Vehicle lane changing method, device and equipment

Abstract

The application provides a vehicle lane changing method, device and equipment, and relates to the technical field of automatic driving. The method comprises the following steps: in response to a received lane changing decision signal, obtaining vehicle operation data of a target vehicle, and obtaining a plurality of initial boundary lines of a target lane to be changed; in the plurality of initial boundary lines, a target boundary line meeting a quality requirement is determined, and a lane center reference line of the target lane is obtained by laterally offsetting the target boundary line; a lane changing path trajectory is determined according to the vehicle operation data and the lane center reference line; and the target vehicle is controlled to travel according to the lane changing path trajectory. The application can improve the adaptability, stability and safety of lane changing planning in a complex traffic environment.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and specifically to a vehicle lane-changing method, device, and equipment. Background Technology

[0002] With the rapid development of autonomous driving technology, automatic lane changing has become one of the core functions of advanced driver assistance systems (ADAS) and autonomous driving systems.

[0003] In existing technologies, when autonomous vehicles change lanes on multi-lane roads, they typically generate and track the lane change path based on lane line information and a preset trajectory model.

[0004] However, existing technologies mostly use fixed parameter trajectories or single reference lines for tracking, which makes it difficult to adjust in real time according to changes in the status of surrounding vehicles and road conditions, thus leading to low lane changing safety. Summary of the Invention

[0005] The purpose of this invention is to provide a vehicle lane-changing method, device, and equipment to solve the problem of low lane-changing safety in the prior art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] In a first aspect, the present invention provides a vehicle lane-changing method, comprising: in response to a received lane-changing decision signal, acquiring vehicle operation data of a target vehicle and acquiring multiple initial boundary lines of the target lane to be changed; determining a target boundary line that meets quality requirements among the multiple initial boundary lines, and laterally offsetting the target boundary line to obtain a lane center reference line of the target lane; determining a lane-changing path trajectory based on the vehicle operation data and the lane center reference line; and controlling the target vehicle to travel according to the lane-changing path trajectory.

[0008] Furthermore, among multiple initial boundary lines, a target boundary line that meets the quality requirements is determined, including: for each initial boundary line, determining at least one of the length quality, stability quality, and continuity quality of the initial boundary line; determining the comprehensive quality of the initial boundary line based on at least one of the length quality, stability quality, and continuity quality; and determining the target boundary line that meets the quality requirements among multiple initial boundary lines based on the comprehensive quality.

[0009] Furthermore, based on the overall quality, a target boundary line that meets the quality requirements is determined from multiple initial boundary lines, including: selecting the initial boundary line with the highest overall quality as a candidate boundary line from multiple initial boundary lines; determining the first lateral position of the candidate boundary line at a preset longitudinal distance, and the second lateral position of the target boundary line in the previous frame at a preset longitudinal distance ahead; and determining the current target boundary line from the candidate boundary line and the target boundary line in the previous frame based on the lateral deviation between the first and second lateral positions.

[0010] Further, based on the lateral deviation between the first lateral position and the second lateral position, the current target boundary line is determined among the candidate boundary line and the target boundary line of the previous frame, including: if the lateral deviation is greater than a lateral deviation threshold, the target boundary line of the previous frame is determined as the current target boundary line; if the lateral deviation is less than or equal to the lateral deviation threshold, the longitudinal deviation between the endpoint of the candidate boundary line and the endpoint of the target boundary line of the previous frame is determined; if the longitudinal length of the candidate boundary line is greater than the longitudinal length of the target boundary line of the previous frame, the longitudinal deviation is greater than or equal to the longitudinal deviation threshold, and the longitudinal length of the target boundary line of the previous frame is less than the longitudinal length threshold, the candidate boundary line is determined as the current target boundary line; otherwise, the target boundary line of the previous frame is determined as the current target boundary line.

[0011] Furthermore, based on vehicle operation data and lane center reference lines, the lane change path trajectory is determined, including: obtaining the target lane change style and determining the target lane change duration based on the target lane change style; generating an offset point sequence for switching to the target lane based on the target lane change duration, vehicle operation data, and lane center reference lines; and performing coordinate transformation processing on the offset point sequence based on the lane center reference lines to obtain the lane change path trajectory.

[0012] Furthermore, it also includes: acquiring the following vehicle's running data in the target lane; determining the yield probability of the following vehicle based on the following vehicle's running data and the vehicle's running data; triggering lane change abort if the yield probability is less than a probability threshold; and determining the target lane change style based on the yield probability if the yield probability is greater than or equal to the probability threshold.

[0013] Furthermore, it also includes: during the lane change process, detecting the first risk level of the target lane and the second risk level of the original lane; if the first risk level is higher than the first risk level threshold and the second risk level is lower than the second risk level threshold, triggering lane change abort and controlling the target vehicle to return to the original lane.

[0014] Furthermore, it also includes: determining the feasibility of reversing based on the distance between the rear wheels and the lane dividing line; if reversing is not feasible, generating a lane-keeping trajectory within the target lane and controlling the target vehicle to drive along the lane-keeping trajectory; if reversing is feasible, generating a reversal trajectory based on the original lane reference line; and controlling the target vehicle to return to the original lane, including: controlling the target vehicle to return to the original lane according to the reversal trajectory.

[0015] Furthermore, based on vehicle operation data and lane center reference lines, the lane change path trajectory is determined, including: obtaining the remaining longitudinal distance for lane change in navigation from the vehicle operation data; when the remaining longitudinal distance for lane change in navigation is less than the standard lane change distance, determining the trajectory generation constraint parameters under the forced lane change condition, including the available longitudinal distance for lane change, the maximum allowable vehicle speed, and the desired acceleration; and generating the lane change path trajectory based on the trajectory generation constraint parameters and the lane center reference lines.

[0016] Furthermore, it also includes: if a preset triggering condition is detected during the lane change process, the lane change path trajectory is adjusted; wherein the preset triggering condition includes at least one of the following: there is a visual merging point ahead, the vehicle in front of the target lane decelerates, and the lateral safety distance between the target vehicle and surrounding vehicles is insufficient.

[0017] In a second aspect, embodiments of the present invention provide a vehicle lane-changing device, comprising:

[0018] The acquisition module is used to acquire vehicle operation data of the target vehicle and multiple initial boundary lines of the target lane to be changed in response to the received lane change decision signal.

[0019] The first determining module is used to determine the target boundary line that meets the quality requirements among multiple initial boundary lines, and to laterally offset the target boundary line to obtain the lane center reference line of the target lane.

[0020] The second determining module is used to determine the lane change path trajectory based on vehicle operation data and lane center reference line;

[0021] The control module is used to control the target vehicle to travel according to the lane change path trajectory.

[0022] Thirdly, embodiments of the present invention provide an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0023] The memory stores the instructions that the computer executes;

[0024] The processor executes computer execution instructions stored in memory to implement the method provided in the first aspect above.

[0025] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method provided in the first aspect above.

[0026] Fifthly, embodiments of the present invention provide a computer program product, comprising: a computer program, which, when executed by a processor, implements the method provided in the first aspect above.

[0027] The beneficial effects of this invention are:

[0028] This invention provides a vehicle lane-changing method, apparatus, and device. The generation of the vehicle lane-changing path in this invention is no longer limited to a single reference line and a fixed trajectory, but instead involves quality screening of multiple boundary lines of the target lane, and determining a more stable lane center reference line based on this. Vehicle operation data and the determined lane center reference line jointly participate in path calculation, enabling the lane-changing trajectory to conform to the current vehicle state and the target lane, thereby improving the adaptability, stability, and safety of lane-changing planning in complex traffic environments. Attached Figure Description

[0029] Figure 1 A schematic diagram illustrating an application scenario provided in an embodiment of the present invention;

[0030] Figure 2 This is a flowchart illustrating the vehicle lane-changing method provided in an embodiment of the present invention.

[0031] Figure 3 A schematic diagram of the lane change path trajectory provided in an embodiment of the present invention;

[0032] Figure 4 This is a schematic diagram of lane-changing traffic conditions provided in an embodiment of the present invention;

[0033] Figure 5 A schematic diagram of a rollback scenario provided in an embodiment of the present invention;

[0034] Figure 6 This is a schematic diagram of a scenario where vehicles are driven close to the side of the road, as provided in an embodiment of the present invention.

[0035] Figure 7 A schematic diagram of a rapid lane change provided in an embodiment of the present invention;

[0036] Figure 8 A schematic diagram of continuous lane changing provided for an embodiment of the present invention;

[0037] Figure 9 A schematic diagram of path replanning provided in an embodiment of the present invention;

[0038] Figure 10 This is a schematic diagram of the vehicle lane-changing device provided in an embodiment of the present invention;

[0039] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0040] The accompanying drawings have illustrated specific embodiments of the invention, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the invention in any way, but rather to illustrate the concept of the invention to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0041] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0042] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0043] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0044] The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, product, or apparatus. Without further limitation, the presence of additional identical or equivalent elements in the process, method, product, or apparatus that includes elements is not excluded. For example, the use of terms such as "first," "second," etc., to indicate names does not imply any particular order.

[0045] Vehicle lane change path planning technology is mainly applied to lateral movement scenarios such as highways and urban multi-lane roads. After receiving a lane change decision signal during driving, the vehicle typically combines the vehicle's operating status, lane line perception results, and control execution module to generate a lane change path and control the vehicle to complete the lateral transfer from the current lane to the target lane.

[0046] Existing autonomous driving lane-changing solutions mostly rely on perceived lane line information to establish reference lines, then combine this with a preset trajectory model to generate a lane-changing path, and finally the lateral control module executes the lane change according to the generated trajectory. The basic idea of ​​this type of solution is to determine the target lane position at the start of the lane change, select a single lane line or center line as a reference, and then construct the lane-changing trajectory according to a fixed time, fixed curvature, or fixed offset pattern.

[0047] While the aforementioned methods can accomplish basic lane-changing tasks in relatively stable traffic environments, their shortcomings are easily exposed under complex and dynamic road conditions. Firstly, fixed-parameter trajectories lack the ability to respond to real-time changes in road conditions. When there are significant changes in the speed of vehicles ahead and behind, or dynamic interference in adjacent lanes, the generated path struggles to consistently match the actual available space, easily reducing the safety margin for lane changes. Secondly, the reliance on a single reference line is high. When lane line perception experiences jitter, obstruction, or switching, the reference line may abruptly change, causing lane-changing trajectory jitter, discontinuous deviations, and in severe cases, even trajectory interruptions, affecting vehicle control stability and ride comfort.

[0048] To address the aforementioned technical issues, the vehicle lane-changing method provided by this invention no longer limits the generation of lane-changing paths to a single reference line and a fixed trajectory. Instead, it focuses on quality screening of multiple boundary lines of the target lane and, based on this, determines a more stable lane center reference line. Vehicle operation data and the determined lane center reference line jointly participate in path calculation, enabling the lane-changing trajectory to conform to the current vehicle state and the target lane, thereby improving the adaptability, stability, and safety of lane-changing planning in complex traffic environments.

[0049] The following is a combination of... Figure 1 The application scenarios of the embodiments of the present invention are described.

[0050] Figure 1 This is a schematic diagram illustrating an application scenario provided by an embodiment of the present invention. For example... Figure 1 As shown, the application scenarios applicable to embodiments of the present invention include multiple lanes, such as lane A, lane B, and lane C. The target vehicle is currently in lane A. If lane-changing conditions are met, the target vehicle needs to switch from lane A to lane B; that is, lane B is the target lane to which it needs to switch. The present invention plans the lane-changing path trajectory for the target vehicle to switch from its current lane to the target lane, and then controls the target vehicle to travel according to this lane-changing path trajectory.

[0051] exist Figure 1 In this invention, the target vehicle can be a passenger car or a commercial vehicle. This embodiment of the invention does not limit the vehicle type or the corresponding applicable driving scenario; the specific requirements can be determined based on actual application needs.

[0052] The following is based on the above. Figure 1 The vehicle-mounted terminal deployed in the target vehicle shown is the execution subject. The specific implementation of the vehicle lane-changing method provided in this embodiment of the invention will be described in detail with reference to specific embodiments.

[0053] Figure 2 This is a flowchart illustrating a vehicle lane-changing method provided in an embodiment of the present invention. Figure 2 As shown, a specific implementation of this vehicle lane-changing method may include the following steps:

[0054] S201, in response to the received lane change decision signal, acquires the vehicle operation data of the target vehicle and acquires multiple initial boundary lines of the target lane to be changed.

[0055] The lane change decision signal is used to trigger the target vehicle to enter the lane change path generation and execution process. The lane change decision signal carries the lane change direction and the lane change triggering reason. The lane change triggering reason can include at least one of the following: navigation lane change, overtaking lane change, obstacle avoidance lane change, and parking lane change. Different lane change triggering reasons correspond to different lane change strategy parameters.

[0056] Furthermore, upon receiving the lane change decision signal, the lane change state machine is initialized, entering the lane change preparation state; then, vehicle operation data is acquired. This vehicle operation data characterizes the current motion state of the target vehicle and includes: the target vehicle's state information and its surrounding environment information. The current motion state includes data such as vehicle speed, heading angle, and position, which characterize the vehicle's current state.

[0057] In this invention, the target lane is used to indicate the lane object that the target vehicle is preparing to enter, and multiple initial boundary lines are used to provide the original boundary description information of the target lane.

[0058] Among these, multiple initial boundary lines can be lane lines of the target lane perceived from the BEV (Bird's-Eye View) based on the lane change direction. For example, multiple initial boundary lines include: the left boundary line and the right boundary line of the target lane.

[0059] In practice, the executing entity can be a controller deployed on the target vehicle. Upon receiving the lane change decision signal, the controller acquires the vehicle's operating data and identifies the target lane to be changed into based on the lane change direction, obtaining multiple initial boundary lines corresponding to the target lane.

[0060] For example, when an upper-level decision module (such as a navigation decision or overtaking decision module) issues a lane change request, S201 is executed. This involves receiving a lane change decision signal, including the lane change direction (left or right) and the lane change triggering reason. The lane change triggering reason includes at least one of the following: navigation lane change (reason=2): a lane change is required according to the navigation route, such as a lane change before a ramp or intersection; overtaking lane change (reason=1): a slower vehicle is ahead of the target vehicle in its lane, requiring a lane change to overtake; obstacle avoidance lane change (reason=4): an obstacle is ahead of the target vehicle in its lane, requiring a lane change to avoid the obstacle; parking lane change (reason=7): parking on the side of the road is required; user-instructed lane change (reason=25): the user triggers the lane change via turn signal or HMI (Human-Machine Interface) command.

[0061] Then, the following initialization operations are performed: the lane change state machine is set to "lane change preparation state"; the lane change counter is initialized: the counter is reset to zero to determine the stability of the lane change direction; lane change parameters are initialized: the total lane change time is set to the default value (e.g., 5 seconds), and the lane change distance is reset to zero; reference line identifiers are initialized: the target lane reference line identifier and the target vehicle lane reference line identifier are reset to zero, indicating that a valid reference line has not yet been tracked; hazard signs are initialized: the lane change abort reason is reset to zero, and the hazard target identifier is reset to zero. The state initialization sub-step S12 simultaneously acquires the following vehicle state information, which includes: target vehicle state information: vehicle speed, heading angle, steering wheel angle, yaw rate, etc.; target lane information: target lane reference line, target lane obstacle list; surrounding environment information: BEV lane lines, fused obstacle list, predicted trajectory, etc.

[0062] In this step, by acquiring the vehicle's own state and the target lane boundary information after the lane change decision signal arrives, subsequent path planning can be based on the vehicle's motion state and the road geometry, reducing trajectory deviations caused by inconsistent information.

[0063] S202, among multiple initial boundary lines, determine the target boundary line that meets the quality requirements, and laterally offset the target boundary line to obtain the lane center reference line of the target lane.

[0064] Among them, the target boundary line is a boundary line selected from multiple initial boundary lines that can characterize the boundary shape of the target lane, and the lane center reference line is a reference benchmark calculated based on the target boundary line for lane change path planning.

[0065] In this invention, the quality requirements can be preset. For example, the boundary line with better quality between the left and right boundary lines can be selected as the target boundary line.

[0066] In practice, multiple initial boundary lines can be quality assessed, and a target boundary line that meets the quality requirements can be determined based on the assessment results. If multiple boundary lines meet the quality requirements, one of them is selected as the target boundary line; if only one boundary line meets the quality requirements, that boundary line is directly selected as the target boundary line.

[0067] After determining the target boundary line, a lateral offset operation is performed on the target boundary line to obtain the lane center reference line. To ensure that the reference line can be used for subsequent trajectory generation, the obtained lane center reference line can also be processed for continuity. Specifically, the lateral offset of the target boundary line is half the lane width of the target lane, where the lane width can be obtained from the perception data of the target vehicle.

[0068] For example, during lane changing, the reference line of the target lane is continuously tracked to provide a stable target reference for generating the lane changing trajectory. First, BEV lane line information is received from the perception module. BEV lane lines include: the left and right lines of the current lane (original lane); the left and right lines of the left adjacent lane; the left and right lines of the right adjacent lane; each lane line includes: lane line identifier, lane line type (solid line / dashed line), and lane line point sequence.

[0069] When changing lanes to the left, and the target lane is the adjacent lane on the left, perform the following operations: From the BEV lane lines, find the lane line corresponding to the left edge of the adjacent lane and obtain its point sequence. If the left edge of the adjacent lane exists, calculate its distance from the target vehicle (lateral distance to the center of the target vehicle) and record its length (number of points in the point sequence). Then extract the right edge of the adjacent lane: From the BEV lane lines, find the lane line corresponding to the right edge of the adjacent lane (i.e., the left edge of the current lane) and obtain its point sequence. If it exists, similarly calculate its distance and length from the target vehicle.

[0070] Furthermore, the reference line used for trajectory generation is selected according to the following principles:

[0071] Principle 1: Length Priority Principle

[0072] Let the minimum length threshold be: ;

[0073] Let the expected length threshold be: ;

[0074] When the length of the left line is greater than the length of the right line plus 20, and the length of the right line is less than... When choosing a position, prioritize the left side.

[0075] When the length of the right line is greater than the length of the left line plus 20, and the length of the left line is less than... When choosing a position, prioritize the right side.

[0076] Principle Two: Lane Changing and Switching Principles

[0077] In the initial stage of lane change (before crossing the lane line), the right line of the target lane (i.e., the left line of the current lane) is used as the reference line; in the later stage of lane change (after crossing the lane line), it is necessary to switch to the left line of the target lane as the reference line; the switching condition is: when the longitudinal distance of the end of the right line is more than 30 meters less than the longitudinal distance of the end of the left line, and the longitudinal distance of the end of the right line is less than 40 meters, the switching is triggered.

[0078] Principle Three: The Principle of Historical Continuity

[0079] If the previous frame used the left line and the left line is still valid, continue using the left line; if the previous frame used the right line and the right line is still valid, continue using the right line.

[0080] Furthermore, the switching of the target reference line is verified to prevent trajectory jitter caused by jumps in the target reference line. The jump verification method is as follows: Let the reference line of the previous frame be prev_refer_vcs, and the candidate reference line of the current frame be update_refer_vcs; at a longitudinal distance of 40 meters (approximately 40 meters ahead), calculate the lateral deviation to_line_dis between the two reference lines; if If the reference line changes, the new reference line is rejected, and the reference line from the previous frame is used instead; if... Meters allow for reference line switching, using a new reference line.

[0081] When the reference line is switched, the following information is updated simultaneously: Target boundary line identifier: Records the lane line identifier of the currently used target boundary line;

[0082] In this step, by first screening target boundary lines that meet the quality requirements, and then calculating the lane center reference line from the target boundary line, the reference benchmark can be kept stable and continuous under the condition of lane line shaking, obstruction or short-term switching, thereby reducing the shaking and interruption caused by abrupt changes in the reference line in the lane change path.

[0083] S203 determines the lane change path trajectory based on vehicle operation data and lane center reference lines.

[0084] Among them, the lane change path trajectory is used to characterize the expected driving path of the target vehicle as it moves from the current lane to the target lane. The trajectory generation takes into account both the current motion state of the target vehicle and the geometric orientation of the center of the target lane.

[0085] In some alternative implementations, the relative relationship between the vehicle's current position and the lane center reference line is first established. The vehicle's current position can be determined by the positioning results and vehicle motion estimation results, and mapped to the corresponding coordinate system. The initial state of the lane change is determined based on the vehicle's operating data, and the planned endpoint is determined on the lane center reference line. Subsequently, a sequence of trajectory points that meets the continuity requirements is generated based on the vehicle's operating data and the lane center reference line, and the lane change path trajectory is obtained.

[0086] Furthermore, in order to ensure that the lane change path trajectory maintains a consistent trend with the lane center reference line, the continuity and smoothness of the generated trajectory can be checked, and adjustments can be made to ensure that the generated trajectory meets the requirements of vehicle lane change execution.

[0087] For example, the target reference line coordinate transformation and post-processing are performed: the target reference line is transformed from the BEV coordinate system (geocentric coordinate system) to the vehicle coordinate system (VCS) to obtain the lane center reference line. Specifically, the center of the target vehicle's rear axle is taken as the origin, the direction of the vehicle's front is the positive X-axis, and the left side is the positive Y-axis; the area behind the target vehicle (…) is removed. The target reference points should be preserved, including those in front of and near the target vehicle; the length of the target reference points must be at least [amount missing]. If insufficient, supplement by extending from the navigation reference lines.

[0088] The coordinate transformation method for reference lines is as follows: ;

[0089] Where ego_pos is the position of the target vehicle in the geodetic coordinate system, and ego_heading is the heading angle of the target vehicle.

[0090] In this step, vehicle operation data and lane center reference lines are used together for trajectory generation, which enables the lane change path to simultaneously reflect the vehicle's current dynamic state and the geometry of the target lane. This makes the trajectory point sequence spatially continuous and executable in terms of attitude changes, thus adapting to lane change requirements under complex dynamic road conditions.

[0091] S204, control the target vehicle to travel according to the lane change path trajectory.

[0092] After the lane change path trajectory is generated, steering, driving and braking control commands can be output according to the lane change path trajectory, so that the target vehicle can complete the lane change along the planned lane change path trajectory.

[0093] In practice, the current trajectory point in the lane change path is read, and based on the current trajectory point, a front wheel turning angle command or a steering wheel turning angle command is generated to control the vehicle to turn and change lanes.

[0094] Furthermore, to achieve closed-loop control, the vehicle continuously reports its current position, real-time heading, and real-time speed during lane change execution. The control module compares the real-time status with the lane change path trajectory at a fixed control cycle and updates the control variables.

[0095] In addition, during the lane change process, the lane change path can be readjusted based on the actual operating status of the target vehicle and the surrounding environmental data.

[0096] In some alternative implementations, the present invention also maintains a lane change state machine to manage the transition logic between various states, output the lane change path trajectory, and provide auxiliary state algorithms.

[0097] For example, the lane change state machine includes: (1) No lane change state: no lane change triggered; (2) Lane change preparation state: lane change has been triggered, waiting for the conditions to be met; (3) Lane change execution state: executing the lane change trajectory; (4) Next lane change preparation state: the lane change is about to be completed, preparing to trigger the next lane change; (5) Lane change abort state: the lane change has been aborted, returning to the original lane; (6) Lane change completed state: the lane change is completed. The output information includes: the sequence of lane change path trajectory points; the lane center reference line; the lane change direction and lane change status; the reason for the lane change abort; and longitudinal decision suggestions (such as needing to decelerate or accelerate).

[0098] The reasons for lane change termination include at least one of the following: stationary obstacle in the lane, the following vehicle turning back into the lane, the following vehicle showing a tendency to turn in, the driver exiting the following vehicle, the lane line becoming a solid line, a large vehicle cutting into the adjacent lane, driver takeover alarm, and traffic divider constraints.

[0099] Based on the above analysis, the present invention provides a vehicle lane-changing method. The lane-changing trigger, boundary screening, reference line generation, trajectory planning, and closed-loop control are executed in a continuous manner according to the same task flow. The target lane does not depend on a single unscreened boundary line. Instead, a target boundary line that meets the quality requirements is first determined from multiple initial boundary lines. Then, the lane center reference line is calculated, and a trackable lane-changing path trajectory is generated by combining the vehicle's real-time operating data. This enables the vehicle to complete stable and continuous lane-changing control under complex dynamic road conditions and improves the safety of the lane-changing process.

[0100] Optionally, among multiple initial boundary lines, a target boundary line that meets the quality requirements is determined, including: for each initial boundary line, determining at least one of the length quality, stability quality, and continuity quality of the initial boundary line; determining the comprehensive quality of the initial boundary line based on at least one of the length quality, stability quality, and continuity quality; and determining the target boundary line that meets the quality requirements among multiple initial boundary lines based on the comprehensive quality.

[0101] The length mass, stability mass, and continuity mass can be weighted and summed to obtain the overall mass of the initial boundary line. For example, the overall mass of the i-th initial boundary line satisfies the following expression (1):

[0102] (1)

[0103] In the above expression (1), The overall quality of the i-th initial boundary line is... Let the length and mass of the i-th initial boundary line be denoted as . Let the stability mass of the i-th initial boundary line be , Let be the continuity quality of the i-th initial boundary line. , and The preset weights.

[0104] The length mass satisfies the following expression (2):

[0105] (2)

[0106] In the above expression (2), Let be the longitudinal length of the i-th initial boundary line, which is the currently perceived longitudinal length. The first length threshold can be preset or determined based on the longitudinal speed of the target vehicle, for example, , The unit can be meters (m). The second length threshold can be preset or determined based on the longitudinal speed of the target vehicle, for example... .

[0107] In this invention, when the longitudinal length of the initial boundary line is less than a first length threshold, the length mass is 0. Between the first and second length thresholds, the length mass increases linearly. When it reaches or exceeds the second length threshold, the length mass is 1. Furthermore, both the first and second length thresholds dynamically increase with the vehicle's speed.

[0108] The stability mass satisfies the following expression (3):

[0109] (3)

[0110] In the above expression (3), N is the total number of sampling points on the initial boundary line. For the j-th sampling point on the current initial boundary line, compare the lateral deviation with the corresponding sampling point on the target boundary line determined in the previous frame. This is a preset distance coefficient, for example, 0.2m.

[0111] In this invention, lateral deviation The smaller, The closer to 1; the greater the lateral deviation. The larger, It approaches 0. Therefore, when the initial boundary line is the same as the target boundary line of the previous frame, A larger value indicates higher stability and quality. When the initial boundary line differs from the target boundary line of the previous frame, A smaller value indicates lower stability and quality.

[0112] Among them, the continuous quality satisfies the following expression (4):

[0113] (4)

[0114] In the above expression (4), The target reference line from the previous frame. For the i-th initial boundary line Target reference line from the previous frame Lateral deviation, This is a preset lateral deviation threshold, such as 2m.

[0115] In this invention, if the initial boundary line is the target reference line of the previous frame, the corresponding continuity quality is relatively high, which is 1; if the initial boundary line is not the target reference line of the previous frame, and the lateral distance between the two is less than the deviation threshold, the corresponding continuity quality is 0.8. If the initial boundary line is not the target reference line of the previous frame, and the lateral distance between the two is greater than or equal to the deviation threshold, the corresponding continuity quality is 0.

[0116] In the specific implementation, if there is no target reference line from the previous frame, meaning the target reference line to be determined is the first frame, then the overall quality can be determined solely based on length quality, with a weight of 1 for length quality. If there is a target reference line from the previous frame, then the overall quality can be determined based on length quality, stability quality, and continuity quality.

[0117] After completing the comprehensive quality calculation, multiple initial boundary lines can be compared and selected based on the comprehensive quality to obtain the target reference line.

[0118] In one optional implementation, the initial boundary line with the highest overall quality can be selected as the target boundary line, or one of the initial boundary lines with an overall quality greater than a quality threshold can be selected as the target boundary line. In this way, the target boundary line is an initial boundary line with sufficient length, minimal fluctuation, and good continuity, thus providing a stable foundation for the subsequent lateral offset to generate the lane center reference line.

[0119] In summary, boundary line selection no longer relies on a single perception result, but determines the target boundary line through a comprehensive evaluation of length quality, stability quality, and continuity quality. This can improve the usability and consistency of the target boundary line and provide a more stable source of reference lines for the generation of lane change path trajectories.

[0120] Optionally, based on the overall quality, a target boundary line that meets the quality requirements is determined from multiple initial boundary lines, including: selecting the initial boundary line with the highest overall quality as a candidate boundary line from multiple initial boundary lines; determining the first lateral position of the candidate boundary line at a preset longitudinal distance, and the second lateral position of the target boundary line in the previous frame at a preset longitudinal distance ahead; and determining the current target boundary line from the candidate boundary line and the target boundary line in the previous frame based on the lateral deviation between the first and second lateral positions.

[0121] If the target boundary line of the first frame is not currently being determined, a jump check is performed on the candidate boundary line after it is determined. If the check passes, the target boundary line can be determined based on the candidate boundary line; if the check fails, the target boundary line of the previous frame is determined as the current target boundary line. If the target boundary line of the first frame is currently being determined, the candidate boundary line is determined as the current target boundary line.

[0122] For example, the lateral deviation satisfies the following expression (5):

[0123] (5)

[0124] In the above expression (5), Let be the lateral coordinate of the i-th initial reference line at a preset vertical distance position (e.g., 40 meters ahead). The horizontal coordinates of the target boundary line in the previous frame at a preset vertical distance position (e.g., 40 meters ahead).

[0125] In this invention, if If the deviation exceeds the lateral deviation threshold (e.g., 2m), the jump check of the candidate boundary line is determined to have failed, meaning the target boundary line of the previous frame will be removed. As the current target boundary line. If If the deviation is less than or equal to the lateral deviation threshold, the candidate boundary line will be used as the target boundary line.

[0126] In practice, after selecting the candidate boundary line with the highest overall quality, its lateral coordinates are determined at a preset distance ahead. Simultaneously, the lateral coordinates of the target boundary line in the previous frame at the same vertical position are read, and the difference between the two is calculated as the lateral deviation. When the lateral deviation is small, it indicates that the switching jump distance is low, and it is directly switched to the candidate boundary line. When the lateral deviation is large, the target boundary line in the previous frame is kept as the current target boundary line to avoid a large jump distance of the target boundary line.

[0127] In summary, this invention uses the boundary line with the highest overall quality as the candidate and combines it with the lateral deviation at a preset distance ahead for historical consistency judgment, ensuring that the current target boundary line balances real-time perception quality and cross-frame continuity. The resulting target boundary line can serve as the basic input for subsequent lane center reference lines, keeping the reference line used for lane change path calculation smooth and stable, and reducing the impact of abrupt boundary line changes on the lane change path trajectory, thereby improving the continuity of lane change planning and the consistency of control execution.

[0128] Further, based on the lateral deviation between the first lateral position and the second lateral position, the current target boundary line is determined among the candidate boundary line and the target boundary line of the previous frame, including: if the lateral deviation is greater than a lateral deviation threshold, determining the target boundary line of the previous frame as the current target boundary line; if the lateral deviation is less than or equal to the lateral deviation threshold, determining the longitudinal deviation between the endpoint of the candidate boundary line and the endpoint of the target boundary line of the previous frame; determining the longitudinal deviation between the endpoint of the candidate boundary line and the endpoint of the target boundary line of the previous frame; and determining the current target boundary line among the candidate boundary line and the target boundary line of the previous frame based on the longitudinal deviation and the longitudinal length of the candidate boundary line.

[0129] Specifically, based on the longitudinal deviation and the longitudinal length of the candidate boundary line, the current target boundary line is determined from the candidate boundary line and the target boundary line of the previous frame, including: determining the candidate boundary line as the current target boundary line when the longitudinal length of the candidate boundary line is greater than the longitudinal length of the target boundary line of the previous frame, the longitudinal deviation is greater than or equal to the longitudinal deviation threshold, and the longitudinal length of the target boundary line of the previous frame is less than the longitudinal length threshold; otherwise, determining the target boundary line of the previous frame as the current target boundary line.

[0130] For example, if the candidate boundary line is the left boundary line and the target boundary line in the previous frame is the right boundary line, then if the following conditions are met... In this case, switch to the right boundary line, meaning the current target boundary line is the right boundary line. If the candidate boundary line is the right boundary line and the target boundary line in the previous frame was the left boundary line, then if the following conditions are met... In this case, switch to the left boundary line, meaning the current target boundary line is the left boundary line. Among these, Indicates the longitudinal length of the left boundary line. Indicates the longitudinal length of the right boundary line. This indicates the preset longitudinal deviation threshold. The longitudinal length threshold is a preset value. In this invention, the longitudinal deviation threshold and the longitudinal length threshold can also be other values, which are not limited here.

[0131] In the specific implementation, candidate boundary lines are first formed from the boundary lines with the highest overall quality. Then, the target boundary line saved in the previous frame is extracted, and the two are sampled and aligned at a preset distance in front, and the lateral deviation of the corresponding sampling points is calculated. When the lateral deviation exceeds the lateral deviation threshold, it is determined that the difference between the candidate boundary line and the historical target boundary line is too large, and the target boundary line of the previous frame is directly used as the current target boundary line to maintain the continuity between frames.

[0132] If the lateral deviation does not exceed the lateral deviation threshold, continue to calculate the longitudinal deviation of the endpoints of the two boundary lines, and then filter the results by combining the longitudinal lengths of the two boundary lines.

[0133] In this invention, the target boundary line determination method is used in lane change planning to perform inter-frame stability screening of the target lane boundary, so that the current target boundary line preferentially inherits the spatial features of the previous frame, and only switches to the candidate boundary line when the lateral and longitudinal consistency conditions are met, thereby keeping the subsequent lane change path generated by the lane center reference line continuous.

[0134] In summary, this invention enables the current target boundary line to maintain high stability even when sensing jitter or changes in boundary line length, reduces fluctuations in the lane center reference line caused by frequent boundary line switching, and thus makes the lane change path trajectory more continuous, improving the stability and consistency of vehicle lane change control.

[0135] Optionally, the lane change path trajectory is determined based on vehicle operation data and lane center reference line, including: obtaining the target lane change style and determining the target lane change duration based on the target lane change style; generating an offset point sequence for switching to the target lane based on the target lane change duration, vehicle operation data and lane center reference line; and performing coordinate transformation processing on the offset point sequence based on the lane center reference line to obtain the lane change path trajectory.

[0136] The target lane change style can be preset by the user or generated in real time, without any limitation.

[0137] In the specific implementation process, the target lane change duration can be determined based on the lane change duration threshold corresponding to the target lane change style and / or the longitudinal speed of the target vehicle.

[0138] For example, if the target lane change style is comfort style, then the target lane change duration satisfies ,in, This indicates the target lane change duration, where 6.6 is the lane change duration threshold for the Comfort style, in seconds. If the target lane change style is Normal, then the target lane change duration meets the following requirements. 5.1 represents the lane change duration threshold for the normal style. If the target lane change style is an aggressive style, then the target lane change duration satisfies... , where 3.5 is the lane change duration threshold corresponding to the aggressive style.

[0139] Furthermore, if an obstacle is detected ahead of the target lane requiring the target vehicle to decelerate, and the acceleration is less than a preset acceleration threshold, the target lane change duration can be adjusted. The adjusted target lane change duration satisfies the following expression (6):

[0140] (6)

[0141] In the above expression (6), The target lane change time is updated, and 0.7 is the reduction factor. This reduction factor can also be any other value less than 1 and greater than 0, which is not limited here. This represents the longitudinal distance between the target vehicle and the obstacle in front.

[0142] In the specific implementation process, the generation of the offset point sequence for switching to the target lane, based on the target lane change duration, vehicle operation data, and lane center reference line, includes the following steps:

[0143] (1) Determine the lateral offset The boundary conditions are satisfied;

[0144] in, The following boundary conditions must be met: ; ; ; Where s is the longitudinal distance traveled along the road, This represents the lateral offset that varies with the longitudinal travel distance s. The total longitudinal travel distance for a complete lane change. This refers to the lateral distance from the target vehicle's current position to the lane center reference line, such as 3.5m. It is the derivative of the lateral offset with respect to the longitudinal travel distance, and represents the slope of the lateral offset relative to the longitudinal offset.

[0145] Furthermore, The lateral offset corresponding to the lane change starting point s=0 is 0, indicating that the target vehicle is in the original lane and has not yet started to move laterally. Indicates the end of the lane change. The lateral offset is the lateral distance from the current position to the lane center reference line. The vehicle has completely changed lanes to the target lane and completed the lane change. This indicates that at the initial moment of lane change, the lateral movement tendency of the target vehicle is 0, and the vehicle body is moving straight along the original lane. This indicates that at the moment the lane change ends, the lateral movement tendency is 0, and the vehicle is moving straight along the target lane.

[0146] (2) Determine the total longitudinal driving distance :

[0147] Among them, based on the target lane change duration and the longitudinal speed of the target vehicle Determine the total longitudinal driving distance, that is, satisfy... .

[0148] Furthermore, the total longitudinal driving distance can be constrained and adjusted based on the target lane-changing style, specifically by determining the total driving distance threshold corresponding to the target lane-changing style. In that and The larger value is selected as the final total longitudinal driving distance.

[0149] For example, if the lane change style is comfort style, The value is 20m; if the lane change style is normal, It is 18m; if the lane-changing style is aggressive, It is 15m.

[0150] In some embodiments, if the longitudinal distance between the current position of the target vehicle and the merging point ahead is detected... When less than a threshold determined based on longitudinal velocity (e.g.) If so, the final total longitudinal driving distance will be reduced, for example: ,in, This can be used as the final total longitudinal travel distance to update the lateral offset. The boundary conditions are satisfied.

[0151] (3) Using the above boundary conditions as constraints, the offset point sequence for switching to the target lane is obtained by fitting.

[0152] The offset point sequence comprises multiple ordered offset points, each including a longitudinal offset and a lateral offset relative to the starting point of the target vehicle. The starting point of the offset point sequence can be represented as ( , ) = (0 The endpoint can be represented as ( , =( ). This represents the vertical offset of the j-th offset point. This represents the lateral offset of the j-th offset point. The offset point sequence has a total of M+1 offset points, with j=0 for the starting point and j=M for the ending point.

[0153] In this invention, polynomials (such as fifth-degree polynomials) can be used. The offset point sequence can be obtained by fitting, but other methods can also be used to obtain the offset point sequence, and there is no limitation on this.

[0154] In some optional implementations, the offset point sequence may further include: multiple extended offset points corresponding to the longitudinal extension distance after the endpoint, wherein the longitudinal offset of these extended offset points is greater than... The lateral offset is The longitudinal extension distance is determined based on the target vehicle's longitudinal speed and lane-changing style, specifically meeting the following requirements. , Indicates the longitudinal extension distance. The extended time corresponding to the lane change style, for example, if the lane change style is comfort style, If the lane-changing style is the normal style, If the lane-changing style is aggressive, .

[0155] In this invention, the longitudinal extension distance is a straight-line trajectory that continues forward from the end of the lane change. That is, after the target vehicle has changed lanes to the target lane, an extra straight-line path is reserved for smooth trajectory connection and to avoid abrupt changes in the trajectory at the end of the lane change.

[0156] Furthermore, after generating the offset point sequence, the offset point sequence is subjected to coordinate transformation based on the lane center reference line to obtain the lane change path trajectory.

[0157] The process of transforming the offset point sequence based on the lane center reference line to obtain the lane change path trajectory includes: calculating the longitudinal distance the target vehicle has moved forward based on the longitudinal speed and the elapsed time after trajectory generation; compensating and updating the longitudinal coordinates of each offset point in the offset point sequence based on the longitudinal distance to offset the coordinate system drift caused by the vehicle's forward movement; performing linear interpolation on the offset point sequence for each sampling point on the lane center reference line to obtain the lateral offset; calculating the azimuth angle of the corresponding point on the lane center reference line and superimposing the interpolated lateral offset along the vertical direction of the reference line onto the lane center reference line point to obtain the original trajectory point; inserting the origin of the target vehicle as the trajectory starting point and performing smoothing filtering on all trajectory points; and confining trajectory points that exceed the lane boundary back within the lane boundary range to finally obtain the lane change path trajectory.

[0158] In the specific implementation process, the lane change path trajectory generation process includes the following steps:

[0159] (1) Determine the longitudinal movement distance:

[0160] The longitudinal movement distance satisfies the following expression (7):

[0161] (7)

[0162] In the above expression (7), The longitudinal distance the target vehicle travels forward within the elapsed time; This represents the time elapsed since the moment this trajectory was generated.

[0163] (2) Update the vertical offset:

[0164] Among them, based on the longitudinal movement distance Update the vertical offset of each offset point in the offset point sequence. The vertical offset after offset satisfies the following expression (8).

[0165] (8)

[0166] In expression (8), the left side of the equation The adjusted vertical offset is shown on the right side of the equation. This represents the original vertical offset in the offset point sequence.

[0167] (3) Generate the trajectory points of the lane change path:

[0168] First, based on the points on the lane center reference line and the points in the offset point sequence, the linear interpolation weight coefficients are determined, where the linear interpolation weight coefficients k satisfy the following expression (9):

[0169] (9)

[0170] In the above expression (9), This represents the longitudinal offset of the p-th point in the lane center reference line. Let j be the vertical offset of the j-th offset point in the offset point sequence. Let be the longitudinal offset of the (j-1)th offset point in the offset point sequence. The j-th and (j-1)th offset points in the offset point sequence are adjacent points of the p-th point in the lane center reference line, satisfying the following condition: .

[0171] Then, linear interpolation is performed on each point on the lane center reference line to obtain the interpolated lateral offset for each point. Specifically, the interpolated lateral offset for the p-th point on the lane center reference line is... Satisfy the following expression (10):

[0172] = (10)

[0173] Furthermore, determine the azimuth angle at point p on the lane center reference line, which satisfies the following expression (11):

[0174] (11)

[0175] The lateral offset is obtained by linear interpolation at each point on the lane center reference line:

[0176] Finally, by superimposing the offset onto the lane center reference line, we can obtain each trajectory point, which satisfies the following expression: ; ;in,: Let p be the longitudinal coordinate of the trajectory point corresponding to the lane center reference line. The lateral coordinates of the trajectory point corresponding to point p on the lane center reference line.

[0177] In this invention, p can take values ​​from 0 to N, where N is the total number of sampling points on the lane center reference line. This results in N+1 trajectory points.

[0178] In some alternative implementations, the origin of the target vehicle can be used as the starting point, and smoothing filters can be applied to each trajectory point.

[0179] Furthermore, boundary limits can be preset, and trajectory points that exceed the boundary limits can be processed back to within the boundary.

[0180] In this embodiment of the invention, the lane change path trajectory is composed of the aforementioned multiple trajectory points.

[0181] For example, refer to Figure 3 This shows a schematic diagram of the lane change path trajectory. Figure 3 In the diagram, the target vehicle is currently traveling in lane A and needs to change lanes to lane B. The dashed lines in the diagram represent the reference line of lane A and the lane center reference line of lane B, respectively. Multiple offset points represent the relative offset relationship with the lane center reference line of lane B. After coordinate transformation, they are spliced ​​together to form a complete lane change path trajectory. The target vehicle can smoothly travel from lane A to lane B along this lane change path trajectory.

[0182] In summary, this invention generates lane-change path trajectories that simultaneously consider lane-change style, duration constraints, and vehicle operating status. Therefore, the trajectory shape remains consistent with specific lane-change requirements, and the distribution of trajectory points is coordinated with the actual movement characteristics of the vehicle. The resulting lane-change path trajectory can be directly called by subsequent control modules to drive the vehicle to complete a lateral transfer along the target lane direction, maintaining trajectory continuity and execution consistency.

[0183] Optionally, it also includes: acquiring the following vehicle operation data of the vehicles behind the target vehicle in the target lane; determining the yield probability of the following vehicles based on the following vehicle operation data and the vehicle operation data; triggering lane change abort if the yield probability is less than a probability threshold; and determining the target lane change style based on the yield probability if the yield probability is greater than or equal to the probability threshold.

[0184] In lane-changing scenarios, a game-theoretic relationship exists between the target vehicle and the vehicles behind it in the target lane: the target vehicle wants to complete the lane change, and the vehicles behind it may or may not yield. This invention establishes a prediction model for the probability of the following vehicle yielding.

[0185] The following vehicle operation data includes: the position of the vehicle behind, the speed of the vehicle behind, etc.

[0186] In this invention, the relative distance from the front of the following vehicle to the rear of the target vehicle is first determined. (m) and relative velocity (m / s). Then, the collision time TTC (s) is determined based on the relative distance and relative velocity.

[0187] In the specific implementation process, the collision time TTC satisfies the following expression (12):

[0188] (12)

[0189] In the above expression (12), This is the speed threshold, which can be a value greater than or equal to 0, such as 0.1; , Let be the longitudinal speed of the vehicle behind. A positive value indicates that the vehicle behind is traveling at a higher speed than the target vehicle. = , This refers to the length of the vehicle behind. A value greater than 0 indicates that the vehicle behind is behind the target vehicle.

[0190] In this invention, the yield probability of a vehicle behind can be obtained by weighting multidimensional features according to a preset yield probability model.

[0191] In the specific implementation process, the yield probability model is represented by the following expression (13):

[0192] (13)

[0193] In the above expression (13), For the probability of yielding; , , , , These are preset weighting coefficients, and the sum of all weighting coefficients is 1. For example, 0.25 0.3 0.2 0.15 It is 0.1.

[0194] Specifically, This represents the distance factor, which satisfies the following expression (14):

[0195] (14)

[0196] In the above expression (14), This refers to a situation where the vehicle behind is behind the target vehicle, and the front of the vehicle behind extends beyond the rear of the target vehicle. The first distance threshold, such as 10m, represents the minimum safe distance. This is the second distance threshold, such as 30m, which represents the comfortable distance.

[0197] In expression (13), Let the collision time factor be the collision time factor, which satisfies the following expression (15):

[0198] (15)

[0199] In the above expression (15), The first collision time threshold, such as 2 seconds; This is the second collision time threshold, such as 5 seconds.

[0200] In expression (13), Let the velocity factor be the velocity factor that satisfies the following expression (16):

[0201] (16)

[0202] In the above expression (16), This represents the speed threshold, indicating the safe speed difference threshold, for example, 5 m / s.

[0203] In expression (13), Let be the intent factor, which satisfies the following expression (17):

[0204] (17)

[0205] In the above expression (17), This represents the probability that vehicles behind will proceed straight. The probability of the vehicle behind intending to slow down. This represents the probability of a vehicle behind changing lanes to the other side of the target vehicle. Among these, , and It is obtained from the preset prediction module, which is used to determine... , and This invention does not limit the specific prediction method.

[0206] In expression (13), This is a lane change necessity factor, which can be determined based on the lane change necessity level. (e.g., any level from 1 to 4), determine the lane change necessity factor. For example, <2 hours, ; hour, ; hour, The level of necessity for lane changing can be carried in the lane change decision signal or determined in real time based on vehicle operation data; there is no limitation on this.

[0207] Furthermore, the probability threshold is preset, such as 0.3. For example, in the probability of yielding... When this occurs, the lane change is aborted. In some implementations, the collision time, i.e., the yield probability, may also be considered. If the collision time TTC is less than the preset collision time threshold, the lane change abort will be triggered.

[0208] In this invention, if the yield probability is greater than or equal to a probability threshold, the target lane-changing style is determined based on the yield probability. The higher the yield probability, the more aggressive the target lane-changing style. For example, if the yield probability is 0.3... When <0.6, the target lane change style can be either comfort style or normal style. The target lane-changing style is an aggressive style.

[0209] In summary, the closer the following vehicle's operating state is to meeting the yield condition for the target vehicle, the higher the probability of yielding. Based on this, a smoother or more comfortable lane-changing style is adopted. When the following vehicle exhibits strong following or imposing characteristics, the probability of yielding decreases, and the lane change is directly aborted to ensure that the lane-changing decision is consistent with the surrounding traffic interaction. Therefore, when executing a lane change, the target vehicle can update its decision based on the dynamic behavior of the following vehicles, making the lane-changing trajectory generation and control output more consistent with the actual traffic interaction and providing a more stable basis for safety judgment during the lane-changing process.

[0210] Optionally, it also includes: during the lane change process, detecting a first risk level of the target lane and a second risk level of the original lane; if the first risk level is higher than the first risk level threshold and the second risk level is lower than the second risk level threshold, triggering lane change abort and controlling the target vehicle to return to the original lane.

[0211] This invention performs real-time risk assessment during lane changes. The target lane can be divided into three risk zones: the area ahead, the parallel area, and the area behind. Corresponding risk assessments are then performed for each of these different risk zones.

[0212] For example, refer to Figure 4 ,exist Figure 4 In the scenario, the target vehicle is currently in lane B and intends to change lanes to lane A. Lane A contains high-risk obstacle vehicles a1, a2, and a3. Lane B contains high-risk obstacle vehicle b1 and a risk-free obstacle vehicle b2. Lane C contains risk-free obstacle vehicles c1, c2, and c3. During the lane change, a risk assessment is conducted in real time. The target lane is divided into a forward area, a parallel area, and a rear area. The risk of obstacle vehicles falling into each area is assessed to evaluate the collision risk of this lane change.

[0213] In the specific implementation process, a relative acceleration model is used for risk assessment of the rear area. This relative acceleration model is expressed by the following expression (18):

[0214] (18)

[0215] In the above expression (18), This is a relative equivalent acceleration, used to estimate how much acceleration or deceleration an obstacle would need to catch up with the target vehicle, assuming both vehicles maintain their current speeds.

[0216] Furthermore, when If the target vehicle has not crossed the lane line between the original lane and the target lane, the first sub-risk level of the rear area is determined to be high risk. The preset relative acceleration threshold, It is a negative value. It can correspond to the longitudinal speed of the target vehicle, changing as the longitudinal speed changes, for example, When =0.1m / s, ; When =0.15m / s, ; When =0.2m / s, ; When =0.25m / s, ; When =0.3m / s, .when If the vehicle has already crossed the lane line, the first sub-risk level of the area behind it is medium risk. In other cases, the first sub-risk level of the area behind it is low risk.

[0217] In this invention, a risk assessment of the parallel area is performed based on parallel obstacles (such as parallel vehicles). The assessment method is as follows: first, the time required for the parallel obstacle to exit the parallel area is determined, wherein the time satisfies the following expression (19):

[0218] (19)

[0219] In the above expression (19), The longitudinal velocity of the parallel obstacles, This represents the longitudinal distance between the parallel obstacles and the target vehicle. The length of the parallel obstacles, The time it takes for the parallel obstacle to exit the parallel region. This indicates that the parallel obstacle, moving slower than the target vehicle, exits from behind. This indicates that the parallel obstacle is faster than the target vehicle and exits from the front.

[0220] Furthermore, the effective lateral distance between the parallel obstacle and the target vehicle satisfies the following expression (20):

[0221] (20)

[0222] In the above expression (20), For effective lateral distance, This represents the absolute lateral distance between the parallel obstacle and the target vehicle. Let be the width of the parallel obstacles. This indicates that there is still space in the lateral direction after deducting the width of parallel obstacles and safety margin. This indicates insufficient horizontal space.

[0223] In this invention, and In the case of [specific condition], the second sub-risk level of the parallel region is determined to be low risk; otherwise, the second sub-risk level of the parallel region is determined to be high risk. Among these, and This is the preset threshold.

[0224] In this invention, a relative acceleration model is used for risk assessment of the forward region, and this relative acceleration model is represented by the following expression (20):

[0225] (20)

[0226] In the above expression (20), The longitudinal velocity of the obstacle ahead. This represents the longitudinal relative distance between the obstacle ahead and the target vehicle. This represents the theoretical acceleration required for the target vehicle to avoid a collision.

[0227] Furthermore, on Predicted acceleration relative to obstacles ahead The fusion process is performed to obtain a fusion acceleration. Specifically, when... At that time, fusion acceleration In other cases, .

[0228] In this invention, if the fusion acceleration At that time, the third sub-risk level corresponding to the area in front is high risk, among which It is a negative number; when When the risk level is medium, the third sub-risk level corresponding to the area in front is medium; otherwise, the third sub-risk level corresponding to the area in front is low.

[0229] In this embodiment of the invention, the first risk level of the target lane can be determined based on the first sub-risk level of the rear region, the second sub-risk level of the parallel region, and the third sub-risk level of the front region. For example, the higher of the first, second, and third sub-risk levels can be taken as the first risk level.

[0230] In some implementations, if there are no obstacles in the target lane, the first level of risk can be no risk or low risk.

[0231] Furthermore, a second level of risk is determined based on obstacles in the target vehicle's original lane. Specifically, the presence of dangerous obstacles in the original lane is first assessed. For example, if a vehicle is rapidly approaching from behind in the original lane, and its relative acceleration to the target vehicle is less than an acceleration threshold (e.g., -1.5 m / s²), then the area behind the original lane is considered to be highly dangerous. Similarly, if an obstacle is present in front of the target vehicle in the original lane, and the obstacle's speed is less than a speed threshold (e.g., 1 m / s), requiring the target vehicle to accelerate less than an acceleration threshold (e.g., -2.6 m / s²) to avoid a collision, then the area in front of the target vehicle is considered to be highly dangerous.

[0232] In this invention, if there is a dangerous obstacle in the original lane (i.e., a second risk level that is higher than the second risk threshold) and there is also a dangerous obstacle in the target lane (i.e., a first risk level that is higher than the first risk threshold), lane change abort is not triggered to avoid a dilemma after lane change abort. If there is a dangerous obstacle in the target lane but no dangerous obstacle in the original lane, lane change abort is triggered to return to the original lane.

[0233] In some implementations, if the risk of oncoming traffic from behind in the target lane is high, and the original lane is safe, the lane change is abandoned and the vehicle retreats to the target lane. If the target vehicle has already crossed the lane line, and there is a high-risk obstacle parallel to it in the target lane, the vehicle does not retreat or continue lateral movement, but remains in place, waiting for the obstacle to move away. If the obstacle in front of the target lane poses a high risk but has not yet crossed the lane line, the lane change is aborted. In other scenarios, the lane change can continue. It can be understood that if the target vehicle has not crossed the lane line, and there is a risk in the target lane, the lane change is aborted first; if the target vehicle has already crossed the lane line, the lane change is continued first to avoid lateral movement that could lead to a collision.

[0234] In summary, this invention, by simultaneously assessing the risk of the target lane and the risk of the original lane during lane changing, and aborting the lane change and reverting to the original lane when the target lane is considered high-risk and the original lane is considered low-risk, enables vehicles to promptly return to a relatively safe driving lane when the environment changes, thereby improving the stability and safety of lane change control.

[0235] Optionally, controlling the target vehicle to return to its original lane includes: determining the feasibility of reversing based on the distance between the rear wheels and the lane dividing line; if reversing is not feasible, generating a lane-keeping trajectory within the target lane and controlling the target vehicle to drive along the lane-keeping trajectory; if reversing is feasible, generating a reversal trajectory based on the original lane reference line; controlling the target vehicle to return to its original lane includes: controlling the target vehicle to return to its original lane according to the reversal trajectory.

[0236] Before controlling the target vehicle to return to its original lane, a feasibility assessment of the return is performed. Specifically, if the rear wheels of the target vehicle are in the original lane, the distance from the rear wheels of the target vehicle to the lane dividing line is determined, and this distance satisfies the following expression (21):

[0237] (twenty one)

[0238] In the above expression (21), It is the lateral distance of the rear wheel in the original lane relative to the lane dividing line (the lane line that needs to be crossed when changing lanes). The lateral distance of the target vehicle's center in the original lane relative to the lane divider line. To change lane direction, It is +1 or -1. Let be the width of the target vehicle body. Expression (21) represents the lateral position of the rear wheel edge relative to the lane dividing line, calculated by superimposing half the vehicle width on the lateral position of the target vehicle center.

[0239] like If preset return conditions are met, the target vehicle can return to its original lane. For example, the return conditions are: This indicates that the absolute value of the lateral distance from the rear wheel to the lane dividing line is no greater than half the lane width plus a preset margin (such as 0.2m). This refers to the original lane width. It's understandable that, before the rear wheels have overstepped into the target lane, and most of the vehicle is still on one side of the original lane, there is a physical possibility of returning to the original lane.

[0240] If the rear wheels of the target vehicle are in the target lane, determine the distance from the rear wheels of the target vehicle to the lane dividing line, which satisfies the following expression (21):

[0241] (twenty one)

[0242] In expression (21), The lateral distance of the rear wheels in the target lane relative to the lane divider line. The lateral distance of the target vehicle's center in the target lane relative to the lane divider.

[0243] like ,and This indicates that the rear wheels have crossed the lane dividing line and are inside the target lane. When this condition is met, returning to the original lane is prohibited, and priority should be given to maintaining the lane line that caused the collision and adjacent vehicles.

[0244] In this invention, if the return condition is met (i.e., reversal is feasible), a reversal trajectory is generated, and the vehicle returns to the original lane according to this trajectory. If the return condition is not met, reversal is not performed, and the system enters the necessary lane-change mode for navigation, generating a lane-keeping trajectory.

[0245] Specifically, the backtrack satisfies the following expression (22):

[0246] (twenty two)

[0247] In the above expression (22), Indicates the retrace trajectory, and indicates the lateral position. This represents the vertical distance. The lateral coordinates of the original lane reference line. For example, the preset offset to the edge. This is understandable; a fixed lateral offset of 0.5m is made based on the original lane reference line to generate the reversal trajectory.

[0248] For example, refer to Figure 5 The target vehicle originally attempted to change lanes from lane B to lane A along the lane change path. When a high-risk obstacle appeared in lane A and the vehicle had not crossed the lane boundary line, the lane change abort condition was triggered. The vehicle then abandoned its attempt to change lanes to lane A and generated a reversal trajectory. The target vehicle returned to its original lane B along this reversal trajectory to avoid a collision with the high-risk obstacle and achieve a safe reversal of the lane change process.

[0249] In this invention, if it is not possible to return to the original lane, i.e., reversing is not feasible, a driving trajectory that is driven to the side of the road is generated, which satisfies the following expression (23):

[0250] (twenty three)

[0251] In the above expression (23), This indicates the driving trajectory while keeping to the side of the road. The lateral coordinates of the lane center reference line for the target lane. This is the lane change direction symbol. This represents the offset to the edge of the target lane. It can be understood as being based on the lane center reference line, combined with the offset from the lane change direction. Generates a driving trajectory that is pulled over to the side of the target lane.

[0252] In this invention, if a risk of parallel obstacles is detected during lane changing, the lane change is not stopped, nor is the lane returned to the original lane. Instead, the lane change is maintained laterally until the parallel obstacle is eliminated, and then the lane change is completed after the obstacle is eliminated.

[0253] For example, refer to Figure 6 The target vehicle originally changed lanes from lane B to lane A according to the lane change path trajectory. When a high-risk parallel obstacle was detected in lane A, the lane change was not stopped or the lane was reversed. Instead, the vehicle switched to a lane-side driving trajectory, and the lateral position of the vehicle remained unchanged, waiting for the parallel risk to be eliminated. After the risk caused by the obstacle was reduced, the remaining lane change actions were performed to complete the lane change to lane A.

[0254] In the specific implementation process, the lane change holding trajectory in lane change holding mode satisfies the following expression (24):

[0255] (twenty four)

[0256] In the above expression (24), This indicates that the vehicle maintains its trajectory while changing lanes. The lateral coordinates of the lane center reference line for the target lane. To trigger the hold moment, the distance between the target vehicle's current lateral position and the lane center reference line is fixed.

[0257] In summary, when the target vehicle has the conditions to return to the original lane, it can smoothly resume the original lane; when the conditions for returning are insufficient, it can promptly switch to a conservative driving state, thereby matching the trajectory planning with the actual feasible space of the vehicle.

[0258] Optionally, it also includes: obtaining the remaining longitudinal distance for lane changing in navigation from vehicle operation data; when the remaining longitudinal distance for lane changing in navigation is less than the standard lane changing distance, determining the trajectory generation constraint parameters under the forced lane changing condition, the trajectory generation constraint parameters including the available longitudinal distance for lane changing, the maximum allowed vehicle speed, and the desired acceleration; and generating the lane changing path trajectory based on the trajectory generation constraint parameters and the lane center reference line.

[0259] Forced lane change conditions include: rapid lane change or continuous lane change.

[0260] In some implementations, a standard lane change distance is determined based on the current longitudinal speed of the target vehicle, which satisfies the following expression (25):

[0261] (25)

[0262] In the above expression (25), This is the standard lane change distance.

[0263] Among them, when the remaining longitudinal distance of the navigation lane change Less than this standard lane change distance When the lane change is initiated, the system enters rapid lane change mode. The rapid lane change distance can be determined using the longitudinal distance, and can be based on... Determine the rapid lane change distance. It should be understood that when generating the lane change path trajectory, the distance the target vehicle needs to move longitudinally during the rapid lane change process is the rapid lane change distance.

[0264] For example, refer to Figure 7 The target vehicle is currently traveling in lane B and needs to exit the main road from the ramp exit. When the target vehicle enters the set distance range from the ramp exit, the rapid lane change strategy is activated, a corresponding lane change path trajectory is generated, and the vehicle is controlled to quickly change to lane A along the lane change path trajectory to ensure that the vehicle completes the lane change before reaching the ramp exit and to prevent missing the exit.

[0265] In some implementations, the number of lane changes ,and 2 When this occurs, continuous lane changing is triggered, where the continuous lane changing distance can be the longitudinal distance, which can be based on... It is confirmed that, among them, This is the preset minimum distance for a single lane change, such as 15m. It should be understood that when generating the lane change path trajectory, the distance the target vehicle needs to move longitudinally during the continuous lane change process is the continuous lane change distance.

[0266] For example, refer to Figure 8The target vehicle is traveling in straight lane C. The navigation plan requires it to turn right at the intersection, so it needs to change to right turn lane A. When the target vehicle is within a set distance from the intersection, a lane change path trajectory is generated, and the vehicle is controlled to cross straight lane B in sequence to complete the continuous lane change from straight lane C to right turn lane A, so that the vehicle can enter the right turn lane in advance and meet the driving requirements for turning right at the intersection.

[0267] During lane changing, the maximum permissible speed satisfies the following expression (26):

[0268] (26)

[0269] In the above expression (26), The maximum permissible speed during a lane change. This is the preset minimum longitudinal speed, such as 5 m / s. For a safety margin, it can be determined based on the current longitudinal velocity, such as... . This is a preset time margin, such as 12 seconds.

[0270] During lane change, the desired acceleration Satisfy the following expression (27):

[0271] (27)

[0272] in, .

[0273] In this invention, the available longitudinal distance for lane changing, the maximum permissible vehicle speed, and the desired acceleration can be used as trajectory generation constraint parameters to generate an offset point sequence. Then, the offset point sequence is subjected to coordinate transformation based on the lane center reference line to obtain the lane changing path trajectory. The specific method for generating the offset point sequence based on the trajectory generation constraint parameters is not limited. Furthermore, the specific process of performing coordinate transformation on the offset point sequence based on the lane center reference line to obtain the lane changing path trajectory can be referred to the aforementioned related content and will not be repeated here.

[0274] In summary, in this invention, the lane change path trajectory can be adaptively adjusted based on the remaining navigation distance and the need for continuous lane changes. The speed constraint during forced lane changes can also be synchronized with the trajectory planning, thereby improving the executability and control consistency of lane changes in scenarios where the remaining navigation distance is limited.

[0275] Optionally, it also includes: if a preset triggering condition is detected during the lane change process, the lane change path trajectory is adjusted; wherein the preset triggering condition includes at least one of the following: there is a visual merging point ahead, the vehicle in front of the target lane decelerates, and the lateral safety distance between the target vehicle and surrounding vehicles is insufficient.

[0276] When a preset trigger condition is detected, the lane change offset point sequence is dynamically replanned to further adjust the lane change path trajectory.

[0277] Specifically, when a visual merging point is detected ahead (such as a ramp merging), and the target vehicle is close to the merging point, the lane change time is shortened and the lane change is accelerated to avoid changing lanes near the merging point. When a vehicle in the target lane is detected to be slowing down during the lane change process, and the target vehicle needs significant acceleration, the lane change time is shortened, the lane change trajectory becomes sharper, and the vehicle enters the target lane more quickly to facilitate longitudinal following. When the lateral safety distance between the target vehicle and surrounding vehicles is insufficient, and the lateral position of the lane change trajectory needs to be adjusted (such as to avoid obstacles at the edge of the target lane), the lane center reference line is re-determined and the offset point sequence is regenerated.

[0278] For example, refer to Figure 9 The target vehicle is traveling in lane A and needs to change lanes to lane B. When the target vehicle travels to a set distance from the merging point, trajectory replanning is triggered. In the figure, the thin dashed line is the path before replanning, and the thick dashed line is the path after replanning. After replanning, the lane change path is adjusted to generate a new lane change path that is adapted to the merging scenario, so that the target vehicle can complete the lane change to lane B before reaching the merging point.

[0279] In this embodiment of the invention, adjusting the lane-change path trajectory includes: saving the original total lane-change time; calculating the remaining lane-change distance and remaining lane-change time based on the triggering conditions; calculating the new lane-change time based on the remaining lane-change distance and the adjustment coefficient; regenerating the offset point sequence using the target vehicle's current position as the starting point and the adjusted lane-change time; then regenerating the lane-change path trajectory based on the offset point sequence, and controlling the target vehicle to continue changing lanes according to the new lane-change path trajectory. A smooth transition mechanism can be used when regenerating the lane-change path trajectory to avoid trajectory jumps caused by replanning.

[0280] In summary, this invention uses real-time triggering and identification of visual convergence points, vehicle deceleration in front of the target lane, and insufficient lateral safety distance to adjust the lane change trajectory online accordingly. This allows the trajectory to be updated synchronously with changes in road geometry and the status of surrounding vehicles, making the lane change control process more in line with the real-time traffic environment and improving trajectory continuity and lane change execution stability.

[0281] The following are embodiments of the apparatus of the present invention, which can be used to execute embodiments of the method of the present invention. For details not disclosed in the embodiments of the apparatus of the present invention, please refer to the embodiments of the method of the present invention.

[0282] Figure 10 This is a schematic diagram of the vehicle lane-changing device provided in an embodiment of the present invention. Figure 10As shown, the vehicle lane changing device 10 includes an acquisition module 1001, a first determination module 1002, a second determination module 1003, and a control module 1004.

[0283] The acquisition module 1001 is used to acquire vehicle operation data of the target vehicle and multiple initial boundary lines of the target lane to be changed in response to the received lane change decision signal.

[0284] The first determining module 1002 is used to determine the target boundary line that meets the quality requirements among multiple initial boundary lines, and to laterally offset the target boundary line to obtain the lane center reference line of the target lane.

[0285] The second determining module 1003 is used to determine the lane change path trajectory based on vehicle operation data and lane center reference line;

[0286] The control module 1004 is used to control the target vehicle to travel according to the lane change path trajectory.

[0287] Furthermore, the first determining module 1002 determines a target boundary line that meets the quality requirements among multiple initial boundary lines. Specifically, it is used to: determine at least one of the length quality, stability quality, and continuity quality of each initial boundary line; determine the comprehensive quality of the initial boundary line based on at least one of the length quality, stability quality, and continuity quality; and determine the target boundary line that meets the quality requirements among multiple initial boundary lines based on the comprehensive quality.

[0288] Furthermore, the first determining module 1002 determines a target boundary line that meets the quality requirements from multiple initial boundary lines based on the overall quality. Specifically, it is used to: select the initial boundary line with the highest overall quality as a candidate boundary line from multiple initial boundary lines; determine the first lateral position of the candidate boundary line at a preset longitudinal distance, and the second lateral position of the target boundary line in the previous frame at a preset longitudinal distance ahead; and determine the current target boundary line from the candidate boundary line and the target boundary line in the previous frame based on the lateral deviation between the first lateral position and the second lateral position.

[0289] Furthermore, the first determining module 1002 determines the current target boundary line from the candidate boundary line and the target boundary line of the previous frame based on the lateral deviation between the first lateral position and the second lateral position. Specifically, it is used to: determine the target boundary line of the previous frame as the current target boundary line when the lateral deviation is greater than the lateral deviation threshold; determine the longitudinal deviation between the endpoint of the candidate boundary line and the endpoint of the target boundary line of the previous frame when the lateral deviation is less than or equal to the lateral deviation threshold; if the longitudinal length of the candidate boundary line is greater than the longitudinal length of the target boundary line of the previous frame, the longitudinal deviation is greater than or equal to the longitudinal deviation threshold, and the longitudinal length of the target boundary line of the previous frame is less than the longitudinal length threshold, determine the candidate boundary line as the current target boundary line; otherwise, determine the target boundary line of the previous frame as the current target boundary line.

[0290] Furthermore, the second determining module 1003 is specifically used for: acquiring the target lane change style and determining the target lane change duration based on the target lane change style; generating an offset point sequence for switching to the target lane based on the target lane change duration, vehicle operation data, and lane center reference line; and performing coordinate transformation processing on the offset point sequence based on the lane center reference line to obtain the lane change path trajectory.

[0291] Furthermore, it also includes a third determining module (not shown in the figure), which is used to obtain the following vehicle operation data of the vehicles behind the target vehicle in the target lane; determine the yield probability of the following vehicles based on the following vehicle operation data and the vehicle operation data; trigger lane change abort if the yield probability is less than the probability threshold; and determine the target lane change style based on the yield probability if the yield probability is greater than or equal to the probability threshold.

[0292] Furthermore, it also includes a lane change abort module (not shown in the figure): used to detect the first risk level of the target lane and the second risk level of the original lane during the lane change process; if the first risk level is higher than the first risk level threshold and the second risk level is lower than the second risk level threshold, it triggers lane change abort and controls the target vehicle to return to the original lane.

[0293] Furthermore, it also includes a trajectory generation module (not shown in the figure): used to determine the feasibility of reversing based on the distance of the rear wheels relative to the lane dividing line; if reversing is not feasible, it generates a driving trajectory to the side of the target lane and controls the target vehicle to drive according to the driving trajectory to the side of the lane; if reversing is feasible, it generates a reversal trajectory based on the original lane reference line; and controls the target vehicle to return to the original lane, including: controlling the target vehicle to return to the original lane according to the reversal trajectory.

[0294] Furthermore, the second determining module 1003 is specifically used for: obtaining the remaining longitudinal distance of the navigation lane change from the vehicle operation data; when the remaining longitudinal distance of the navigation lane change is less than the standard lane change distance, determining the trajectory generation constraint parameters under the forced lane change condition, the trajectory generation constraint parameters including the available longitudinal distance of the lane change, the maximum allowed vehicle speed and the expected acceleration; and generating the lane change path trajectory based on the trajectory generation constraint parameters and the lane center reference line.

[0295] Furthermore, it also includes a trajectory adjustment module (not shown in the figure): used to adjust the lane change path trajectory if a preset trigger condition is detected during the lane change process; wherein the preset trigger condition includes at least one of the following: there is a visual merging point ahead, the vehicle in front of the target lane decelerates, and the lateral safety distance between the target vehicle and surrounding vehicles is insufficient.

[0296] The vehicle lane-changing device provided in this embodiment of the invention can be used to execute the method steps of the above method embodiment. The specific implementation and technical effects are similar, and will not be repeated here.

[0297] The model training apparatus provided in this embodiment of the invention can be used to execute the method steps of the above method embodiments. The specific implementation and technical effects are similar, and will not be repeated here.

[0298] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 110 includes at least one processor 1101 and a memory 1102.

[0299] Optionally, the electronic device 110 also includes a communication component 1103. The processor 1101, memory 1102, and communication component 1103 are connected via a bus 1104.

[0300] In a specific implementation, at least one processor 1101 executes computer execution instructions stored in memory 1102, causing at least one processor 1101 to perform the above-described method.

[0301] The specific implementation process of processor 1101 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0302] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0303] The memory may include random access memory (RAM) and non-volatile memory (NVM), such as at least one disk storage device.

[0304] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0305] This invention also provides a vehicle, including a vehicle body and the aforementioned electronic equipment.

[0306] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0307] This invention also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the aforementioned model training method.

[0308] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as ferromagnetic random access memory (FRAM), static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic surface memory, flash memory, magnetic disk, optical disk, or compact disc read-only memory (CD-ROM), etc. The readable storage medium can be any available medium accessible to general-purpose or special-purpose computers.

[0309] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0310] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0311] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0312] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0313] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0314] Those skilled in the art will understand that the present invention may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0315] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0316] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0317] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0318] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0319] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for changing lanes for vehicles, characterized in that, include: In response to the received lane change decision signal, the vehicle operation data of the target vehicle and multiple initial boundary lines of the target lane to be changed are obtained. Among the multiple initial boundary lines, a target boundary line that meets the quality requirements is determined, and the target boundary line is laterally offset to obtain the lane center reference line of the target lane; The lane change path trajectory is determined based on the vehicle operation data and the lane center reference line; Control the target vehicle to travel along the lane-changing path trajectory.

2. The method according to claim 1, characterized in that, The step of determining the target boundary line that meets the quality requirements among the plurality of initial boundary lines includes: For each initial boundary line, determine at least one of the following: length mass, stability mass, and continuity mass. The overall quality of the initial boundary line is determined based on at least one of the length quality, stability quality, and continuity quality. Based on the overall quality, a target boundary line that meets the quality requirements is determined among the plurality of initial boundary lines.

3. The method according to claim 2, characterized in that, The step of determining a target boundary line that meets the quality requirements from among the plurality of initial boundary lines based on the comprehensive quality includes: Among the multiple initial boundary lines, the initial boundary line with the highest comprehensive quality is selected as the candidate boundary line; Determine the first lateral position of the candidate boundary line at a preset vertical distance, and the second lateral position of the target boundary line in the previous frame at the preset vertical distance in front; Based on the lateral deviation between the first lateral position and the second lateral position, the current target boundary line is determined among the candidate boundary lines and the target boundary line of the previous frame.

4. The method according to claim 3, characterized in that, The step of determining the current target boundary line among the candidate boundary lines and the target boundary line of the previous frame based on the lateral deviation between the first lateral position and the second lateral position includes: If the lateral deviation is greater than the lateral deviation threshold, the target boundary line of the previous frame is determined as the current target boundary line; If the lateral deviation is less than or equal to the lateral deviation threshold, the longitudinal deviation between the endpoint of the candidate boundary line and the endpoint of the target boundary line in the previous frame is determined. If the longitudinal length of the candidate boundary line is greater than the longitudinal length of the target boundary line in the previous frame, the longitudinal deviation is greater than or equal to the longitudinal deviation threshold, and the longitudinal length of the target boundary line in the previous frame is less than the longitudinal length threshold, the candidate boundary line is determined to be the current target boundary line; otherwise, the target boundary line in the previous frame is determined to be the current target boundary line.

5. The method according to any one of claims 1 to 4, characterized in that, The step of determining the lane-changing path trajectory based on the vehicle operation data and the lane center reference line includes: Obtain the target lane change style and determine the target lane change duration based on the target lane change style; Based on the target lane change duration, the vehicle operation data, and the lane center reference line, a sequence of offset points for switching to the target lane is generated; The lane change path trajectory is obtained by performing coordinate transformation on the offset point sequence based on the lane center reference line.

6. The method according to claim 5, characterized in that, Also includes: Acquire the following vehicle movement data of the vehicles behind the target vehicle in the target lane; Based on the following vehicle's running data and the vehicle's running data, determine the probability of the following vehicle yielding; If the probability of yielding is less than the probability threshold, the lane change is aborted. If the yield probability is greater than or equal to a probability threshold, the target lane change style is determined based on the yield probability.

7. The method according to any one of claims 1 to 4, characterized in that, Also includes: During lane changing, the first risk level of the target lane and the second risk level of the original lane are detected; If the first risk level is higher than the first risk level threshold and the second risk level is lower than the second risk level threshold, the lane change is aborted and the target vehicle is controlled to return to its original lane.

8. The method according to claim 7, characterized in that, Also includes: The feasibility of reversing is determined based on the distance between the rear wheels and the lane dividing line. If reversing is not feasible, a driving trajectory for staying on the side of the target lane is generated, and the target vehicle is controlled to drive according to the driving trajectory for staying on the side of the lane. If reversal is feasible, a reversal trajectory is generated based on the original lane reference line; The step of controlling the target vehicle to return to its original lane includes: controlling the target vehicle to return to its original lane according to the reversal trajectory.

9. The method according to any one of claims 1 to 4, characterized in that, The step of determining the lane-changing path trajectory based on the vehicle operation data and the lane center reference line includes: Obtain the remaining longitudinal distance for lane changing during navigation from the vehicle operation data; When the remaining longitudinal distance for lane change in the navigation is less than the standard lane change distance, the trajectory generation constraint parameters under the forced lane change condition are determined. The trajectory generation constraint parameters include the available longitudinal distance for lane change, the maximum allowed vehicle speed, and the desired acceleration. The lane change path trajectory is generated based on the trajectory generation constraint parameters and the lane center reference line.

10. The method according to any one of claims 1 to 4, characterized in that, Also includes: If a preset trigger condition is detected during the lane change process, the lane change path trajectory will be adjusted. The preset triggering conditions include at least one of the following: the presence of a visual confluence point ahead, the vehicle ahead of the target lane slowing down, and insufficient lateral safety distance between the target vehicle and surrounding vehicles.

11. A vehicle lane-changing device, characterized in that, The device includes: The acquisition module is used to acquire vehicle operation data of the target vehicle and multiple initial boundary lines of the target lane to be changed in response to the received lane change decision signal. The first determining module is used to determine a target boundary line that meets the quality requirements among the multiple initial boundary lines, and to laterally offset the target boundary line to obtain the lane center reference line of the target lane. The second determining module is used to determine the lane change path trajectory based on the vehicle operation data and the lane center reference line; The control module is used to control the target vehicle to travel according to the lane-changing path trajectory.

12. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1 to 10.

13. A vehicle, characterized in that, Includes the vehicle body and the electronic equipment as described in claim 12.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 10.

15. A computer program product, characterized in that, include: A computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 10.

Images