Vehicle autonomous lane changing method, system, device and storage medium

Abstract

The application discloses a vehicle autonomous lane-changing method, system, device and storage medium, and relates to the technical field of automatic driving. The method comprises the following steps: acquiring road traffic information; constructing a lane-changing obstacle topology relationship of a vehicle according to the road traffic information, wherein the lane-changing obstacle topology relationship comprises current lane information and adjacent lane information; determining a lane-changing collision risk, a self-vehicle benefit before and after lane changing and a self-vehicle surrounding traffic flow disturbance according to the lane-changing obstacle topology relationship; making a lane-changing decision according to the lane-changing collision risk, the self-vehicle benefit before and after lane changing and the self-vehicle surrounding traffic flow disturbance, wherein the lane-changing decision comprises a target lane; based on the lane-changing decision, performing horizontal and longitudinal joint planning constraint to obtain an initial lane-changing trajectory according to the lane-changing obstacle topology relationship and self-vehicle parameters; and performing secondary dynamic planning according to the initial lane-changing trajectory to obtain a smooth trajectory in a lane-changing process. The application can improve lane-changing efficiency and safety and reasonably plan a lane-changing trajectory.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a method, system, device, and storage medium for autonomous lane changing of a vehicle. Background Technology

[0002] Autonomous driving technology, as an important research topic in the transportation field, is considered one of the key solutions for improving driving safety and traffic efficiency. With the continuous advancement of IoT technology, autonomous driving technology has also developed rapidly. In driving behavior, following other vehicles and lane changing are the two most common behaviors. Compared to following other vehicles, autonomous lane changing can improve traffic efficiency and alleviate localized traffic congestion to some extent, but it also brings greater safety hazards. Inappropriate lane-changing decisions can directly lead autonomous vehicles into dangerous driving situations. Therefore, research on lane-changing decisions in autonomous driving is crucial. Furthermore, how to effectively execute lane-changing decisions and generate appropriate lane-changing paths, enabling vehicles to complete the lane-changing process quickly, comfortably, and safely, is also a key focus of lane-changing behavior research.

[0003] Traditional lane-changing decision-making processes often idealize lane-changing behavior, failing to consider the impact of autonomous vehicle lane-changing actions on local traffic flow. Furthermore, lane-changing models only consider simple traffic scenarios, ignoring complex interactions and game theory among multiple vehicles, thus limiting lane-changing efficiency. Specifically, lane-changing may disrupt local traffic flow, affecting traffic efficiency and passenger experience. In addition, traditional autonomous driving lane-changing decision-making and planning are decoupled horizontally and vertically, failing to effectively utilize lane-changing decision information, resulting in planned lane-changing trajectories that do not meet actual needs. Summary of the Invention

[0004] This invention aims to solve at least one of the technical problems existing in the prior art. To this end, this invention proposes a method, system, device, and storage medium for autonomous lane changing of vehicles, which can improve lane changing efficiency and safety, and rationally plan lane changing trajectories.

[0005] On one hand, embodiments of the present invention provide a method for autonomous lane changing for vehicles, comprising the following steps:

[0006] Obtain road traffic information;

[0007] The topological relationship of lane-changing obstacles for vehicles is constructed based on the road traffic information, wherein the topological relationship of lane-changing obstacles includes current lane information and adjacent lane information;

[0008] Based on the topological relationship of the lane-changing obstacles, determine the lane-changing collision risk, the vehicle's benefits before and after lane changing, and the traffic flow disturbance around the vehicle.

[0009] Lane-changing decisions are made based on the lane-changing collision risk, the vehicle's benefits before and after the lane change, and the traffic flow disturbance around the vehicle, wherein the lane-changing decision includes the target lane.

[0010] Based on the lane-changing decision, the initial lane-changing trajectory is obtained by performing lateral and longitudinal joint planning constraints according to the topological relationship of the lane-changing obstacles and the vehicle parameters.

[0011] The smooth trajectory during the lane-changing process is obtained by performing secondary dynamic programming based on the initial lane-changing trajectory.

[0012] According to some embodiments of the present invention, constructing the lane-changing obstacle topology relationship of the vehicle based on the road traffic information includes the following steps:

[0013] The current lane information is filtered based on the road traffic information, and the current lane topology is constructed based on the current lane information. The current lane topology includes the information of the first obstacle in front of and behind the vehicle in the current lane, the information of the first leading vehicle, and the information of the first following vehicle.

[0014] Adjacent lane information is filtered based on the road traffic information, and adjacent lane vehicle topology is constructed based on the adjacent lane information, wherein the adjacent lane vehicle topology includes lane change gap information, second lead vehicle information and second follower vehicle information of the adjacent lane;

[0015] The lane-changing obstacle topology of the vehicle is determined based on the current lane topology and the adjacent lane vehicle topology.

[0016] According to some embodiments of the present invention, determining the lane-changing collision risk, the vehicle's benefits before and after the lane-changing obstacle topology, and the traffic flow disturbance around the vehicle based on the lane-changing obstacle topology includes the following steps:

[0017] Construct a potential field of vehicle lane-changing risk in front of and behind the vehicle in the target lane-changing gap based on the topological relationship of the lane-changing obstacles.

[0018] Based on the lane change risk potential field, the lane change collision risk of each spatiotemporal trajectory point is obtained by spatiotemporal trajectory prediction.

[0019] Based on the topological relationship of the lane-changing obstacles, determine the preset speed of the lead vehicle in the target lane, the speed of the vehicle in the current lane, the acceleration of the following vehicle before and after the lane change, the acceleration of the lead vehicle after the lane change, the acceleration of the vehicle after the lane change, the size of the target lane change gap, and the remaining space in front of the vehicle, and calculate the traffic flow disturbance around the vehicle.

[0020] The vehicle's gains before and after a lane change are determined based on its current speed in the current lane and the predicted speed in the target lane.

[0021] According to some embodiments of the present invention, the lane-changing decision based on the lane-changing collision risk, the vehicle's benefits before and after the lane change, and the traffic flow disturbance around the vehicle includes the following steps:

[0022] The lane change collision risk of the aforementioned spatiotemporal trajectory points is compared with the lane change prediction to determine whether to change lanes;

[0023] The vehicle's gains before and after lane changing are compared with the traffic flow disturbances around the vehicle to determine whether to change lanes.

[0024] According to some embodiments of the present invention, the traffic flow disturbance around the vehicle is calculated using the following formula:

[0025]

[0026] Where w1, w2, w3, and w4 are preset weighting coefficients, v target v is the preset speed of the leading vehicle in the target lane. ego Let a be the vehicle's speed in the current lane. follow Let a be the acceleration of the following vehicle after it changes lanes in the target lane. follow The acceleration of the following vehicle in the target lane before it changes lanes. Acceleration of the lead vehicle after the lead vehicle in the target lane changes lanes. For the acceleration of the car after changing lanes, gap target For the target lane change gap size, gap ego This represents the remaining space in front of the vehicle. Weighting coefficients are used to quantify the balance between speed advantage and traffic impact on other vehicles. They also provide drivers with various driving styles, making lane-changing behavior in autonomous vehicles more aligned with driver habits.

[0027] According to some embodiments of the present invention, the step of obtaining the initial lane-changing trajectory based on the lane-changing decision and the lateral and longitudinal joint planning constraints according to the topological relationship of the lane-changing obstacles and the vehicle parameters includes the following steps:

[0028] Based on the longitudinal trajectory planning of the previous time node, the lateral trajectory curvature constraint is obtained by combining the vehicle speed and the wheel friction ellipse algorithm.

[0029] Based on the lateral planning results, longitudinal planning is performed according to the associated trajectory obstacle information in the topological relationship of the lane-changing obstacles to obtain the longitudinal trajectory velocity curvature constraint.

[0030] The lateral trajectory and longitudinal trajectory are defined by the lateral trajectory curvature constraint and the longitudinal trajectory velocity curvature constraint, respectively, to obtain the initial lane-changing trajectory.

[0031] According to some embodiments of the present invention, the step of obtaining a smooth trajectory during the lane-changing process by performing secondary dynamic programming based on the initial lane-changing trajectory includes the following steps:

[0032] Based on the SL coordinate system, the road space of the initial lane-changing trajectory is discretized into an ordered lattice at intervals, and the lattice nodes are marked as level nodes.

[0033] Iterate through each level node and perform collision detection between the current node and obstacles;

[0034] If the collision detection result indicates that a collision has occurred, the level node is discarded.

[0035] When the collision detection result is no collision, the loss value of each predecessor node of the current level node is calculated, wherein the loss value includes the lateral offset of the predecessor node from the target reference line, the lateral rate of change, and the lateral deviation of the endpoint position.

[0036] The predecessor node with the smallest loss value is taken as the parent node of the current level node;

[0037] Traverse until the last level node, find the parent node through the linked list, and continue until the initial node is found to obtain the smooth trajectory.

[0038] On the other hand, embodiments of the present invention also provide a vehicle autonomous lane-changing system, comprising:

[0039] The first module is used to obtain road traffic information;

[0040] The second module is used to construct the lane-changing obstacle topology relationship of the vehicle based on the road traffic information, wherein the lane-changing obstacle topology relationship includes current lane information and adjacent lane information;

[0041] The third module is used to determine the lane change collision risk, the vehicle's benefits before and after lane change, and the traffic flow disturbance around the vehicle based on the topological relationship of the lane change obstacles.

[0042] The fourth module is used to make lane-changing decisions based on the lane-changing collision risk, the vehicle's benefits before and after the lane-changing, and the traffic flow disturbance around the vehicle, wherein the lane-changing decision includes the target lane.

[0043] The fifth module is used to obtain the initial lane-changing trajectory by performing lateral and longitudinal joint planning constraints based on the lane-changing decision, the topological relationship of the lane-changing obstacles, and the vehicle parameters.

[0044] The sixth module is used to perform secondary dynamic programming based on the initial lane-changing trajectory to obtain a smooth trajectory during the lane-changing process.

[0045] On the other hand, embodiments of the present invention also provide a vehicle autonomous lane-changing device, comprising:

[0046] At least one processor;

[0047] At least one memory for storing at least one program;

[0048] When the at least one program is executed by the at least one processor, the at least one processor implements the vehicle autonomous lane-changing method as described above.

[0049] On the other hand, embodiments of the present invention also provide a computer-readable storage medium storing computer-executable instructions for causing a computer to execute the vehicle autonomous lane-changing method described above.

[0050] The technical solution of this invention has at least one of the following advantages or beneficial effects: It constructs a vehicle lane-changing obstacle topology relationship based on road traffic information, including current lane information and adjacent lane information; then, it determines the lane-changing collision risk, vehicle benefits before and after lane changing, and traffic flow disturbances around the vehicle based on the lane-changing obstacle topology relationship; and makes a lane-changing decision based on the lane-changing collision risk, vehicle benefits before and after lane changing, and traffic flow disturbances around the vehicle. Based on the lane-changing decision, it performs lateral and longitudinal joint planning constraints based on the lane-changing obstacle topology relationship and vehicle parameters to obtain an initial lane-changing trajectory; then, it performs secondary dynamic planning based on the initial lane-changing trajectory to obtain a smooth trajectory during the lane-changing process. In the lane-changing decision-making process, it considers lane-changing collision risk, vehicle benefits before and after lane changing, and traffic flow disturbances around the vehicle, improving lane-changing efficiency and safety. The lateral and longitudinal joint planning constraints based on the lane-changing decision and the secondary planning to obtain a smooth trajectory enable reasonable planning of lane-changing trajectories. Attached Figure Description

[0051] Figure 1 This is a flowchart of the autonomous lane-changing method for vehicles provided in an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of the vehicle autonomous lane-changing device provided in an embodiment of the present invention. Detailed Implementation

[0053] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar originals or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0054] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, left, right, etc., are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0055] In the description of this invention, the use of terms such as "first," "second," etc., is merely for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of technical features indicated, or implicitly indicating the order of the technical features indicated.

[0056] This invention provides a method for autonomous lane changing for vehicles. This method can be applied to a terminal, a server, or software running on either a terminal or server. The terminal can be a tablet, laptop, desktop computer, or vehicle controller, but is not limited to these. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The autonomous lane changing method of this invention can also be applied to a system composed of a vehicle terminal and a server.

[0057] Reference Figure 1 This invention provides a method for autonomous lane changing for vehicles, including but not limited to the following steps:

[0058] Step S110: Obtain road traffic information;

[0059] Step S120: Construct the topological relationship of lane-changing obstacles for vehicles based on road traffic information, wherein the topological relationship of lane-changing obstacles includes current lane information and adjacent lane information;

[0060] Step S130: Determine the lane change collision risk, the vehicle's benefits before and after lane change, and the traffic flow disturbance around the vehicle based on the topological relationship of lane change obstacles.

[0061] Step S140: Make a lane change decision based on the lane change collision risk, the vehicle's benefits before and after the lane change, and the traffic flow disturbance around the vehicle. The lane change decision includes the target lane.

[0062] Step S150: Based on the lane change decision, the initial lane change trajectory is obtained by performing lateral and longitudinal joint planning constraints according to the topological relationship of lane change obstacles and the vehicle parameters.

[0063] Step S160: Perform secondary dynamic programming based on the initial lane-changing trajectory to obtain the smooth trajectory during the lane-changing process.

[0064] In some embodiments of step S110, road traffic information can be obtained from an upper-layer perception server, which interacts with cameras, infrared sensors, or vehicles on the road to obtain various vehicle information and vehicle conditions on the road traffic network.

[0065] In some embodiments of step S120, the step of constructing the topological relationship of lane-changing obstacles for vehicles based on road traffic information includes, but is not limited to, the following steps:

[0066] Step S210: Filter the current lane information according to the road traffic information, and construct the current lane topology relationship according to the current lane information. The current lane topology relationship includes the information of the first obstacle in front of and behind the vehicle in the current lane, the information of the first leading vehicle, and the information of the first following vehicle.

[0067] Step S220: Filter adjacent lane information based on road traffic information, and construct adjacent lane vehicle topology relationship based on adjacent lane information. The adjacent lane vehicle topology relationship includes lane change gap information, second lead vehicle information and second follow vehicle information of adjacent lanes.

[0068] Step S230: Determine the lane-changing obstacle topology relationship of the vehicle based on the current lane topology relationship and the vehicle topology relationship of adjacent lanes.

[0069] Specifically, in step S210, relevant information such as key obstacles in front of and behind the vehicle in the current lane is filtered from the acquired road traffic information to construct the current lane topology. It is understood that, for the vehicle itself, other dynamically moving vehicles or barriers on the road can all be considered obstacles. The current lane topology includes information on the first obstacle in front of and behind the vehicle in the current lane, information on the first leading vehicle, and information on the first following vehicle.

[0070] In step S220, the lane-changing gap between adjacent vehicles in the obtained road traffic information is filtered out, and the vehicles that affect the interaction between the vehicles can be determined based on the appropriate lane-changing gap.

[0071] For the current lane topology, its topological structure is as follows:

[0072]

[0073] The data structure `front_obstacle:car_lead` represents the information of the first leading vehicle in the current lane, including but not limited to the speed and position of the leading vehicle; the data structure `ear_obstacle:car_follow` represents the information of the first following vehicle, including but not limited to the speed and position of the following vehicle; and the data structure `key_gaps_list` represents the list of lane change gaps.

[0074] The data structure for the target gap, i.e., the topological relationship between vehicles in adjacent lanes, is as follows:

[0075]

[0076] Here, car_lead represents the information of the second lead vehicle during lane change intervals; car_follow represents the information of the second following vehicle.

[0077] In some embodiments of step S130, the step of determining the lane-changing collision risk, the vehicle's benefits before and after the lane-changing obstacle topology, and the traffic flow disturbance around the vehicle, including but not limited to the following steps:

[0078] Step S310: Construct the potential field of lane-changing risk for vehicles in front of and behind the vehicle in the target lane-changing gap based on the topological relationship of lane-changing obstacles.

[0079] Step S320: Based on the lane change risk potential field, perform spatiotemporal trajectory prediction to obtain the lane change collision risk of each spatiotemporal trajectory point;

[0080] Step S330: Determine the preset speed of the lead vehicle in the target lane, the speed of the vehicle in the current lane, the acceleration of the following vehicle before and after the lane change of the following vehicle in the target lane, the acceleration of the lead vehicle after the lane change of the lead vehicle in the target lane, the acceleration of the vehicle after the lane change, the size of the target lane change gap, and the remaining space in front of the vehicle based on the topological relationship of the lane change obstacle, and calculate the traffic flow disturbance around the vehicle.

[0081] Step S340: Determine the vehicle's benefits before and after lane changing based on the vehicle's speed in the current lane and the predicted vehicle speed in the target lane.

[0082] Specifically, in steps S310 and S320, the lane change collision risk analysis process is as follows:

[0083] Based on various information in the topological relationship of lane-changing obstacles, the safe distance is determined based on the lane-changing risk potential field of vehicles in front of and behind the target lane-changing gap. The safe distance is calculated as shown in formula (3):

[0084] safe_dis = abs(v ego -v target (3) × 2.5 + μ

[0085] Among them, v ego v is the vehicle's speed. target The target lane change interval is defined by the vehicle speeds before and after the lane change, and μ is the basic safety distance, typically set to 3m.

[0086] Set the extension length of the potential field for lane changing risk, as shown in formula (4):

[0087] RPT_length = abs(v ego -(v target +2.75×a target (4) × 5.5

[0088] Among them, a target The instantaneous acceleration of vehicles before and after the lane change gap is considered.

[0089] For the vehicle and the target vehicle during the lane change process, a spatiotemporal trajectory of a certain time length (e.g., 8 seconds) can be selected, and the spatiotemporal trajectory can be discrete with a resolution of unit time length (e.g., 0.5 seconds). The relative distances between the front and rear of the vehicle and the target vehicle projected on the current reference line at the same time scale are calculated respectively, thereby calculating the collision risk of the front and rear of the vehicle. The minimum value of the collision risk of the front and rear of the vehicle at all spatiotemporal trajectory points is taken as the final lane change collision risk.

[0090] In step S330, the traffic flow disturbance around the vehicle is used to characterize the disturbance to the leading and following vehicles in the target lane after lane change. The traffic flow disturbance R around the vehicle is calculated using the following formula:

[0091]

[0092] Where w1, w2, w3, and w4 are preset weighting coefficients, v target v is the preset speed of the leading vehicle in the target lane. ego Let a be the vehicle's speed in the current lane. follow Let a be the acceleration of the following vehicle after it changes lanes in the target lane. follow The acceleration of the following vehicle in the target lane before it changes lanes. Acceleration of the lead vehicle after the lead vehicle in the target lane changes lanes. For the acceleration of the car after changing lanes, gap target For the target lane change gap size, gap ego This refers to the remaining space in front of the vehicle.

[0093] In step S340, the right-hand side λ of the lane-changing model represents the cross-lane speed gain. When the cross-lane speed gain is significant, it reduces the lane-changing resistance generated by changing lanes towards the lane with the significant gain. The lane-changing decision system will be more inclined to change lanes to obtain the substantial lane-changing gain after the change. The vehicle's gain λ before and after the lane change is calculated using the following formula:

[0094] λ=w(v' target -v ego (6)

[0095] Where w is the speed gain coefficient after crossing lanes, used to balance the benefits of crossing lanes and changing lanes in adjacent lanes; v′ target -v ego The speed gain after changing lanes, i.e., v ego v′ represents the vehicle's speed in the current lane. target This indicates the predicted vehicle speed in the target lane.

[0096] In some embodiments of step S140, the step of making a lane-changing decision based on the lane-changing collision risk, the vehicle's benefits before and after the lane change, and the traffic flow disturbance around the vehicle includes, but is not limited to, the following steps:

[0097] Step S410: Compare the lane change collision risk of the spatiotemporal trajectory point with the lane change prediction to determine whether to change lanes;

[0098] Step S420: Compare the benefits of the vehicle before and after lane changing with the traffic flow disturbance around the vehicle to determine whether to change lanes.

[0099] Specifically, in step S410, when the vehicle is in the pre-lane-change preparation and lane-change process states, the lane-change collision risk value of the vehicle's spatiotemporal trajectory point on the lane line in the lane-change direction is used as the basis for judging the lane-change safety. If the vehicle is currently in the lane-change preparation state, and P < 0.3, the lane change is considered safe, and the vehicle enters the lane-change in progress state; if the vehicle is in the lane-change process and P > 0.3 + η, the vehicle enters the lane-change withdrawal state and returns to the original lane, where P is the lane-change collision risk calculated in step S320, and η is the lane-change inertia.

[0100] In step S420, the minimum lane-changing inertia and lane-changing tolerance factor are further combined to determine whether to change lanes. The specific criteria for this determination are as follows:

[0101] R>ξ+ρ(t)+λ;(7)

[0102] Where R is the traffic flow disturbance around the vehicle calculated by formula (5); λ is the vehicle's benefit before and after lane changing calculated by formula (6); ξ is the minimum lane changing inertia; ρ(t) is a time-dependent inertia function, representing the tolerance for the current lane, as follows:

[0103]

[0104] ρ(t) decreases exponentially over a certain period of time. Coefficient a represents the basic lane-changing inertia. Coefficient b is adjusted to suit different personalities of drivers and passengers. Coefficient c represents the general tolerance time. Being in a lane-changing tendency state for a long time will reduce lane-changing inertia and make it easier to make lane-changing decisions.

[0105] In some embodiments of step S150, the step of obtaining the initial lane-changing trajectory based on lane-changing decision and by performing lateral and longitudinal joint planning constraints according to the topological relationship of lane-changing obstacles and the vehicle parameters includes, but is not limited to, the following steps:

[0106] Step S510: Based on the longitudinal trajectory planning of the previous time node, lateral planning is performed by combining the vehicle speed and wheel friction ellipse algorithm to obtain the lateral trajectory curvature constraint.

[0107] Step S520: Based on the lateral planning results, longitudinal planning is performed according to the associated trajectory obstacle information in the topological relationship of lane-changing obstacles to obtain the longitudinal trajectory velocity curvature constraint.

[0108] Step S530: Define the lateral trajectory and longitudinal trajectory according to the lateral trajectory curvature constraint and the longitudinal trajectory velocity curvature constraint respectively to obtain the initial lane-changing trajectory.

[0109] Specifically, the vehicle speed information is obtained from the vehicle speed sensor and combined with the wheel friction ellipse to perform lateral trajectory curvature constraints, as follows:

[0110] Since the tire turning angle is not very large at high speeds, and we can ignore the lateral force caused by the tire turning angle, the main factor affecting lateral acceleration is the radius of curvature of the trajectory, resulting in the following lateral force:

[0111] F y =v 2 ·k;(9)

[0112] Among them, F y ν is the lateral force; v is the longitudinal vehicle speed; k is the radius of curvature.

[0113] By restricting the friction ellipse to a friction circle, we can obtain:

[0114]

[0115] Where R is the radius of the friction circle, F x_max This represents the maximum longitudinal force of a single wheel.

[0116] After a round of lane-changing decisions, based on the longitudinal planning trajectory of the previous time node, the dynamic model of formulas (9) and (10) can be used to converge the SL hard boundary in the current lateral planning, thereby optimizing the SL solution space of the lateral planning, and applying curvature constraints to the lateral trajectory to plan a safe and reasonable trajectory.

[0117] Secondly, based on the results of lateral planning, the longitudinal trajectory is calculated by combining the information of the associated trajectory obstacles in the topological relationship of lane-changing obstacles, and the longitudinal trajectory is subject to speed constraints.

[0118] Then, based on the Frenet coordinate system, obstacle and road boundary information are projected onto the reference line to construct the lateral planning solution space SL_boundary. For lane-changing scenarios of autonomous vehicles on structured roads, the boundary constraints of lane-changing planning are analyzed. Based on the longitudinal planning results of the previous time node, the vehicle's coarse TS trajectory for a future period can be obtained. Combined with the velocity attributes of dynamic obstacles at the current time node, the boundary constraints of dynamic obstacles can be obtained.

[0119] Based on the calculated lateral trajectory curvature constraints, the lateral and longitudinal trajectories are defined, and a lane-changing trajectory that satisfies the current constraints is obtained by combining the trajectory planning algorithm.

[0120] In some embodiments of step S160, the step of performing secondary dynamic programming based on the initial lane-changing trajectory to obtain a smooth trajectory during the lane-changing process includes, but is not limited to, the following steps:

[0121] Step S610: Based on the SL coordinate system, the road space of the initial lane-changing trajectory is discretized into an ordered lattice at intervals, and the lattice nodes are marked as level nodes.

[0122] Step S620: Iterate through each level node and perform collision detection between the current node and obstacles;

[0123] Step S630: If the collision detection result indicates a collision has occurred, then the level node is discarded.

[0124] Step S640: When the collision detection result is no collision, calculate the loss value of each predecessor node of the current level node, wherein the loss value includes the lateral offset of the predecessor node from the target reference line, the lateral rate of change, and the lateral deviation of the endpoint position.

[0125] Step S650: Select the predecessor node with the smallest loss value as the parent node of the current level node;

[0126] Step S660: Traverse until the last level node, find the parent node through the linked list, and continue until the initial node is found to obtain the smooth trajectory.

[0127] Specifically, firstly, an SL coordinate system is constructed, and the current road space is discretized into an ordered matrix at intervals and initialized. Each level node is iterated through, and collision detection is performed on the current level node in conjunction with obstacle information. If a collision occurs, the node is discarded; otherwise, the cost (loss value) of each predecessor node of the current node is calculated. The cost function includes the lateral offset of the current level node from the target reference line, the lateral rate of change, and the lateral deviation from the endpoint position. The predecessor node with the smallest cost is selected as the parent node. The child nodes of all levels are traversed, and the child node with the smallest cost in the last level is selected. The parent node is found through a linked list until the initial node is found, completing the backtracking of the entire trajectory.

[0128] Furthermore, the road space of the initial lane-changing trajectory is discretized into an ordered lattice at regular intervals, and the penalty function is designed as follows:

[0129] J = J smooth +J obs +J ref_diff (11)

[0130] Lateral smoothness J in the penalty function smooth as follows:

[0131] J smooth =w l′ ·∫l′(s) 2 ds;(12)

[0132] Where l'(s) is the first derivative of the lateral deviation of the current trajectory from the current reference line, w l′ These are the weighting coefficients.

[0133] The penalty function specifies the safe distance J between the vehicle and the obstacle at the current level node. obs ,as follows:

[0134]

[0135] Where w is the weighting coefficient, K max The maximum cost of the current node is given by d, where dis_obs is the distance between the vehicle and the obstacle, and d is the distance between the vehicle and the obstacle. max d min These are the maximum and minimum safe distances, respectively.

[0136] The deviation J between the current level node of the vehicle and the target reference line in the penalty function. ref_diff as follows:

[0137] J ref_diff =w diff_l ·∫diff_l target(s) 2 ds;(14)

[0138] Among them, ∫diff_l target This refers to the lateral deviation between the vehicle and the target lane reference line.

[0139] In the quadratic lateral programming process, the cost function is as follows:

[0140] J = J cost_1 +J cost_2 +J cost_3 (15)

[0141] The first term of the cost function is as follows:

[0142] J cost_1 =w l ·∫l(s) 2 ds+w l′ ·∫l′(s) 2 ds+w l″ ·∫l″(s) 2 ds+w l″′ ·∫l″′(s) 2 ds;(16)

[0143] Where l, l′, l″, and l″′ are the lateral deviation relative to the reference line and the first, second, and third derivatives of l with respect to s, respectively; w l The weighting factor for the lateral deviation from the reference line; w l′ w l″ and w l″′ These are the weighting coefficients of the first, second, and third derivatives of l with respect to s, respectively.

[0144] The second term of the cost function is as follows:

[0145] J cost_2 =w obs ·∫(l(s)-0.5*(l soft (s) left +l soft (s) right )) 2 ds;(17)

[0146] Among them, w obs Represents the weighting coefficient, l soft (s) left and l soft (s) right These represent the upper and lower soft boundaries of sl_boundary, respectively.

[0147] The third term of the cost function is as follows:

[0148]

[0149] Among them, l endref 、l′ endref and l″ endref These represent constraints on the endpoint state.

[0150] The constraints for the quadratic programming of the lateral trajectory planning are as follows:

[0151]

[0152] Among them, l left l right These represent the left and right boundaries of sl_boundary, respectively; l′ represents the rate of change of the lateral and longitudinal displacements about the reference line; l” i l″′ represents lateral acceleration; l″′ represents lateral jerk.

[0153] A longitudinal trajectory planning algorithm is designed based on solving the initial solution using longitudinal dynamic programming and smoothing the initial solution using quadratic programming. The penalty function during the quadratic longitudinal programming process is as follows:

[0154] J = J v +J acc +J obs (20)

[0155] In the penalty function, assuming the current parent-child node moves at a constant speed, then:

[0156] J v =w v ·(v i -v ref ) 2 ;(twenty one)

[0157] Among them, v i -v ref w is the difference between the actual vehicle speed and the desired vehicle speed. v These are the weighting coefficients.

[0158] In the penalty function, the acceleration of the current node is limited as follows:

[0159]

[0160] Among them, w acc v is the weighting coefficient. i -v i-1 This represents the speed difference between the two nodes.

[0161] In the penalty function, set J obs as follows:

[0162]

[0163] Among them, w obs These are the weighting coefficients.

[0164] Set the constraints for the quadratic programming of the longitudinal trajectory planning as follows:

[0165]

[0166] Where s, s′, s″, and s″′ represent the position, velocity, acceleration, and jerk at time t, respectively.

[0167] In this embodiment of the invention, the impact of lane changing on local traffic flow is introduced, considering lane changing decisions not only from the perspective of vehicle benefits but also from the perspective of local traffic efficiency. Lane changing tolerance is also added to increase the rationality of lane changing decisions. Based on traditional decoupled lateral and longitudinal planning, a lateral and longitudinal planning method that combines lane changing decision information is proposed. In lateral planning, a discrete solution space is constructed based on the SL coordinate system, and an initial solution is calculated using a dynamic programming algorithm. The trajectory curvature is constrained and the current SL solution space is optimized by combining the decision process and the longitudinal planning results from the previous time point, performing nonlinear optimization. Subsequently, longitudinal planning is performed, projecting obstacles into the ST space based on the lateral trajectory, determining the distribution of the solution space based on the lane changing decision results, and finally planning a safe, efficient, and comfortable lateral and longitudinal trajectory. The lateral and longitudinal joint planning algorithm of this embodiment can plan smooth lane changing trajectories and comfortable speed trajectories in real time. The lane changing decision process of this embodiment can improve lane changing efficiency, safety, and effectiveness.

[0168] On the other hand, embodiments of the present invention also provide a vehicle autonomous lane-changing system, comprising:

[0169] The first module is used to obtain road traffic information;

[0170] The second module is used to construct the topological relationship of lane-changing obstacles for vehicles based on road traffic information. The topological relationship of lane-changing obstacles includes current lane information and adjacent lane information.

[0171] The third module is used to determine the collision risk of lane changing, the benefits of the vehicle before and after lane changing, and the traffic flow disturbance around the vehicle based on the topological relationship of lane changing obstacles.

[0172] The fourth module is used to make lane-changing decisions based on the collision risk of lane changing, the benefits of the vehicle before and after lane changing, and the traffic flow disturbance around the vehicle. The lane-changing decision includes the target lane.

[0173] The fifth module is used to obtain the initial lane-changing trajectory based on lane-changing decision-making, by performing lateral and longitudinal joint planning constraints according to the topological relationship of lane-changing obstacles and the vehicle parameters.

[0174] The sixth module is used to perform secondary dynamic programming based on the initial lane-changing trajectory to obtain a smooth trajectory during the lane-changing process.

[0175] It is understood that the content of the above-described autonomous lane-changing method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above-described autonomous lane-changing method embodiments, and the beneficial effects achieved are also the same as those achieved in the above-described autonomous lane-changing method embodiments.

[0176] Reference Figure 2 , Figure 2 This is a schematic diagram of a vehicle autonomous lane-changing device according to an embodiment of the present invention. The vehicle autonomous lane-changing device of this embodiment includes one or more control processors and a memory. Figure 2 The example consists of a control processor and a memory.

[0177] The control processor and memory can be connected via a bus or other means. Figure 2 Taking the example of a connection between China and Israel via a bus.

[0178] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the control processor, and these remote memories can be connected to the vehicle's autonomous lane-changing device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0179] Those skilled in the art will understand that Figure 2 The device structure shown does not constitute a limitation on the vehicle's autonomous lane-changing device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0180] The non-transient software program and instructions required to implement the autonomous lane-changing method applied to the autonomous lane-changing device in the above embodiments are stored in the memory. When executed by the control processor, the autonomous lane-changing method applied to the autonomous lane-changing device in the above embodiments is executed.

[0181] Furthermore, one embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions that are executed by one or more control processors, causing the one or more control processors to perform the vehicle autonomous lane-changing method in the above method embodiment.

[0182] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0183] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A method for autonomous lane changing for vehicles, characterized in that, Includes the following steps: Obtain road traffic information; The topological relationship of lane-changing obstacles for vehicles is constructed based on the road traffic information, wherein the topological relationship of lane-changing obstacles includes current lane information and adjacent lane information; Based on the topological relationship of the lane-changing obstacles, determine the lane-changing collision risk, the vehicle's benefits before and after lane changing, and the traffic flow disturbance around the vehicle. Lane-changing decisions are made based on the lane-changing collision risk, the vehicle's benefits before and after the lane change, and the traffic flow disturbance around the vehicle, wherein the lane-changing decision includes the target lane. Based on the lane-changing decision, the initial lane-changing trajectory is obtained by performing lateral and longitudinal joint planning constraints according to the topological relationship of the lane-changing obstacles and the vehicle parameters. The smooth trajectory during the lane-changing process is obtained by performing secondary dynamic programming based on the initial lane-changing trajectory.

2. The autonomous lane-changing method for vehicles according to claim 1, characterized in that, The process of constructing the lane-changing obstacle topology relationship of the vehicle based on the road traffic information includes the following steps: The current lane information is filtered based on the road traffic information, and the current lane topology is constructed based on the current lane information. The current lane topology includes the information of the first obstacle in front of and behind the vehicle in the current lane, the information of the first leading vehicle, and the information of the first following vehicle. Adjacent lane information is filtered based on the road traffic information, and adjacent lane vehicle topology is constructed based on the adjacent lane information, wherein the adjacent lane vehicle topology includes lane change gap information, second lead vehicle information and second follower vehicle information of the adjacent lane; The lane-changing obstacle topology of the vehicle is determined based on the current lane topology and the adjacent lane vehicle topology.

3. The autonomous lane-changing method for vehicles according to claim 2, characterized in that, The process of determining the lane-changing collision risk, vehicle benefits before and after lane changing, and traffic flow disturbance around the vehicle based on the topological relationship of the lane-changing obstacles includes the following steps: Construct a potential field of vehicle lane-changing risk in front of and behind the vehicle in the target lane-changing gap based on the topological relationship of the lane-changing obstacles. Based on the lane change risk potential field, the lane change collision risk of each spatiotemporal trajectory point is obtained by spatiotemporal trajectory prediction. Based on the topological relationship of the lane-changing obstacles, determine the preset speed of the lead vehicle in the target lane, the speed of the vehicle in the current lane, the acceleration of the following vehicle before and after the lane change, the acceleration of the lead vehicle after the lane change, the acceleration of the vehicle after the lane change, the size of the target lane change gap, and the remaining space in front of the vehicle, and calculate the traffic flow disturbance around the vehicle. The vehicle's gains before and after a lane change are determined based on its current speed in the current lane and the predicted speed in the target lane.

4. The autonomous lane-changing method for vehicles according to claim 3, characterized in that, The process of making a lane-changing decision based on the lane-changing collision risk, the vehicle's benefits before and after the lane-changing, and the traffic flow disturbance around the vehicle includes the following steps: The lane change collision risk of the aforementioned spatiotemporal trajectory points is compared with the lane change prediction to determine whether to change lanes; The vehicle's gains before and after lane changing are compared with the traffic flow disturbances around the vehicle to determine whether to change lanes.

5. The autonomous lane-changing method for vehicles according to claim 4, characterized in that, The traffic flow disturbance around the vehicle is calculated using the following formula: ； in, , , and All are preset weighting coefficients. The preset speed for the lead vehicle in the target lane. The speed of the vehicle in the current lane. The acceleration of the following vehicle after it changes lanes in the target lane. The acceleration of the following vehicle in the target lane before it changes lanes. Acceleration of the lead vehicle after the lead vehicle in the target lane changes lanes. The acceleration of the car after changing lanes, For the target lane change clearance size, This refers to the remaining space in front of the vehicle.

6. The autonomous lane-changing method for vehicles according to claim 5, characterized in that, The process of obtaining the initial lane-changing trajectory based on the lane-changing decision, according to the topological relationship of the lane-changing obstacles and the vehicle parameters, includes the following steps: Based on the longitudinal trajectory planning of the previous time node, the lateral trajectory curvature constraint is obtained by combining the vehicle speed and the wheel friction ellipse algorithm. Based on the lateral planning results, longitudinal planning is performed according to the associated trajectory obstacle information in the topological relationship of the lane-changing obstacles to obtain the longitudinal trajectory velocity curvature constraint. The lateral trajectory and longitudinal trajectory are defined by the lateral trajectory curvature constraint and the longitudinal trajectory velocity curvature constraint, respectively, to obtain the initial lane-changing trajectory.

7. The autonomous lane-changing method for vehicles according to claim 6, characterized in that, The process of obtaining a smooth trajectory during the lane-changing process through secondary dynamic programming based on the initial lane-changing trajectory includes the following steps: Based on the SL coordinate system, the road space of the initial lane-changing trajectory is discretized into an ordered lattice at intervals, and the lattice nodes are marked as level nodes. Iterate through each level node and perform collision detection between the current node and obstacles; If the collision detection result indicates that a collision has occurred, the level node is discarded. When the collision detection result is no collision, the loss value of each predecessor node of the current level node is calculated, wherein the loss value includes the lateral offset of the predecessor node from the target reference line, the lateral rate of change, and the lateral deviation of the endpoint position. The predecessor node with the smallest loss value is taken as the parent node of the current level node; Traverse until the last level node, find the parent node through the linked list, and continue until the initial node is found to obtain the smooth trajectory.

8. A vehicle autonomous lane-changing system, characterized in that, include: The first module is used to obtain road traffic information; The second module is used to construct the lane-changing obstacle topology relationship of the vehicle based on the road traffic information, wherein the lane-changing obstacle topology relationship includes current lane information and adjacent lane information; The third module is used to determine the lane-changing collision risk, the vehicle's benefits before and after lane changing, and the traffic flow disturbance around the vehicle based on the topological relationship of the lane-changing obstacles. The fourth module is used to make lane-changing decisions based on the lane-changing collision risk, the vehicle's benefits before and after the lane-changing, and the traffic flow disturbance around the vehicle, wherein the lane-changing decision includes the target lane. The fifth module is used to obtain the initial lane-changing trajectory by performing lateral and longitudinal joint planning constraints based on the lane-changing decision, the topological relationship of the lane-changing obstacles, and the vehicle parameters. The sixth module is used to perform secondary dynamic programming based on the initial lane-changing trajectory to obtain a smooth trajectory during the lane-changing process.

9. A vehicle autonomous lane-changing device, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the autonomous lane-changing method for vehicles as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a processor-executable program, characterized in that, When the processor executes the program, it is used to implement the autonomous lane-changing method for vehicles as described in any one of claims 1 to 7.

Images