat(0) is the lateral acceleration of the vehicle at the planned starting point; l(t sample v_lat(s) represents the lateral distance of the vehicle at the end of the planned time slice; sample ) represents the lateral velocity of the vehicle at the end of the planned time slice; acc_lat(s sample () represents the lateral acceleration of the vehicle at the end of the planned time slice;

[0142] Step S520: The second-to-fifth degree polynomial is used in the lateral trajectory planning, denoted as l(s) = a l s 5 +b l s 4 +c l s 3 +d l s 2 +e l s+f l Where l(t) is the lateral distance as a function of time, a l ,b l ,c l ,d l ,e l ,f l The coefficients of the transverse trajectory equation;

[0143] Step S530: The second-to-fifth degree polynomial can be differentiated with respect to the independent variable to obtain the corresponding horizontal objective equation, i.e. Where v_lat(s) is the first derivative of l(s), which is the lateral velocity of the vehicle as a function of time; acc_lat(s) is the second derivative of l(t), which is the lateral acceleration of the vehicle as a function of time.

[0144] Step S540: Substitute the lateral constraints into the lateral objective equation, and adjust the coefficients of the lateral equation (a) l ,b l ,c l ,d l ,e l ,f l Solve the equation to obtain the horizontal trajectory.

[0145] Specifically, the lateral constraints are divided into start-point constraints and end-point constraints. The start-point constraints include lateral distance, lateral velocity, and lateral acceleration, while the end-point constraints also include lateral distance, lateral velocity, and lateral acceleration. Therefore, there are a total of six constraints on the lateral trajectory. Calculating the lateral trajectory requires satisfying all six constraints while simultaneously minimizing the rate of change of lateral acceleration using a functional. The trajectory function that meets these requirements is a fifth-degree polynomial, denoted as l(s) = as. 5 +bs 4 +cs 3 +ds 2 +es+f, where: l(t) is the longitudinal position as a function of time; a,b,c,d,e,f are the coefficients of the trajectory equation.

[0146] The starting horizontal constraint is Endpoint lateral constraint is The calculation method for the lateral constraints of a fifth-degree polynomial can be obtained by differentiating the independent variable, i.e. Where v_lat(s) is the first derivative of l(s), representing the lateral velocity of the vehicle as a function of time; acc_lat(s) is the second derivative of l(t), representing the lateral acceleration of the vehicle as a function of time; substituting the six lateral constraints into the objective equation, and based on the obtained six parametric equation constraints, the parameters (a) can be adjusted. l ,b l ,c l ,d l ,e l ,f l Solve the equation. The resulting equation for the transverse trajectory can then be obtained.

[0147] Furthermore, the lateral trajectory equations and longitudinal trajectory equations are matched to obtain a trajectory cluster, including:

[0148] The longitudinal trajectory equation and the longitudinal trajectory equation are matched and calculated by sampling time slices;

[0149] The obtained trajectory clusters are:

[0150]

[0151] In a specific implementation, the function l(s) theoretically allows s to take any value, but generally only a certain time frame is needed, such as the future 6 or 8 seconds. This is a time constraint; calculating too much not only reduces computational efficiency but also lowers reliability. Therefore, t is set... limit ;t sample This refers to the intermediate sampling time. For example, if the sampling time is 8 seconds, sampling can be performed at intervals of 1, 2, 3, 4... 7, 8. sample It can be viewed as a set, t max It is a variable, t limit It is a fixed value, and can be set manually. To a certain extent, t limit The smaller the value, the shorter the computation time and the higher the reliability of the computation.

[0152] Specifically, for each horizontal and vertical trajectory function, matching and calculation are performed using time alignment to obtain a series of trajectory clusters. The trajectory clusters are obtained as follows:

[0153] Furthermore, the trajectory clusters are evaluated, and the optimal trajectory is selected and output, including:

[0154] Step S710: Construct the trajectory evaluation function based on the trajectory cluster as f(trajectory) = w1·cost_objective + w2·cost_smooth + w3·cost_referenceline;

[0155] Step S720: Evaluate each trajectory in the trajectory cluster according to the trajectory evaluation function and obtain a score;

[0156] Step S730: Select the optimal trajectory based on the score and output the optimal trajectory as the final trajectory;

[0157] Where w1, w2, and w3 are weighting coefficients; cost_objective is the distance loss to each obstacle; cost_smooth is the smoothing loss of the final trajectory; and cost_referenceline is the deviation loss from the reference line.

[0158] Specifically, the trajectory evaluation function, as a crucial step in trajectory selection, primarily focuses on the trajectory's ability to avoid obstacles and whether the trajectory itself meets the requirements for ride comfort. Therefore, the loss function is defined as: f(trajectory) = w1·cost_objective + w2·cost_smooth + w3·cost_referenceline, where: w1, w2, and w3 are weighting coefficients; cost_objective is the distance loss to each obstacle; cost_smooth is the smoothing loss of the final trajectory; and cost_referenceline is the deviation loss from the reference line.

[0159] Based on the constructed loss function, each trajectory is evaluated, a score is obtained for each trajectory, and the trajectory that is optimal under this criterion is selected as the final trajectory for this planning.

[0160]

Example 2

[0161] See Figure 3 This embodiment provides a vehicle 100, including: a processor 120, a memory 110, and a program or instruction 111 stored in the memory 110 and executable on the processor. When the program or instruction 111 is executed by the processor 120, it implements a high-speed single-lane driving method with spatiotemporal decoupling as in any embodiment of the present invention.

[0162] This is a schematic diagram of the structure of a vehicle 100 provided in the second embodiment of the present invention. The vehicle 100 includes, for example, a processor 120 and a memory 110 electrically connected to the processor. The memory 110 stores a computer program 111. The processor 120 loads the computer program 111 to implement the high-speed single-lane driving method with spatiotemporal decoupling as in the first embodiment.

[0163] The vehicle 100 of this embodiment is used to implement the high-speed single-lane driving method with spatiotemporal decoupling as in any embodiment of this invention, and therefore has all the beneficial effects of the high-speed single-lane driving method with spatiotemporal decoupling as in any embodiment of this invention, which will not be repeated here.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A spatiotemporally decoupled high-speed single-lane driving method, characterized in that, The high-speed single-lane driving method includes: Step S100: Select the decision boundary based on the vehicle information; Step S200: Make a decision on the target vehicles appearing within the decision boundary; Step S300: Obtain the longitudinal final state (s) of the vehicle based on the target predicted trajectory of the target vehicle. 纵 ,v 纵 ,a 纵 ,t sample ) and transverse final state (l sample ,v 横 ,a 横 ,s sample ); Step S400: Based on the position of the vehicle and the longitudinal final state, obtain the longitudinal trajectory equation of the vehicle; Step S500: Based on the position of the vehicle and the lateral final state, obtain the lateral trajectory equation of the vehicle; Step S600: Match the lateral trajectory equation and the longitudinal trajectory equation to obtain a trajectory cluster; Step S700: Evaluate the trajectory cluster, select and output the optimal trajectory; The decision-making process includes: determining that the target vehicle is a following vehicle and determining that the target vehicle is an overtaking vehicle; s 纵 v represents the longitudinal distance of the vehicle at the end of the sampling point. 纵 a represents the longitudinal velocity of the vehicle at the end of the sampling time slice. 纵 t represents the final acceleration of the vehicle in the longitudinal direction. sample For the sampling time slice, l sample v represents the lateral distance between the sampling points of the vehicle's final state. 横 a represents the lateral velocity of the vehicle at the end of the sampling time slice. 横 s represents the final acceleration of the vehicle in the lateral direction. sample This represents the longitudinal distance between the sampling points of the vehicle's final state.

2. The high-speed single-lane driving method according to claim 1, characterized in that, Step S100 includes: Step S110: Obtain the lane length and lane width of the lane where the vehicle is located; Step S120: Define the length of the decision boundary region as l based on the lane length. lane ; Step S130: Define the width of the decision boundary area as (d) based on the lane width and lateral safety distance. lane +d safe ); Among them, l lane d represents the furthest planned length. lane d represents the lane width. safe This represents the lateral safety distance.

3. The high-speed single-lane driving method according to claim 1, characterized in that, Step S200 includes: The vehicle closest to the autonomous vehicle within the decision boundary is identified as the target vehicle. Step S210: If the target predicted trajectory of the target vehicle is a following vehicle decision, then the target vehicle is determined to be the following vehicle; Step S220: If the target predicted trajectory of the target vehicle is an overtaking decision, then the target vehicle is determined to be the overtaking vehicle.

4. The high-speed single-lane driving method according to claim 1, characterized in that, Step S300 includes: Step S310: Based on the target vehicle's predicted trajectory, obtain the longitudinal distance s of the target vehicle in each time slice in the natural coordinate system. 纵 and speed limit v 纵 ; Step S320: Based on the longitudinal distance and the upper speed limit, construct a sampling range, sample within the sampling range, and obtain the longitudinal final state (s) of the vehicle. 纵 ,v 纵 ,a 纵 ,t sample ); Step S330: Obtain the transverse final state (l) based on the longitudinal final state. sample ,v 横 ,a 横 ,s sample ); Among them, s sample =s 纵 v 横 =0, a 横 =0.

5. The high-speed single-lane driving method according to claim 4, characterized in that, Step S310 includes: Step S311: When the target vehicle is the following vehicle, v 纵 =min(v lane_limit ,(v ego +a ego_max t sample ),v vehicle_follow ); a 纵 ＝0； Among them, v lane_limit V represents the road speed limit. ego a represents the real-time speed of the vehicle. ego_max t represents the maximum acceleration of the vehicle. sample Represents the sampling time slice, v vehicle_follow The longitudinal speed of the following vehicle is represented by s. lane_limit s represents the pre-set feasible planning distance. vehicle_follow s represents the longitudinal end point distance of the following vehicle. vehicle_follow_safe This represents the safe distance to maintain when following another vehicle. Step S312: When the target vehicle is the overtaking vehicle, a 纵 ＝0； Among them, v vehicle_over The longitudinal velocity s of the overtaking vehicle in the sampling time slice represents the longitudinal velocity of the vehicle during that sampling time slice. vehicle_over The longitudinal distance of the overtaking vehicle during the sampling time slice; s vehicle_over_safe The safe distance that the self-driving vehicle needs to maintain from the overtaking vehicle after the overtaking is completed.

6. The high-speed single-lane driving method according to claim 1, characterized in that, Step S400 includes: Step S410: Set longitudinal constraints, namely... The longitudinal constraint of the vehicle is Endpoint longitudinal constraint is Where s(0) is the longitudinal distance of the vehicle at the planned starting point; v(0) is the longitudinal velocity of the vehicle at the planned starting point; acc(0) is the longitudinal acceleration of the vehicle at the planned starting point; s(t) is the longitudinal velocity ...acc(0) is the longitudinal acceleration of the vehicle at the planned starting point; s(t) is the longitudinal velocity of the vehicle at the planned starting point; sample ) represents the longitudinal distance of the vehicle at the end of the planned time slice; v(t) sample ) represents the longitudinal velocity of the vehicle at the end of the planned time slice; acc(t) sample ) represents the longitudinal acceleration of the vehicle at the end of the planned time slice; Step S420: The longitudinal trajectory planning uses a first-to-fifth degree polynomial, denoted as s(t) = a s t 5 +b s t 4 +c s t 3 +d s t 2 +e s t+f s Where s(t) is the longitudinal distance as a function of time, a s ,b s ,c s ,d s ,e s ,f s The coefficients of the longitudinal trajectory equation; Step S430: The first fifth-degree polynomial can be differentiated with respect to the independent variable to obtain the corresponding vertical objective equation, i.e. Where v(t0) is the first derivative of s(t), which is the longitudinal velocity of the vehicle as a function of time; acc(t) is the second derivative of s(t), which is the longitudinal acceleration of the vehicle as a function of time. Step S440: Substitute the longitudinal constraint into the longitudinal objective equation, and adjust the coefficients (a) of the longitudinal equation. s ,b s ,c s ,d s ,e s ,f s Solve the equation to obtain the longitudinal trajectory equation.

7. The high-speed single-lane driving method according to claim 1, characterized in that, Step S500 includes: Step S510: Set lateral constraints, namely... The lateral constraint of the vehicle is Endpoint lateral constraint is Where, l(0) is the lateral distance of the vehicle at the planned starting point; v_lat(0) is the lateral velocity of the vehicle at the planned starting point; acc_lat(0) is the lateral acceleration of the vehicle at the planned starting point; l(t sample v_lat(s) represents the lateral distance of the vehicle at the end of the planned time slice; sample ) represents the lateral velocity of the vehicle at the end of the planned time slice; acc_lat(s) sample ) represents the lateral acceleration of the vehicle at the end of the planned time slice; Step S520: The second-to-fifth degree polynomial is used in the lateral trajectory planning, denoted as l(s) = a l s 5 +b l s 4 +c l s 3 +d l s 2 +e l s+f l Where l(t) is the lateral distance as a function of time, a l ,b l ,c l ,d l ,e l ,f l The coefficients of the transverse trajectory equation; Step S530: The second fifth-degree polynomial can be differentiated with respect to the independent variable to obtain the corresponding horizontal objective equation, i.e. Where v_lat(s) is the first derivative of l(s), which is the lateral velocity of the vehicle as a function of time; acc_lat(s) is the second derivative of l(t), which is the lateral acceleration of the vehicle as a function of time. Step S540: Substitute the lateral constraints into the lateral objective equation, and adjust the coefficients (a) of the lateral equation. l ,b l ,c l ,d l ,e l ,f l Solve the equation to obtain the transverse trajectory.

8. The high-speed single-lane driving method according to claim 1, characterized in that, Step S600 includes: The longitudinal trajectory equation and the longitudinal trajectory equation are matched and calculated using the sampling time slice; The trajectory cluster is obtained as follows: where t max =t limit 。 9. The high-speed single-lane driving method according to claim 1, characterized in that, Step S700 includes: Step S710: Construct a trajectory evaluation function based on the trajectory cluster as f(trajectory) = w1·cost_objective + w2·cost_smooth + w3·cost_referenceline; Step S720: Evaluate each trajectory in the trajectory cluster according to the trajectory evaluation function and obtain a score; Step S730: Select the optimal trajectory based on the score and output the optimal trajectory as the final trajectory; Where w1, w2, and w3 are weighting coefficients; cost_objective is the distance loss to each obstacle; cost_smooth is the smoothing loss of the final trajectory; and cost_referenceline is the deviation loss from the reference line.

10. A vehicle, characterized in that, The vehicle includes: a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein when the program or instructions are executed by the processor, they implement the steps of the spatiotemporally decoupled high-speed single-lane driving method as described in any one of claims 1 to 9.

Images