A method, apparatus, device, and storage medium for changing speed and lanes.

Abstract

This application discloses a method, apparatus, device, and storage medium for lane changing, relating to the field of autonomous driving technology. The method includes: determining a first vehicle distance between two adjacent vehicles in the vehicle's lane and a second vehicle distance between two adjacent vehicles in a target lane based on the vehicle's own driving data and the driving data of surrounding vehicles; the target lane is the lane to be changed; acquiring driving prediction data for surrounding vehicles and the vehicle's acceleration output by a novel gating model; the novel gating model takes the driving data of surrounding vehicles as input; and determining the vehicle's lane-changing trajectory and target acceleration based on the first vehicle distance, the second vehicle distance, the driving prediction data, and the acceleration, so as to perform lane changing according to the target acceleration and the lane-changing trajectory. This method can control the vehicle's acceleration or deceleration to complete the lane-changing process, improving the comfort of the lane-changing process and increasing the success rate of lane changes.

Description

Technical Field

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

[0002] Currently, the vehicle speed change and lane change process for autonomous vehicles is not yet perfect. During the speed change and lane change process, the change in the vehicle's speed may affect the driving of other vehicles. In addition, the comfort of the vehicle during lane change is relatively low. Therefore, how to improve the comfort of the vehicle during lane change and further improve the success rate of lane change are urgent problems to be solved. Summary of the Invention

[0003] In view of this, the purpose of this invention is to provide a method, apparatus, device, and medium for variable speed lane changing, capable of controlling vehicle acceleration or deceleration to complete the lane changing process, improving lane changing comfort, and increasing lane changing success rate. The specific solution is as follows: In a first aspect, this application discloses a speed-changing and lane-changing method, applied to autonomous vehicles, comprising: Based on the vehicle's own driving data and the driving data of surrounding vehicles, determine the first vehicle distance between two adjacent vehicles in the vehicle's own lane and the second vehicle distance between two adjacent vehicles in the target lane; the target lane is the lane where the lane change is to be made. The novel gating model outputs driving prediction data for surrounding vehicles and acceleration data for the vehicle itself; the novel gating model takes driving data of surrounding vehicles as input. Based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration, the lane-changing trajectory and target acceleration of the vehicle are determined so that the vehicle can change speed and lanes according to the lane-changing trajectory based on the target acceleration.

[0004] Optionally, before determining the first vehicle distance between two adjacent vehicles in the self-lane and the second vehicle distance between two adjacent vehicles in the target lane based on the self-vehicle's driving data and the driving data of surrounding vehicles, the method further includes: Determine whether the conditions for starting a lane change with increased speed are met; the conditions for starting a lane change with increased speed include the presence of vehicles in the vicinity, the radius of the road curve being less than a preset radius threshold, and the lane being changeable. If the conditions are met, then the steps of determining the first vehicle distance between two adjacent vehicles in the vehicle lane and the second vehicle distance between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles are executed.

[0005] Optionally, determining the first vehicle distance between two adjacent vehicles in the vehicle's own lane and the second vehicle distance between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles includes: The lane change type is determined based on lane change behavior data, and a lane change instruction is generated based on the lane change type; the lane change type includes left lane change and right lane change; The target lane is determined based on the lane change instruction, and the driving data of vehicles in the target lane is extracted from the driving data of surrounding vehicles. Based on the vehicle driving data in the self-lane, determine the first vehicle distance between two adjacent vehicles in the self-lane; Based on the vehicle driving data of the target lane, determine the second vehicle spacing between two adjacent vehicles in the target lane.

[0006] Optionally, acquiring the driving prediction data for surrounding vehicles and the acceleration for the vehicle itself output by the novel gating model includes: Obtain driving prediction data for surrounding vehicles output by the first novel gating model; the first novel gating model takes the driving data of the surrounding vehicles in the previous N cycles as input and the predicted driving data of the surrounding vehicles in the next T cycles as output. Obtain the acceleration of the vehicle as output by the second novel gating model; the second novel gating model takes the driving data of the surrounding vehicles for the previous N cycles and the target acceleration as input, and takes the acceleration of the vehicle as output.

[0007] Optionally, determining the lane-changing trajectory and target acceleration of the vehicle based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration includes: Based on the distance to the vehicle, the second vehicle spacing is selected in order from closest to farthest for lane change verification, and the second target vehicle spacing that passes the verification is obtained. Select a first target vehicle distance corresponding to the second target vehicle distance, determine the initial lane-changing trajectory of the vehicle based on the first target vehicle distance and the second target vehicle distance, and determine the initial acceleration; Based on the driving prediction data, determine whether there is a collision risk in the initial lane change trajectory, and determine the lane change trajectory based on the judgment result; The initial acceleration is adjusted according to the acceleration to obtain the target acceleration.

[0008] Optionally, adjusting the initial acceleration based on the acceleration to obtain the target acceleration includes: The weights corresponding to the acceleration and the initial acceleration are determined based on the emphasis of the lane change behavior. The target acceleration is obtained by calculating the weighted sum of the acceleration and the initial acceleration.

[0009] Optionally, after determining the lane-changing trajectory and target acceleration of the vehicle, the method further includes: Update the vehicle control flag; the vehicle control flag indicates whether the autonomous vehicle is in the gear change and lane change control phase. If the vehicle is in the speed change and lane change control phase, the vehicle will determine in real time whether there is a collision risk during the speed change and lane change process based on the driving prediction data of surrounding vehicles output by the new gating model. If there is a collision risk, the target acceleration will be adjusted.

[0010] Secondly, this application discloses a speed-changing and lane-changing device for use in autonomous vehicles, comprising: The vehicle spacing determination module is used to determine the first vehicle spacing between two adjacent vehicles in the vehicle's own lane and the second vehicle spacing between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles; the target lane is the lane to be changed. The driving prediction data acquisition module is used to acquire driving prediction data for surrounding vehicles and acceleration for the vehicle itself, output by the novel gating model; the novel gating model takes the driving data of surrounding vehicles as input. The speed change and lane change parameter determination module is used to determine the lane change trajectory and target acceleration of the vehicle based on the first vehicle spacing, the second vehicle spacing, the driving prediction data and the acceleration, so as to perform speed change and lane change according to the target acceleration and the lane change trajectory.

[0011] Thirdly, this application discloses an electronic device, including: Memory, used to store computer programs; A processor is used to execute the computer program to implement the aforementioned speed change and lane change method.

[0012] Fourthly, this application discloses a computer-readable storage medium for storing a computer program; wherein the computer program, when executed by a processor, implements the aforementioned speed-changing and lane-changing method.

[0013] In this application, based on the vehicle's driving data and the driving data of surrounding vehicles, a first vehicle distance between two adjacent vehicles in the vehicle's lane and a second vehicle distance between two adjacent vehicles in the target lane are determined; the target lane is the lane to be changed; driving prediction data for surrounding vehicles and acceleration for the vehicle are obtained from the output of a novel gating model; the novel gating model takes the driving data of surrounding vehicles as input; based on the first vehicle distance, the second vehicle distance, the driving prediction data, and the acceleration, the vehicle's lane-changing trajectory and target acceleration are determined, so that the vehicle can change speed and lanes according to the target acceleration and the lane-changing trajectory. As can be seen, by comprehensively analyzing the lane-changing trajectory and target acceleration of the vehicle based on the first vehicle distance between two adjacent vehicles in the self-lane, the second vehicle distance between two adjacent vehicles in the target lane, and the driving prediction data and acceleration of surrounding vehicles output by the new gating model, the vehicle can be controlled to accelerate or decelerate to complete the lane-changing process. The lane-changing process does not affect the driving of other vehicles. Furthermore, the acceleration and deceleration determined by the new gating model make the entire lane-changing process smoother and improve the comfort of the lane-changing process. The lane-changing trajectory is determined by predicting the trajectory of surrounding vehicles based on the new gating model, which greatly improves the success rate of lane-changing. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0015] Figure 1 A flowchart of a speed change and lane change method provided in this application; Figure 2 This application provides a specific schematic diagram of speed change and lane change; Figure 3 A schematic diagram of a specific NGCU model structure is provided for this application; Figure 4 This application provides a schematic diagram of a specific speed change and lane change system structure; Figure 5 This application provides a schematic diagram of a speed-changing and lane-changing device. Figure 6 This application provides a structural diagram of an electronic device. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] In existing technologies, the vehicle speed change and lane change process for fully automated vehicles is not yet perfect. During the speed change and lane change process, the change in vehicle speed may affect the driving of other vehicles, and the comfort of the vehicle during lane change is relatively low. To overcome the above technical problems, this application proposes a speed change and lane change method that can control the vehicle to accelerate or decelerate to complete the lane change process, thereby improving the comfort of the lane change process and increasing the success rate of lane change.

[0018] This application discloses a speed-changing and lane-changing method. See also Figure 1 As shown, the method may include the following steps: Step S11: Based on the vehicle's driving data and the driving data of surrounding vehicles, determine the first vehicle distance between two adjacent vehicles in the vehicle's lane and the second vehicle distance between two adjacent vehicles in the target lane; the target lane is the lane to be changed.

[0019] In this embodiment, the vehicle's driving data and the driving data of surrounding vehicles are first acquired. Then, based on the acquired driving data, a first vehicle distance between two adjacent vehicles in the vehicle's lane and a second vehicle distance between two adjacent vehicles in the target lane of the lane to be changed are determined. Specifically, the surrounding vehicle data, including but not limited to speed, position, and acceleration information, can be obtained from the vehicle's intelligent forward-facing camera and front and rear corner radars. This allows for the collection of driving data within a preset distance in front and behind the vehicle. The first and second vehicle distances are longitudinal distances. For example... Figure 2 As shown in the figure, when the autonomous vehicle needs to change lanes to the left, it collects vehicle driving data within a range of approximately 120m before and after the vehicle in its own lane and the target lane, and calculates all vehicle distances Dxi and Vxi. Among them, Dx1, Dx2, and Dx3 are the second vehicle distances in the target lane, and Dx4 and Dx5 are the first vehicle distances in the vehicle's own lane.

[0020] In this embodiment, determining the first vehicle spacing between two adjacent vehicles in the vehicle's own lane and the second vehicle spacing between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles may include: determining the lane change type based on lane change behavior data and generating a lane change command based on the lane change type; the lane change type includes changing lanes to the left and changing lanes to the right; determining the target lane based on the lane change command and extracting the target lane vehicle driving data from the driving data of surrounding vehicles; determining the first vehicle spacing between two adjacent vehicles in the vehicle's own lane based on the vehicle driving data of the vehicle's own lane; and determining the second vehicle spacing between two adjacent vehicles in the target lane based on the vehicle driving data of the target lane. Understandably, the system determines whether a lane change is to the left or right based on lane change data from lever-operated or active lane change, thus identifying the target lane. If a left lane change is required, the left lane becomes the target lane. Subsequent calculations only require the driving data of vehicles in the user's own lane and the vehicles in the left lane; the driving data of vehicles in the right lane can be disregarded. Then, based on the user's own driving data and the driving data of other vehicles in the user's own lane, the first vehicle distance between two adjacent vehicles in the user's own lane is determined. Based on the driving data of vehicles in the target lane, the second vehicle distance between two adjacent vehicles in the target lane is determined.

[0021] In this embodiment, before determining the first vehicle distance between two adjacent vehicles in the vehicle's own lane and the second vehicle distance between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles, the method may further include: determining whether the conditions for initiating a lane change are met; the conditions for initiating a lane change include the presence of surrounding vehicles, a road curve radius less than a preset radius threshold, and the current lane being changeable; if these conditions are met, then the step of determining the first vehicle distance between two adjacent vehicles in the vehicle's own lane and the second vehicle distance between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles is executed. That is, before initiating a lane change, it is necessary to determine whether the conditions for initiating a lane change are met, which include the presence of surrounding vehicles, a road curve radius less than a preset radius threshold, and the current lane being changeable. If there are no vehicles around the vehicle, the lane change will proceed directly without initiating the speed change / lane change procedure. If the radius of the road curve is greater than a preset radius threshold, meaning the road curvature is small and it is a sharp turn, then it is not suitable to change lanes, and the speed change / lane change operation will not be performed. If there are solid lines on both sides of the lane, then according to traffic rules, lane change is not allowed, so the speed change / lane change operation will not be performed. Specifically, the above-mentioned speed change / lane change initiation conditions can be determined using the safety check module. This module can be responsible for outputting whether there is a collision risk in the surrounding area. Performing the above-mentioned condition determination before speed change / lane change can save computing resources.

[0022] Step S12: Obtain the driving prediction data for surrounding vehicles and the acceleration of the vehicle itself output by the new gating model; the new gating model takes the driving data of surrounding vehicles as input.

[0023] In this embodiment, during lane change and acceleration shifting, the vehicle acquires driving prediction data for surrounding vehicles and acceleration for its own vehicle, output by a novel gating model. The novel gating model is an NGCU (New Gate Control Unit) model, which takes the driving data of surrounding vehicles as input. It can be understood that, as a continuously trained model, the novel gating model, by predicting the driving of surrounding vehicles, can effectively guide the lane change trajectory and acceleration during lane change and acceleration shifting.

[0024] In this embodiment, obtaining the driving prediction data for surrounding vehicles and the acceleration of the vehicle itself output by the novel gating model may include: obtaining the driving prediction data for surrounding vehicles output by a first novel gating model; the first novel gating model takes the driving data of the surrounding vehicles for the previous N cycles as input and the predicted driving data of the surrounding vehicles for the next T cycles as output; obtaining the acceleration of the vehicle itself output by a second novel gating model; the second novel gating model takes the driving data of the surrounding vehicles for the previous N cycles and a target acceleration as input and the acceleration of the vehicle itself as output; the target acceleration is... The target acceleration of the vehicle Specifically, it includes a first novel gating model and a second novel gating model. The first novel gating model takes the driving data of surrounding vehicles for the previous N periods as input and outputs the predicted driving data of surrounding vehicles for the next T periods. In other words, the first novel gating model, NGCUⅠ, as a time series prediction model, can predict potential collisions in advance by performing lane change planning based on its output driving prediction data, and then adjust the lane change trajectory accordingly, further improving the success rate of lane changes. The second novel gating model takes the driving data of surrounding vehicles for the previous N periods and the target acceleration as input and outputs the acceleration relative to the vehicle itself. That is, NGCU model Ⅱ, as a continuously trained model, can learn the speed range that the vehicle can accept, and when a lane change occurs, it can appropriately adjust the acceleration to achieve a smooth lane change while ensuring the success of the lane change.

[0025] Specifically, the structure of the above NGCU model is as follows: Figure 3 As shown, the forgetting gate Its function is to remove data that is not highly similar to the current time step, that is, to remove data that has little impact on the result. The formula is: ;in It is the input vector at the t-th time step. With weight matrix Multiplication performs a linear transformation; Store the data from the previous time step t-1. By using the weight matrix The product undergoes a linear transformation, adding the two data parts together, and then the result is activated by the Sigmoid function. The calculation is performed, and the output is between 0 and 1. It is a constant. Input gate Selectively retain data with higher features at the current time, using the following formula: ,in It is the input vector at the t-th time step. With weight matrix Multiplication performs a linear transformation; Store the data from the previous time step t-1. By using the weight matrix Multiplication performs a linear transformation, adding the two data parts together. It is a constant; and it is related to the forget gate. In comparison, NGCU's input gates Introducing data stream information from the previous time period Therefore, the input gates have a memory function for retaining the current time data; the NGCU's input gates use... Sigmoid Activation function; input gate when outputting results. Increased Module, its use tanh As an activation function, compared to Sigmoid Activation function, when x When the target value is between -3 and 3 y The changes are more pronounced, improving the sensitivity of learning; To retain data with higher features at the current moment, the result is equal to ,in, .

[0026] Step S13: Determine the lane-changing trajectory and target acceleration of the vehicle based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration, so as to change speed and lanes according to the lane-changing trajectory based on the target acceleration.

[0027] In this embodiment, based on the first and second vehicle distances calculated from the current actual road conditions, and combined with the driving prediction data and acceleration output by the NGCU model, the final lane-changing trajectory and target acceleration of the vehicle are determined. The acceleration and deceleration learned by the novel gating model are then combined to make the entire lane-changing process smoother and improve comfort. Furthermore, the lane-changing trajectory is determined by predicting the trajectories of surrounding vehicles using the novel gating model, which significantly improves the success rate of lane changes. The aforementioned acceleration can be negative, meaning the vehicle can choose to decelerate and change lanes based on the situation, for example... Figure 2As shown, if the lane change trajectory determined for the autonomous vehicle is from Dx4 to Dx3, then acceleration is required for the lane change; if the determined lane change trajectory is from Dx5 to Dx1, then deceleration is required for the lane change.

[0028] In this embodiment, determining the lane-changing trajectory and target acceleration of the vehicle based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration may include: selecting second vehicle spacings sequentially from closest to farthest based on the distance to the vehicle for lane-changing verification to obtain a second target vehicle spacing that passes the verification; selecting a first target vehicle spacing corresponding to the second target vehicle spacing; determining the initial lane-changing trajectory of the vehicle based on the first target vehicle spacing and the second target vehicle spacing, and determining the initial acceleration; judging whether there is a collision risk in the initial lane-changing trajectory based on the driving prediction data, and determining the lane-changing trajectory based on the judgment result; adjusting the initial acceleration based on the acceleration to obtain the target acceleration. That is, firstly, the initial lane-changing trajectory of the vehicle is determined based on the first vehicle spacing and the second vehicle spacing, and the initial acceleration is determined; then, the final lane-changing trajectory is adjusted and determined based on the driving prediction data output by the novel gating model, and the initial acceleration is adjusted based on the acceleration output by the novel gating model to obtain the target acceleration.

[0029] Specifically, when determining the initial lane-changing trajectory of the vehicle based on the first and second vehicle spacings, and determining the initial acceleration, the second vehicle spacing is selected sequentially from closest to farthest distances for lane-changing verification, resulting in a second target vehicle spacing that passes the verification. For example... Figure 2 As shown, according to the workshop data Dx i and Vx i Based on the calculation results, the system selects the interval that the safety detection module can pass through. For example, Dx2 is an interval, but the safety detection module cannot pass the verification of this interval, meaning that changing lanes from this interval may lead to a collision and is dangerous. In this case, it will calculate Dx1 or Dx3, specifically selecting the calculation order according to the distance. If Dx3 can be passed by the safety detection module, the next step is to select the first target vehicle distance corresponding to the second target vehicle distance, that is, the first target vehicle distance that needs to be passed to reach the second target vehicle distance. For example, if the second target vehicle distance is Dx3, then the first target vehicle distance is Dx4. Then, it calculates whether the target position is reachable, that is, whether the first target vehicle distance is long enough. For example, if the distance to the vehicle in front in the lane is close, it is necessary to accelerate when changing lanes, but the vehicle in front may brake during acceleration, which may cause a collision. To avoid a collision, Dx3 will be discarded, and the next Dx3 will be calculated. i until a specific Dx is selected. i Alternatively, all lanes can be traversed to obtain the initial lane change trajectory.

[0030] In this embodiment, adjusting the initial acceleration based on the acceleration to obtain the target acceleration may include: determining the weights corresponding to the acceleration and the initial acceleration based on the emphasis of the lane change behavior; and obtaining the target acceleration by calculating the weighted sum of the acceleration and the initial acceleration. It is understood that, without considering the NGCU model, the calculated initial acceleration... It is the most efficient However, this can lead to sudden acceleration or deceleration of the vehicle, resulting in a lack of comfort. Introducing the NGCU model prioritizes comfort in its output; therefore, the output acceleration of the NGCU model... It is related to the initial acceleration. The target acceleration of the final output is different from that of the target acceleration. .parameter and It is an adjustable parameter. The larger the size, the more efficient the lane-changing behavior; conversely, the smaller the size, the more comfortable the behavior.

[0031] In this embodiment, after determining the lane-changing trajectory and target acceleration of the autonomous vehicle, the process may further include: updating the vehicle control flag; the vehicle control flag indicates whether the autonomous vehicle is in the lane-changing control phase; if the vehicle is in the lane-changing control phase, then based on the driving prediction data for surrounding vehicles output by the novel gating model, it is determined in real time whether there is a collision risk during the lane-changing process; if so, the target acceleration is adjusted. That is, immediately after initiating the lane-changing, after calculating the lane-changing trajectory and target acceleration according to the above process, the vehicle control flag is updated. During the remaining lane-changing process, based on the driving prediction data for surrounding vehicles output by the novel gating model, it is continuously determined in real time whether there is a collision risk during the lane-changing process; if so, the target acceleration is adjusted, and final vehicle control is achieved by correcting the acceleration. Specifically, the target acceleration can be the vehicle speed before the lane-changing, i.e., returning to the original speed and abandoning the lane change.

[0032] As can be seen from the above, in this embodiment, based on the vehicle's driving data and the driving data of surrounding vehicles, the first vehicle spacing between two adjacent vehicles in the vehicle's lane and the second vehicle spacing between two adjacent vehicles in the target lane are determined; the target lane is the lane to be changed; the driving prediction data for surrounding vehicles and the acceleration for the vehicle are obtained from the novel gating model; the novel gating model takes the driving data of surrounding vehicles as input; based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration, the vehicle's lane-changing trajectory and target acceleration are determined, so that the vehicle can change speed and lane according to the lane-changing trajectory based on the target acceleration. As can be seen, by comprehensively analyzing the lane-changing trajectory and target acceleration of the vehicle based on the first vehicle distance between two adjacent vehicles in the self-lane, the second vehicle distance between two adjacent vehicles in the target lane, and the driving prediction data and acceleration of surrounding vehicles output by the new gating model, the vehicle can be controlled to accelerate or decelerate to complete the lane-changing process. The lane-changing process does not affect the driving of other vehicles. Furthermore, the acceleration and deceleration determined by the new gating model make the entire lane-changing process smoother and improve the comfort of the lane-changing process. The lane-changing trajectory is determined by predicting the trajectory of surrounding vehicles based on the new gating model, which greatly improves the success rate of lane-changing.

[0033] Based on the above embodiments, this application also discloses a specific speed-changing and lane-changing system, for example... Figure 4 As shown, the container module receives data from the intelligent forward-facing camera, front and rear corner radars, and other vehicle-related data. The safetycheck module outputs whether there is a collision risk in the surrounding area, and the drivetask module outputs the road curvature. The gear shifting and lane changing module is the core of this application, specifically including modules A to J, and ultimately outputs the target acceleration. The operation process is as follows: When a vehicle prepares to change lanes, the safetycheck module checks if a direct lane change is possible. If the vehicle is changing lanes to the left and there are no vehicles on the left, there is no need to activate the lane change mechanism. If there are vehicles, it also checks if the vehicle is on a road section with a large curve (e.g., curve radius r < 150). If both conditions are met, the lane change module is activated. When a lane change is required, module A in the lane change module receives the results from the safetycheck and the container. When the intelligent forward-looking camera identifies the lane ahead as a solid line, the lane change function is disabled. Module B classifies the vehicle's lane change behavior, determining whether the vehicle is changing lanes to the left or right, and outputs the result to module C. Module C receives the lane change command and simultaneously packages and sends the vehicle's own driving data, the driving data of vehicles in its own lane, and the driving data of vehicles in the target lane to downstream modules. When the lane change function is first activated during this lane change process, the default parameter of module D is false, and it is set to true when a vehicle control flag signal is received.

[0034] Module E first filters out scenarios where lane changes are not possible, such as when other vehicles surround the vehicle and there is no effective lane-changing area, and outputs a flag to the downstream module. Module F calculates the lane-changing area based on the first and second vehicle spacings, and outputs the initial lane-changing trajectory and initial acceleration. Module G receives the information from Module F and, combined with the driving prediction data and acceleration from the novel gating model, determines the final lane-changing trajectory and target acceleration, and outputs it to Module J. Module J sends a vehicle control flag and outputs the target acceleration to the vehicle control module. It is evident that the acceleration output by Module G based on the results of Module F's operation is the maximum acceleration under vehicle kinematic constraints. This method aims to improve the success rate of lane changes, but while highly efficient, it sacrifices comfort. Therefore, the NGCU model is introduced to learn the acceptable speed range for the vehicle. When a lane change occurs, the ratio of lane-changing efficiency to comfort can be adjusted by regulating the weights.

[0035] During subsequent speed change and lane change, module D receives the vehicle control signal and, in conjunction with the trajectory of surrounding vehicles output by NGCU Model II, continuously sends the results to module H. Module H performs acceleration correction and sends the results to module J, and finally outputs the vehicle control signal.

[0036] Furthermore, during system operation, the output of module J causes longitudinal displacement of the vehicle, and its output is also received by the safetycheck module. When the safetycheck module detects a collision, it can directly output updated acceleration to module J to ensure driving safety. Simultaneously, the output of module J is also transmitted to NGCU model II. NGCU model II, as a continuously trained model, outputs acceleration. Similarly, information on other vehicle behaviors, such as lane changes and sudden acceleration / deceleration, can be directly output to module H. Upon receiving this information, module H can terminate the lane change to prevent an accident.

[0037] As can be seen, by comprehensively analyzing the lane-changing trajectory and target acceleration of the vehicle based on the first vehicle distance between two adjacent vehicles in the self-lane, the second vehicle distance between two adjacent vehicles in the target lane, and the driving prediction data and acceleration of surrounding vehicles output by the new gating model, the vehicle can be controlled to accelerate or decelerate to complete the lane-changing process. The lane-changing process does not affect the driving of other vehicles. Furthermore, the acceleration and deceleration determined by the new gating model make the entire lane-changing process smoother and improve the comfort of the lane-changing process. The lane-changing trajectory is determined by predicting the trajectory of surrounding vehicles based on the new gating model, which greatly improves the success rate of lane-changing.

[0038] Accordingly, this application also discloses a speed-changing and lane-changing device for use in autonomous vehicles, see [link to relevant documentation]. Figure 5 As shown, the device includes: The vehicle spacing determination module 11 is used to determine the first vehicle spacing between two adjacent vehicles in the vehicle lane and the second vehicle spacing between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles; the target lane is the lane to be changed. The driving prediction data acquisition module 12 is used to acquire driving prediction data for surrounding vehicles and acceleration for the vehicle itself, output by the novel gating model; the novel gating model takes the driving data of surrounding vehicles as input. The speed change and lane change parameter determination module 13 is used to determine the lane change trajectory and target acceleration of the vehicle based on the first vehicle spacing, the second vehicle spacing, the driving prediction data and the acceleration, so as to perform speed change and lane change according to the target acceleration and the lane change trajectory.

[0039] As can be seen from the above, in this embodiment, based on the vehicle's driving data and the driving data of surrounding vehicles, the first vehicle spacing between two adjacent vehicles in the vehicle's lane and the second vehicle spacing between two adjacent vehicles in the target lane are determined; the target lane is the lane to be changed; the driving prediction data for surrounding vehicles and the acceleration for the vehicle are obtained from the novel gating model; the novel gating model takes the driving data of surrounding vehicles as input; based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration, the vehicle's lane-changing trajectory and target acceleration are determined, so that the vehicle can change speed and lane according to the lane-changing trajectory based on the target acceleration. As can be seen, by comprehensively analyzing the lane-changing trajectory and target acceleration of the vehicle based on the first vehicle distance between two adjacent vehicles in the self-lane, the second vehicle distance between two adjacent vehicles in the target lane, and the driving prediction data and acceleration of surrounding vehicles output by the new gating model, the vehicle can be controlled to accelerate or decelerate to complete the lane-changing process. The lane-changing process does not affect the driving of other vehicles. Furthermore, the acceleration and deceleration determined by the new gating model make the entire lane-changing process smoother and improve the comfort of the lane-changing process. The lane-changing trajectory is determined by predicting the trajectory of surrounding vehicles based on the new gating model, which greatly improves the success rate of lane-changing.

[0040] In some specific embodiments, the speed change and lane changing device may specifically include: The activation condition judgment unit is used to determine whether the speed change and lane change activation conditions are met before determining the first vehicle distance between two adjacent vehicles in the self-lane and the second vehicle distance between two adjacent vehicles in the target lane based on the self-vehicle's driving data and the driving data of surrounding vehicles. The speed change and lane change activation conditions include the presence of surrounding vehicles, the road curve radius being less than a preset radius threshold, and the lane being changeable. An execution unit is configured to, if the conditions for starting a lane change are met, execute the steps of determining the first vehicle distance between two adjacent vehicles in the vehicle lane and the second vehicle distance between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles.

[0041] In some specific embodiments, the vehicle spacing determination module 11 may specifically include: The lane change classification unit is used to determine the lane change type based on lane change behavior data and generate a lane change instruction based on the lane change type; the lane change type includes left lane change and right lane change; The target lane determination unit is used to determine the target lane according to the lane change instruction and extract the target lane vehicle driving data from the driving data of surrounding vehicles. The first vehicle spacing determination unit is used to determine the first vehicle spacing between two adjacent vehicles in the lane based on the vehicle driving data of the lane. The second vehicle spacing determination unit is used to determine the second vehicle spacing between two adjacent vehicles in the target lane based on the vehicle driving data of the target lane.

[0042] In some specific embodiments, the driving prediction data acquisition module 12 may specifically include: The driving prediction data acquisition unit is used to acquire driving prediction data for surrounding vehicles output by the first novel gating model; the first novel gating model takes the driving data of the surrounding vehicles in the previous N cycles as input and the predicted driving data of the surrounding vehicles in the next T cycles as output. An acceleration acquisition unit is used to acquire the acceleration of the vehicle as output by the second novel gating model; the second novel gating model takes the driving data of the surrounding vehicles for the previous N cycles and the target acceleration as input, and takes the acceleration of the vehicle as output.

[0043] In some specific embodiments, the speed change and lane change parameter determination module 13 may specifically include: The verification unit is used to select the second vehicle spacing in order from near to far based on the distance to the vehicle itself for lane change verification, and obtain the second target vehicle spacing that passes the verification. The initial lane change parameter determination unit is used to select the first target vehicle distance corresponding to the second target vehicle distance, determine the initial lane change trajectory of the vehicle based on the first target vehicle distance and the second target vehicle distance, and determine the initial acceleration; The lane change trajectory determination unit is used to determine whether there is a collision risk in the initial lane change trajectory based on the driving prediction data, and to determine the lane change trajectory based on the determination result; A target acceleration determination unit is used to adjust the initial acceleration according to the acceleration to obtain the target acceleration.

[0044] In some specific embodiments, the target acceleration determination unit may specifically include: The weight determination unit is used to determine the weight corresponding to the acceleration and the weight corresponding to the initial acceleration based on the focus of the lane change behavior. The target acceleration calculation unit is used to obtain the target acceleration by calculating a weighted sum of the initial acceleration and the target acceleration.

[0045] In some specific embodiments, the speed-changing and lane-changing device may further include: The vehicle control flag update unit is used to update the vehicle control flag after determining the lane change trajectory and target acceleration of the vehicle; the vehicle control flag indicates whether the autonomous vehicle is in the speed change and lane change control phase. The target acceleration adjustment unit is used to determine in real time whether there is a collision risk during the vehicle's shifting and lane changing process, based on the driving prediction data of surrounding vehicles output by the novel gating model, if the vehicle is in the shifting and lane changing control phase. If there is a collision risk, the target acceleration is adjusted.

[0046] Furthermore, this application also discloses an electronic device, see [link to relevant documentation]. Figure 6 As shown, the content in the figure should not be considered as any limitation on the scope of use of this application.

[0047] Figure 6 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the speed change and lane-changing method disclosed in any of the foregoing embodiments.

[0048] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0049] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon include operating system 221, computer program 222 and data 223 including driving data, etc. The storage method can be temporary storage or permanent storage.

[0050] The operating system 221 manages and controls the various hardware devices on the electronic device 20 and the computer program 222 to enable the processor 21 to perform calculations and processing on the massive amounts of data 223 in the memory 22. The operating system 221 can be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the speed-changing and lane-changing method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0051] Furthermore, this application also discloses a computer storage medium storing computer-executable instructions. When the computer-executable instructions are loaded and executed by a processor, they implement the speed change and lane change method steps disclosed in any of the foregoing embodiments.

[0052] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0053] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0054] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, 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 a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0055] The above provides a detailed description of the speed-changing and lane-changing method, apparatus, equipment, and medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for changing speed and lanes, characterized in that, Applied to autonomous vehicles, including: Based on the vehicle's own driving data and the driving data of surrounding vehicles, determine the first vehicle distance between two adjacent vehicles in the vehicle's own lane and the second vehicle distance between two adjacent vehicles in the target lane; the target lane is the lane where the lane change is to be made. The novel gating model outputs driving prediction data for surrounding vehicles and acceleration data for the vehicle itself; the novel gating model takes driving data of surrounding vehicles as input. Based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration, the lane-changing trajectory and target acceleration of the vehicle are determined so that the vehicle can change speed and lanes according to the lane-changing trajectory based on the target acceleration.

2. The speed change and lane changing method according to claim 1, characterized in that, Before determining the first vehicle distance between two adjacent vehicles in the self-lane and the second vehicle distance between two adjacent vehicles in the target lane based on the self-vehicle's driving data and the driving data of surrounding vehicles, the method further includes: Determine whether the conditions for starting a lane change with increased speed are met; the conditions for starting a lane change with increased speed include the presence of vehicles in the vicinity, the radius of the road curve being less than a preset radius threshold, and the lane being changeable. If the conditions are met, then the steps of determining the first vehicle distance between two adjacent vehicles in the vehicle lane and the second vehicle distance between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles are executed.

3. The speed change and lane changing method according to claim 1, characterized in that, The step of determining the first vehicle distance between two adjacent vehicles in the vehicle's own lane and the second vehicle distance between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles includes: The lane change type is determined based on lane change behavior data, and a lane change instruction is generated based on the lane change type; the lane change type includes left lane change and right lane change; The target lane is determined based on the lane change instruction, and the driving data of vehicles in the target lane is extracted from the driving data of surrounding vehicles. Based on the vehicle driving data in the self-lane, determine the first vehicle distance between two adjacent vehicles in the self-lane; Based on the vehicle driving data of the target lane, determine the second vehicle spacing between two adjacent vehicles in the target lane.

4. The speed change and lane changing method according to claim 1, characterized in that, The acquisition of driving prediction data for surrounding vehicles and acceleration for the vehicle itself, output by the novel gating model, includes: Obtain driving prediction data for surrounding vehicles output by the first novel gating model; the first novel gating model takes the driving data of the surrounding vehicles in the previous N cycles as input and the predicted driving data of the surrounding vehicles in the next T cycles as output. Obtain the acceleration of the vehicle as output by the second novel gating model; the second novel gating model takes the driving data of the surrounding vehicles for the previous N cycles and the target acceleration as input, and takes the acceleration of the vehicle as output.

5. The speed change and lane changing method according to claim 1, characterized in that, The step of determining the lane-changing trajectory and target acceleration of the vehicle based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration includes: Based on the distance to the vehicle, the second vehicle spacing is selected in order from closest to farthest for lane change verification, and the second target vehicle spacing that passes the verification is obtained. Select a first target vehicle distance corresponding to the second target vehicle distance, determine the initial lane-changing trajectory of the vehicle based on the first target vehicle distance and the second target vehicle distance, and determine the initial acceleration; Based on the driving prediction data, determine whether there is a collision risk in the initial lane change trajectory, and determine the lane change trajectory based on the judgment result; The initial acceleration is adjusted according to the acceleration to obtain the target acceleration.

6. The speed change and lane changing method according to claim 5, characterized in that, The step of adjusting the initial acceleration according to the acceleration to obtain the target acceleration includes: The weights corresponding to the acceleration and the initial acceleration are determined based on the emphasis of the lane change behavior. The target acceleration is obtained by calculating the weighted sum of the acceleration and the initial acceleration.

7. The speed change and lane changing method according to any one of claims 1 to 6, characterized in that, After determining the vehicle's lane-changing trajectory and target acceleration, the process also includes: Update the vehicle control flag; the vehicle control flag indicates whether the autonomous vehicle is in the gear change and lane change control phase. If the vehicle is in the speed change and lane change control phase, the driving prediction data of surrounding vehicles output by the new gating model is used to determine in real time whether there is a collision risk during the speed change and lane change process. If there is, the target acceleration is adjusted.

8. A speed-changing and lane-changing device, characterized in that, Applied to autonomous vehicles, including: The vehicle spacing determination module is used to determine the first vehicle spacing between two adjacent vehicles in the vehicle's own lane and the second vehicle spacing between two adjacent vehicles in the target lane based on the vehicle's own driving data and the driving data of surrounding vehicles; the target lane is the lane to be changed. The driving prediction data acquisition module is used to acquire driving prediction data for surrounding vehicles and acceleration for the vehicle itself, output by the novel gating model; the novel gating model takes the driving data of surrounding vehicles as input. The speed change and lane change parameter determination module is used to determine the lane change trajectory and target acceleration of the vehicle based on the first vehicle spacing, the second vehicle spacing, the driving prediction data, and the acceleration, so as to perform speed change and lane change according to the target acceleration and the lane change trajectory.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the speed change and lane change method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store computer programs; wherein the computer programs, when executed by a processor, implement the speed change and lane change method as described in any one of claims 1 to 7.

Images