Vehicle autonomous lane-changing decision planning method and related device

Abstract

The embodiment of the application provides a vehicle autonomous lane-changing decision planning method and related equipment, and belongs to the technical field of vehicle automatic driving. In the lane-changing decision planning method, after lane-changing decision is made according to the current traffic state in the lane-changing decision stage, in the case that the lane-changing intention is to change lanes to a target lane, a lane-changing trajectory interval is obtained by performing preliminary trajectory planning according to the current traffic state and the target lane, then safety rule verification is performed on the lane-changing trajectory interval, in the case that the lane-changing trajectory interval does not satisfy the safety rule, the vehicle continues to keep in the current lane, and in the case that the lane-changing trajectory interval satisfies the safety rule, the lane-changing decision intention to change lanes to the target lane is executed. The scheme can improve the safety of the autonomous lane-changing behavior of the vehicle.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, and in particular to a method and related equipment for autonomous lane-changing decision-making and planning of vehicles. Background Technology

[0002] Changing lanes in a timely manner can effectively improve traffic efficiency, and it is also an important way to avoid collisions and improve driving safety. However, changing lanes is a complex driving maneuver that requires a high level of decision-making and judgment from the driver. Improper lane changes can cause vehicle instability or collisions, resulting in traffic accidents. Related studies have shown that some traffic accidents are related to lane changes, with driver errors in lane-changing decisions being one of the main causes. With the development of intelligent vehicles, autonomous lane changing is gradually becoming a core function of autonomous driving. Lane-changing decisions and corresponding trajectory planning have a direct and significant impact on traffic flow, driving safety, and passenger comfort.

[0003] Please refer to Figure 2 The diagram illustrates the stages of vehicle lane-changing behavior, which can be divided into two phases: the decision-making phase and the execution phase. The decision-making phase refers to the driver observing traffic conditions and determining whether a lane change is necessary before proceeding. The execution phase refers to the process of the driver moving the vehicle from the original lane to the target lane after deciding to change lanes. For autonomous lane changing in intelligent vehicles, the decision-making phase involves issuing a lane-changing instruction based on current environmental information, while the execution phase involves lane-changing trajectory planning and trajectory tracking control.

[0004] Typically, lane-changing trajectory planning is performed after a lane-changing decision is made. In related technologies, the lane-changing decision and trajectory planning processes are often considered to be independent parts. This means that if an erroneous lane-changing decision that could lead to a collision is made during the lane-changing decision-making stage, it is difficult to avoid a collision during the lane-changing execution stage. Therefore, the safety of autonomous lane-changing behavior of vehicles is still relatively low at present. Summary of the Invention

[0005] The main objective of this application is to propose a vehicle autonomous lane-changing decision-making and planning method and related equipment, which aims to improve the safety of vehicle autonomous lane-changing behavior.

[0006] To achieve the above objectives, one aspect of this application proposes a vehicle autonomous lane-changing decision-making and planning method, comprising the following steps: The lane-change decision model makes a lane-change decision based on the current traffic conditions to obtain the lane-change intention; the lane-change intention is used to determine whether to change lanes to the target lane. When the lane-changing intention is to change lanes to the target lane, preliminary trajectory planning is performed based on the current traffic conditions and the target lane to obtain the lane-changing trajectory range; The safety rules of the lane-changing trajectory section are verified, and the verification results are obtained. If the verification result indicates that the lane-changing trajectory section does not meet the safety rules, then the lane will remain in the current lane, and the safety rules of the lane-changing trajectory section will be re-verified based on the real-time updated current traffic conditions. If the verification result indicates that the lane-changing trajectory section meets the safety rules, then the decision to change lanes to the target lane is executed.

[0007] In some embodiments, the lane-changing decision model is trained through the following steps: Based on the correlation between traffic information around the main vehicle and sample labels, the quality of each sample data is assessed and high-quality samples are selected; the sample labels are for performing lane changing or lane keeping. A training dataset is formed based on multiple selected high-quality samples; The long short-term memory neural network was trained using the training dataset to obtain the lane-changing decision model.

[0008] In some embodiments, the process of evaluating the quality of each sample data based on the correlation between traffic information around the main vehicle and sample labels, and selecting high-quality samples, includes the following steps: Based on the sample labels, the sample data is divided into either the first sample corresponding to lane changing or the second sample corresponding to lane keeping; The first quality score of the first sample is determined based on the relative state of the main vehicle and surrounding vehicles at the starting point of the lane change, and high-quality samples are selected from multiple first samples based on the first quality score. The second quality score of the second sample is determined based on the average relative state between the main vehicle and surrounding vehicles during the lane change decision period, and high-quality samples are selected from multiple second samples based on the second quality score.

[0009] In some embodiments, the preliminary trajectory planning based on the current traffic conditions and the target lane to obtain the lane-changing trajectory interval includes the following steps: Initialize the destination state of the main vehicle after lane changing based on the current traffic conditions and the target lane; Trajectory planning is performed based on the endpoint state and the starting state of the main vehicle to obtain a trajectory equation that takes safety factors into account. The trajectory equation is used to represent the trajectory sub-interval corresponding to the endpoint state, and multiple trajectory sub-intervals of the endpoint states share a common representation of the lane-changing trajectory interval.

[0010] In some embodiments, the safety rule verification of the lane-changing trajectory section to obtain the verification result includes the following steps: Perform safety rule verification on the trajectory sub-intervals of the current endpoint state; If the trajectory sub-interval of the current destination state does not meet the safety rules, the destination state is updated according to the current traffic state and the target lane. The steps of verifying the safety rules of the trajectory sub-interval of the current destination state and performing trajectory planning according to the destination state and the starting state of the main vehicle to obtain the trajectory equation considering the safety factor are repeated until the trajectory sub-interval of the current destination state meets the safety rules. Then, the verification result is determined to be that the lane-changing trajectory interval meets the safety rules. If none of the trajectory sub-intervals in all endpoint states meet the safety rules, then the verification result is determined to be that the lane-changing trajectory interval does not meet the safety rules.

[0011] In some embodiments, the safety rules include a first constraint rule regarding the displacement of the main lane at a first critical moment when the main vehicle does not collide in its own lane, a second constraint rule regarding the speed and displacement of the target lane at a second critical moment when the main vehicle collides in the target lane, and a third constraint rule regarding the speed and displacement of the target lane at the moment the lane change is completed. The decision intent to change lanes to the target lane includes the following steps: If the current time is before the first critical time, then the first constraint rule, the second constraint rule and the third constraint rule are used to replan the trajectory of the lane-changing trajectory interval updated based on the current traffic state to obtain a dynamic tracking trajectory. If the current time is after the first critical time, the second constraint rule and the third constraint rule are used to replan the trajectory of the lane-changing trajectory interval updated based on the current traffic state to obtain a dynamic tracking trajectory.

[0012] To achieve the above objectives, another aspect of this application proposes a vehicle autonomous lane-changing decision-making and planning system, comprising: The first module is used to make lane-changing decisions based on the current traffic conditions through a lane-changing decision model to obtain lane-changing intentions; the lane-changing intentions are used to determine whether to change lanes to the target lane. The second module is used to perform preliminary trajectory planning based on the current traffic conditions and the target lane when the lane-changing intention is to change lanes to the target lane, and to obtain the lane-changing trajectory range. The third module is used to perform safety rule verification on the lane-changing trajectory section and obtain the verification result. The fourth module is used to continue to stay in the current lane and re-verify the safety rules of the lane-changing trajectory section according to the real-time updated current traffic status if the verification result shows that the lane-changing trajectory section does not meet the safety rules. The fifth module is used to execute the decision to change lanes to the target lane if the verification result shows that the lane-changing trajectory section meets the safety rules.

[0013] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0014] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0015] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0016] The embodiments of this application include at least the following beneficial effects: This application provides a method, system, electronic device, storage medium, and program product for autonomous lane-changing decision-making and planning. In the lane-changing decision-making stage, after making a lane-changing decision based on the current traffic conditions, if the lane-changing intention is to change lanes to the target lane, preliminary trajectory planning is performed based on the current traffic conditions and the target lane to obtain a lane-changing trajectory interval. Then, the lane-changing trajectory interval is verified against safety rules. If the lane-changing trajectory interval does not meet the safety rules, the vehicle continues to remain in the current lane; otherwise, if the lane-changing trajectory interval meets the safety rules, the decision to change lanes to the target lane is executed. This solution coordinates the lane-changing decision-making and execution stages through safety rules, using safety rules to verify the safety of the planned trajectory interval under the lane-changing intention. If it is unsafe, the vehicle does not enter the lane-changing execution stage, thus improving the safety of autonomous lane-changing behavior. Attached Figure Description

[0017] Figure 1 This is a flowchart of the vehicle autonomous lane-changing decision-making and planning method provided in the embodiments of this application; Figure 2 This is a schematic diagram of the vehicle lane-changing behavior stages provided in the embodiments of this application; Figure 3 This is a schematic diagram of a critical collision scenario provided in an embodiment of this application; Figure 4 This is a schematic diagram of the overall concept of the vehicle autonomous lane-changing decision-making and planning method provided in the embodiments of this application; Figure 5 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0020] Before providing a detailed description of the embodiments of this application, some of the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this application are subject to the following interpretations.

[0021] Long Short-Term Memory (LSTM) is a type of recurrent neural network designed to address the long-term dependency problem inherent in general RNNs (Recurrent Neural Networks). All RNNs have a chain-like structure of repeating neural network modules.

[0022] For related technologies, please refer to Figure 2 The diagram illustrates the stages of vehicle lane-changing behavior, which can be divided into two phases: the decision-making phase and the execution phase. The decision-making phase involves the driver observing traffic conditions and determining whether a lane change is necessary before proceeding. The execution phase refers to the process of the driver moving the vehicle from the original lane to the target lane after deciding to change lanes. For autonomous lane changing in intelligent vehicles, the decision-making phase involves issuing a command based on current environmental information, while the execution phase involves lane-changing trajectory planning and trajectory tracking control. Typically, lane-changing trajectory planning occurs after the lane-changing decision is made. In related technologies, the lane-changing decision-making and trajectory planning processes are often considered independent. This leads to a situation where if a collision-prone erroneous lane-changing decision is made during the decision-making phase, it is difficult to prevent a collision during the execution phase. Therefore, the safety of autonomous lane-changing behavior in vehicles remains relatively low.

[0023] In view of this, this application provides a vehicle autonomous lane-changing decision-making and planning method and related equipment. This solution associates and coordinates the lane-changing decision-making stage and the execution stage through safety rules. It uses safety rules to verify the safety of the planned trajectory range under the lane-changing intention. If it is not safe, it will not enter the lane-changing execution stage, thereby improving the safety of the vehicle's autonomous lane-changing behavior.

[0024] The vehicle autonomous lane-changing decision-making and planning method provided in this application relates to the field of vehicle autonomous driving technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or in-vehicle terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application that implements the vehicle autonomous lane-changing decision-making and planning method, but is not limited to the above forms.

[0025] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0026] Figure 1 This is an optional flowchart of the vehicle autonomous lane-changing decision-making and planning method provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S101 to S105.

[0027] Step S101: Based on the current traffic conditions, a lane-changing decision is made using the lane-changing decision model to obtain the lane-changing intention; the lane-changing intention is used to determine whether to change lanes to the target lane. Step S102: When the lane-changing intention is to change lanes to the target lane, preliminary trajectory planning is performed based on the current traffic conditions and the target lane to obtain the lane-changing trajectory range. Step S103: Perform safety rule verification on the lane-changing trajectory section and obtain the verification result; Step S104: If the verification result shows that the lane-changing trajectory section does not meet the safety rules, then continue to stay in the current lane and re-verify the safety rules of the lane-changing trajectory section according to the real-time updated current traffic conditions. Step S105: If the verification result shows that the lane-changing trajectory section meets the safety rules, then execute the decision to change lanes to the target lane.

[0028] Steps S101 to S105 as shown in the embodiments of this application link the lane-changing decision-making stage and the execution stage through safety rules. The safety of the planned trajectory range under the lane-changing intention is verified by using safety rules. If it is not safe, the lane-changing execution stage is not entered, and the decision-making model of the decision-making stage is waited for to make a new decision, thereby improving the safety of the vehicle's autonomous lane-changing behavior.

[0029] In step S101 of some embodiments, the lane-changing decision model is used to classify the current traffic state-related data to obtain the lane-changing intention to change lanes to the target lane. The lane-changing decision model can be a machine learning-based algorithm model, such as a support vector machine or decision tree, or a deep learning-based neural network model, such as an LSTM model. In this embodiment, the current traffic state input to the lane-changing decision model may include the speed of the main vehicle during the lane-changing decision time (e.g., 2 seconds), the relative speed and distance with adjacent vehicles (the vehicle in front in this lane, the vehicle in front in the target lane, and the vehicle behind).

[0030] In step S102 of some embodiments, upon deciding to change lanes to the target lane, preliminary trajectory planning is performed based on the current traffic conditions and the target lane to obtain a lane-changing trajectory interval. The lane-changing trajectory interval represents the drivable section after completing the lane-changing process. It indicates the possible endpoint state, i.e., the relative state of the vehicle in the target lane after the lane change, such as longitudinal position, longitudinal speed, longitudinal acceleration, lateral position, lateral speed, and lateral acceleration. The lane-changing trajectory interval also indicates the possible driving trajectory between the starting state and the endpoint state. Specifically, the lane-changing trajectory interval can be represented by trajectory equations corresponding to multiple possible endpoint states. These trajectory equations represent the possible driving trajectories given the starting and endpoint states. The possible driving trajectories are determined by the safety factor in the trajectory equations; that is, determining the safety factor in the trajectory equations determines the driving trajectory from the starting state to the endpoint state. The trajectory equations can be constructed using trajectory planning algorithms such as polynomial fitting or model predictive control.

[0031] In step S103 of some embodiments, a safety rule verification is performed on the lane-changing trajectory interval to obtain a verification result. Safety rule verification refers to determining whether a trajectory within the lane-changing trajectory interval conforms to defined safety rules. If such a trajectory exists, the verification result is that the lane-changing trajectory interval meets the safety rules; otherwise, the verification result is that the lane-changing trajectory interval does not meet the safety rules. The safety rules in this embodiment may include a first constraint rule regarding the displacement of the main lane at a first critical moment when the main vehicle does not collide in its current lane, a second constraint rule regarding the speed and displacement of the target lane at a second critical moment when the main vehicle collides in the target lane, and a third constraint rule regarding the speed and displacement of the target lane at the moment the lane change is completed.

[0032] In step S104 of some embodiments, if the verification result indicates that the lane-changing trajectory section does not meet the safety rules, it means that the lane-changing decision made in the decision-making stage is likely to cause a collision in the execution stage. Therefore, the vehicle does not enter the lane-changing execution stage, but continues to drive in the current lane, waiting for the lane-changing decision model to re-determine the next step. If the verification is successful, the vehicle can enter the lane-changing execution stage.

[0033] In step S105 of some embodiments, if the verification result shows that the lane-changing trajectory interval meets the safety rules, it indicates that the lane-changing decision in the decision-making stage is highly safe. At this point, the lane-changing execution stage begins. When executing the lane change, the lane-changing trajectory interval is first optimized with comfort and communication efficiency as the optimization objectives to obtain the safest target lane-changing trajectory. During the tracking of the target lane-changing trajectory, trajectory replanning can be performed based on the real-time updated current traffic conditions and the aforementioned safety rules to improve the safety of the lane-changing process.

[0034] According to some embodiments of this application, the lane-changing decision model in step S101 is trained through the following steps: Step S201: Based on the correlation between traffic information around the main vehicle and sample labels, the quality of each sample data is evaluated and high-quality samples are selected; the sample label is to perform lane change or lane keeping. Step S202: A training dataset is formed based on the selected high-quality samples; Step S203: Train the long short-term memory neural network based on the training dataset to obtain the lane-changing decision model.

[0035] In step S201 of some embodiments, lane-changing decisions in related technologies mainly include rule-based methods, game-theory-based methods, and data-driven methods. Rule-based methods determine lane-changing feasibility by setting a fixed threshold, exhibiting stable performance and being easy to interpret and implement, but they are only applicable to specific scenarios and lack generalization performance. Game-theory-based methods accurately model multi-vehicle interactions in lane-changing scenarios to find the optimal decision. Compared to rule-based methods, game-theory has stronger generalization capabilities and can be applied to various lane-changing scenarios, but solving game theory is relatively complex and may affect real-time performance when involving the interaction of many vehicles. Data-driven decision models, especially deep learning decision models, demonstrate high prediction accuracy by learning decision strategies through simulation. These models are trained on large amounts of data, can adapt to complex traffic environments, and have strong generalization capabilities. However, data-driven methods are greatly affected by the size and quality of the dataset, making the lane-changing decision model unreliable as the decision-making intention.

[0036] Based on this, in training the lane-changing decision model, this implementation first selects higher-quality samples from the sample data to train the model, thereby improving the robustness of the lane-changing decision model. Specifically, when the surrounding traffic information of the main vehicle shows a strong positive correlation with the lane-changing execution and a strong negative correlation with the lane-keeping execution, it is defined as high-quality data. If the following conditions are met, the main vehicle is more suitable for lane changing, can obtain higher benefits, and ensures safety: (1) The distance between the main vehicle and the vehicle in front in this lane is small and much smaller than the distance between the main vehicle and the vehicle in the target lane; (2) The distance between the main vehicle and the vehicle in front or behind in the target lane is large, or there is no vehicle in front or behind in the target lane; (3) The vehicle in front of this lane has a lower speed, the vehicle in front of the target lane has a higher speed, and the vehicle behind has a lower speed.

[0037] To obtain a high-quality dataset, it is necessary to increase the proportion of high-quality data in the dataset. Based on the above conditions, this implementation evaluates the quality of sample data by constructing decision evaluation functions with different sample labels.

[0038] Specifically, step S201 may include, but is not limited to, the following steps: Step S301: Divide the sample data into the first sample corresponding to lane change or the second sample corresponding to lane keeping based on the sample label; Step S302: Determine the first quality score of the first sample based on the relative state of the main vehicle and surrounding vehicles at the starting time of lane change, and select high-quality samples from multiple first samples based on the first quality score. Step S303: Determine the second quality score of the second sample based on the average relative state between the main vehicle and surrounding vehicles during the lane change decision time period, and select high-quality samples from multiple second samples based on the second quality score.

[0039] In this embodiment, the sample data is first divided into a first sample corresponding to lane changing or a second sample corresponding to lane keeping based on the sample label.

[0040] For the quality assessment of the first sample, its first quality score is calculated using the lane-change decision data evaluation function, which is as follows: ; in, This is the starting time for lane changing; , , These represent the distances between the main vehicle and the vehicle in front in its current lane, the vehicle behind in the target lane, and the vehicle in front of the target lane at the start of the lane change. The relative speed between the vehicle in the target lane and the vehicle in front of the vehicle in the current lane at the start of the lane change; The relative speed between the vehicle behind and the vehicle in the target lane at the moment of lane change; The denominators of each term represent the maximum value of the corresponding numerator in the lane change dataset.

[0041] First quality score corresponding to the sample data The higher the score, the higher the quality of the lane-changing decision sample. Specifically, the top few first samples with the highest first quality score can be identified as high-quality samples, or the top few first samples with the first quality score greater than the first score threshold can be identified as high-quality samples.

[0042] For the quality assessment of the second sample, its second quality score is calculated using the lane-keeping decision data evaluation function, which is as follows: ; in, This represents the average distance between the main vehicle and the vehicle in front in the same lane during the lane-changing decision-making period; similarly, , , , This represents the average value of the corresponding feature parameters; The denominators of each term represent the maximum value of the corresponding numerator in the lane-keeping dataset.

[0043] Second quality score corresponding to sample data The higher the score, the higher the quality of the lane-keeping sample. Specifically, the top few second samples with the highest second quality scores can be identified as high-quality samples, or the few second samples with second quality scores greater than the second score threshold can be identified as high-quality samples.

[0044] This embodiment filters samples by evaluating the magnitude of the function value, thus obtaining a high-quality dataset (i.e., a training dataset) in which the input and output show a strong correlation.

[0045] In steps S202 to S203 of some embodiments, based on the high-quality dataset extracted above, an LSTM neural network lane-changing decision model with seven feature parameters can be established. During training, in order to obtain the optimal decision model, different numbers of intermediate layer nodes are selected for training, and the average prediction accuracy of each structure model is calculated. In this embodiment, considering both model structure complexity and performance, a lane-changing decision model with 8 intermediate layer nodes is selected.

[0046] According to some embodiments of this application, step S102 may include, but is not limited to, the following steps: Step S401: Initialize the destination state of the main vehicle after lane changing based on the current traffic conditions and the target lane; Step S402: Based on the endpoint state and the starting state of the main vehicle, trajectory planning is performed to obtain a trajectory equation that takes safety factors into account. The trajectory equation is used to represent the trajectory sub-interval corresponding to the endpoint state. Multiple trajectory sub-intervals of the endpoint states share the same representation to represent the lane-changing trajectory interval.

[0047] In this embodiment, based on the current traffic conditions (i.e., the positions of vehicles in front and behind in the target lane) and the initial destination state of the main vehicle after lane changing in the target lane, lane-changing trajectory planning can be performed when the current starting and ending states are determined. This embodiment uses a polynomial fitting method for lane-changing trajectory planning. Specifically, lane changing is the process of interaction between the main vehicle and the vehicle in front in its current lane, and the vehicles in front and behind in the target lane. Therefore, the lane-changing trajectory not only requires smoothness and continuous curvature but also must meet certain safety requirements to avoid collisions with surrounding vehicles. For a fifth-order polynomial, once the vehicle's lane-changing boundary conditions (i.e., the positions of the starting and ending points, vehicle speed, and acceleration) are determined, the trajectory equation is uniquely determined, meaning only one trajectory can satisfy the lane-changing boundary conditions. This may lead to the main vehicle being unable to maintain a safe distance from surrounding vehicles, causing safety hazards. If the polynomial order is greater than 5, there are multiple trajectories that satisfy the boundary conditions. By analyzing the relative positions and speeds of the main vehicle and surrounding vehicles, the trajectory with the highest safety can be selected. This satisfies the boundary conditions and improves safety during the lane-changing process compared to a fifth-order polynomial. Considering that the collision risk during lane changing is mainly concentrated in longitudinal rear-end collisions, and to avoid a sharp increase in computational load due to excessively high polynomial order, this embodiment uses the polynomial curve generation method for lane changing trajectory planning, and sets the order of the lateral trajectory equation to 5 and the order of the longitudinal trajectory equation to 6. The trajectory equation in this embodiment is as follows: ; in, and They are respectively The vehicle's longitudinal and lateral positions at all times. , , , , , , and , , , , , The fitting coefficients for the trajectory equation are . This refers to the time required for the lane change process.

[0048] Since the starting state (including the vehicle's longitudinal position, longitudinal velocity, longitudinal acceleration, lateral position, lateral velocity, and lateral acceleration at the starting point) and the ending state (including the vehicle's longitudinal position, longitudinal velocity, longitudinal acceleration, lateral position, lateral velocity, and lateral acceleration at the ending point) are determined, the lateral trajectory equation can be fitted. The six fitting coefficients (i.e., the lateral trajectory equation is determined) and the longitudinal trajectory equation The six fitting coefficients and the unknown variables in the longitudinal trajectory equation are still unknown. Unknown variables This is the safety factor. The current longitudinal trajectory equation has only one unknown variable. The performance of the longitudinal trajectory can be comprehensively evaluated by establishing a multi-objective cost function, and the optimal solution can be obtained. However, before establishing the cost function, a lane-switching safety analysis needs to be performed to construct safety constraints, thereby determining... The range of values ​​for . During lane changing, there is a potential collision risk between the main vehicle and the vehicle in front in the current lane, the vehicle in front in the target lane, and the vehicle behind in the target lane. Therefore, it is necessary to perform collision analysis (i.e., safety rule verification) on each of the three vehicles separately.

[0049] According to some embodiments of this application, step S103 may include, but is not limited to, the following steps: Step S501: Perform safety rule verification on the trajectory sub-intervals of the current endpoint state; Step S502: If the trajectory sub-section of the current destination state does not meet the safety rules, update the destination state according to the current traffic state and the target lane, and repeat steps S402 and S501 until the trajectory sub-section of the current destination state meets the safety rules. Then, determine that the verification result is that the lane-changing trajectory section meets the safety rules. In step S503, if none of the trajectory sub-intervals in all endpoint states meet the safety rules, then the verification result is determined to be that the lane-changing trajectory interval does not meet the safety rules.

[0050] In this embodiment, please refer to Figure 3 The analysis identifies three critical collision scenarios that a primary vehicle may encounter when changing lanes. These scenarios involve the possibility of collisions with vehicles in the primary lane, vehicles in the target lane, and vehicles behind the primary vehicle. ; in, It is the time it takes for the main vehicle to travel one vehicle width laterally from the starting point (that is, the first critical moment when the main vehicle does not collide with the vehicle in its own lane). The moment when the main vehicle is about to collide with the vehicle behind or in front of the target vehicle (i.e., the second critical moment). for The safe distance that the main vehicle should maintain from the vehicle in front in its lane at all times. , They are respectively The safe distance that the main vehicle should maintain from the vehicle behind it in the target lane and the vehicle in front of it in the target lane at all times; , They are respectively The distance between the main vehicle and the vehicle in front in the same lane at the start of the lane change; , They are respectively At any given moment, the vehicle in front in this lane experiences longitudinal displacement of the main vehicle; , They are respectively Real-time distance between the main vehicle and the vehicle behind in the target lane, and distance between the vehicle in front in the target lane; , These are the distance between the main vehicle and the vehicle behind it in the target lane at the start of the lane change, and the distance between the vehicle in front of it in the target lane. , , They are respectively The longitudinal displacement of the main vehicle is measured at all times, including the vehicle behind and in front of the target lane. , They are respectively The longitudinal displacement of the main vehicle is always measured in relation to the vehicle in front in the target lane. for The safe distance that the main vehicle should maintain between the vehicle in the target lane and the vehicle in front at the moment of lane change (i.e., the moment of completion of lane change).

[0051] in, , , The calculation formula is as follows: ; ; ; in, It is a constant; Minimum safe distance; , , They are respectively At any given moment, the longitudinal speed of the vehicle behind, in front of, and in the target lane; , They are respectively The longitudinal speed of the main vehicle and the vehicle in front in the target lane at any given time.

[0052] Based on the above analysis, the safety rules (i.e., the collision avoidance inequalities) constructed in this embodiment are expressed as follows: ; In this set of collision avoidance inequalities, the first term represents the first constraint rule on the displacement of the main lane at the first critical moment when the main vehicle does not collide in its own lane; the second and third terms represent the second constraint rules on the speed and displacement of the target lane at the second critical moment when the main vehicle collides in the target lane; and the fourth term represents the third constraint rule on the speed and displacement of the target lane at the moment when the lane change is completed.

[0053] The trajectory equation is constrained by the aforementioned collision avoidance inequalities (equivalent to constraining the range of the safety factor). By solving the trajectory equation, if there is no solution, it means that the trajectory equation corresponding to the endpoint state does not satisfy the safety rules, i.e., the initial speed determined based on the vehicle's initial speed, target speed, and lane change time. (i.e., the endpoint state) does not meet the security requirements, so update using the following formula: ; The trajectory equation corresponding to the updated endpoint state is then solved to determine whether it satisfies the safety rules. If it does, the performance of the longitudinal trajectory can be comprehensively evaluated using a multi-objective cost function based on the trajectory equation corresponding to the endpoint state, thereby determining the optimal safety factor. By determining the unique longitudinal trajectory and combining it with the aforementioned lateral trajectory, the target lane-changing trajectory is obtained.

[0054] If none of the trajectory equations corresponding to all endpoint states have a solution, it means that the lane-changing trajectory interval does not meet the safety rules. In this case, the vehicle continues to travel in the current lane and the lane-changing decision is not executed. It is understood that in this embodiment, there can be one or more endpoint states.

[0055] According to some embodiments of this application, the decision to change lanes to the target lane in step S105 of this application embodiment may include, but is not limited to, the following steps: Step S601: If the current time is before the first critical time, then the first constraint rule, the second constraint rule and the third constraint rule are used to replan the trajectory of the lane change trajectory after the current traffic state is updated to obtain the dynamic tracking trajectory. Step S602: If the current time is after the first critical time, then the second constraint rule and the third constraint rule are used to replan the trajectory of the lane-changing trajectory after the current traffic state update to obtain the dynamic tracking trajectory.

[0056] This embodiment performs safety checks on scenarios predicted by LSTM as lane changes to correct erroneous lane-changing decisions, achieving integrated decision-making and planning. If the lane-changing decision meets the safety check, the lane-changing execution phase begins. The specific process of the lane-changing execution phase is as follows: After confirming that the initially planned lane-changing trajectory section meets safety rules, the system enters the lane-changing execution phase. During lane-changing, the system first optimizes the lane-changing trajectory section based on comfort and communication efficiency to obtain the safest target lane-changing trajectory. While tracking the target lane-changing trajectory, the system periodically collects speed and position data of surrounding vehicles and calculates the vehicle distance at critical moments. If each vehicle distance is greater than the specified safe distance, the main vehicle continues to travel according to the target lane-changing trajectory; otherwise, the trajectory is updated again according to the planning algorithm, as follows: To address the changes in vehicle motion during lane changes and ensure lane-changing safety, a three-segment trajectory replanning algorithm based on collision avoidance analysis is proposed. Based on the trajectory execution status, corresponding data is collected, and the trajectory is updated in a targeted manner to minimize computational power consumption and improve trajectory timeliness. when At that time, data from three adjacent vehicles is collected and the distance between them is calculated to determine whether to perform replanning. During replanning, the complete set of collision avoidance inequalities mentioned above is used to determine the appropriate method. , The range of values ​​for; when At that time, only data on the vehicles in front and behind in the target lane are collected for judgment. During replanning, the last three terms of the collision avoidance inequality are used to determine the collision avoidance. , The range of values ​​for; when At that time, only data on the vehicle in front in the target lane is collected for judgment, and during replanning, only the last two terms of the collision avoidance inequality are used to determine the collision avoidance. The range of values, in addition to the target vehicle speed Set to the speed of the vehicle in front.

[0057] This embodiment improves system computational efficiency and reduces computing power consumption by continuously and reasonably reducing the number of constraint rules used for security verification and trajectory planning over time.

[0058] Please refer to Figure 4 This application provides a high-quality data-driven integrated decision-making and planning method for safe lane changing in intelligent vehicles. This includes a lane-changing decision evaluation function, an integrated decision-making and planning lane-changing method, and a corresponding trajectory replanning method. First, high-quality data is selected based on the lane-changing decision evaluation function, and a lane-changing decision model is trained based on this data to address the interference of abnormal logical samples in the training data on the model's correct output. Second, a trajectory planning model based on polynomial curves is established, and collision scenarios are analyzed to derive collision avoidance inequalities. A decision module based on safety rules is built as a post-processing layer for the decision model, resolving the correlation between decision-making and planning and achieving integrated decision-making and planning. Finally, a trajectory replanning method is proposed to address the lack of adaptability of planning algorithms to dynamic scenarios. The main vehicle can dynamically respond to changes in the speed of interfering vehicles and adjust its trajectory in a timely manner to maintain a safe distance.

[0059] Compared with the prior art, the embodiments of this application integrate the lane-changing decision module and the trajectory planning module, and propose an integrated decision-making and planning method, which has at least the following beneficial effects: (1) In view of the problem that there are abnormal samples in the lane-changing samples in the dataset, this application proposes a lane-changing decision evaluation function based on the lane-changing logic to screen the samples and effectively improve the quality of the lane-changing decision dataset.

[0060] (2) Regarding the issue of the relationship between the decision-making module and the planning module, this application embodiment constructs a safety rule module as LSTM post-processing based on the derived collision avoidance inequality to achieve integrated decision-making. (3) This application further proposes a trajectory replanning algorithm based on the integrated decision planning method of this embodiment, which improves lane changing safety.

[0061] This application also proposes a vehicle autonomous lane-changing decision-making and planning system, including: The first module is used to make lane-changing decisions based on the current traffic conditions through a lane-changing decision model, and to obtain the lane-changing intention; the lane-changing intention is used to determine whether to change lanes to the target lane. The second module is used to perform preliminary trajectory planning based on the current traffic conditions and the target lane when the lane-changing intention is to change lanes to the target lane, and to obtain the lane-changing trajectory range. The third module is used to verify the safety rules of the lane-changing trajectory section and obtain the verification results. The fourth module is used to continue in the current lane and re-verify the safety rules of the lane-changing trajectory section if the verification result shows that the lane-changing trajectory section does not meet the safety rules. The fifth module is used to execute the decision to change lanes to the target lane if the verification result shows that the lane change trajectory section meets the safety rules.

[0062] It is understood that the methods described in the above method embodiments are applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0063] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0064] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0065] Please see Figure 5 , Figure 5 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 901 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 902 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 902 and is called and executed by the processor 901. The 903 input / output interface is used to implement information input and output. The communication interface 904 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 905 transmits information between various components of the device (e.g., processor 901, memory 902, input / output interface 903, and communication interface 904); The processor 901, memory 902, input / output interface 903, and communication interface 904 are connected to each other within the device via bus 905.

[0066] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0067] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0068] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0069] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0070] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0071] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0072] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0073] The system embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0074] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0075] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or modules is not necessarily limited to those steps or modules explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0076] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0077] In the embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules 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.

[0078] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; 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.

[0079] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0080] If the integrated module is implemented as a software functional module 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 application, in essence, or the part that contributes to the prior art, or all or 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 multiple 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 application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0081] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A vehicle autonomous lane-changing decision-making and planning method, characterized in that, Includes the following steps: The lane-changing decision model is used to make lane-changing decisions based on the current traffic conditions, thereby obtaining the lane-changing intention. The lane change intention is used to determine whether to change lanes to the target lane; When the lane-changing intention is to change lanes to the target lane, preliminary trajectory planning is performed based on the current traffic conditions and the target lane to obtain the lane-changing trajectory range; The safety rules of the lane-changing trajectory section are verified, and the verification results are obtained. If the verification result indicates that the lane-changing trajectory section does not meet the safety rules, then continue to stay in the current lane; If the verification result indicates that the lane-changing trajectory section meets the safety rules, then the decision to change lanes to the target lane is executed. The safety rules include a first constraint rule regarding the displacement of the main lane at the first critical moment when the main vehicle does not collide in its own lane, a second constraint rule regarding the speed and displacement of the target lane at the second critical moment when the main vehicle collides in the target lane, and a third constraint rule regarding the speed and displacement of the target lane at the moment the lane change is completed. The decision intention to change lanes to the target lane includes the following steps: If the current time is before the first critical time, then the first constraint rule, the second constraint rule and the third constraint rule are used to replan the trajectory of the lane-changing trajectory interval updated based on the current traffic state to obtain a dynamic tracking trajectory. If the current time is after the first critical time, the second constraint rule and the third constraint rule are used to replan the trajectory of the lane-changing trajectory interval updated based on the current traffic state to obtain a dynamic tracking trajectory.

2. The method according to claim 1, characterized in that, The lane-changing decision model is trained through the following steps: Based on the correlation between traffic information around the main vehicle and sample labels, the quality of each sample data is assessed and high-quality samples are selected; the sample labels are for performing lane changing or lane keeping. A training dataset is formed based on multiple selected high-quality samples; The long short-term memory neural network was trained using the training dataset to obtain the lane-changing decision model.

3. The method according to claim 2, characterized in that, The process of evaluating the quality of each sample data based on the correlation between traffic information around the main vehicle and sample labels, and selecting high-quality samples, includes the following steps: Based on the sample labels, the sample data is divided into either the first sample corresponding to lane changing or the second sample corresponding to lane keeping; The first quality score of the first sample is determined based on the relative state of the main vehicle and surrounding vehicles at the starting point of the lane change, and high-quality samples are selected from multiple first samples based on the first quality score. The second quality score of the second sample is determined based on the average relative state between the main vehicle and surrounding vehicles during the lane change decision period, and high-quality samples are selected from multiple second samples based on the second quality score.

4. The method according to claim 1, characterized in that, The preliminary trajectory planning based on the current traffic conditions and the target lane to obtain the lane-changing trajectory interval includes the following steps: Initialize the destination state of the main vehicle after lane changing based on the current traffic conditions and the target lane; Trajectory planning is performed based on the endpoint state and the starting state of the main vehicle to obtain a trajectory equation that takes safety factors into account. The trajectory equation is used to represent the trajectory sub-interval corresponding to the endpoint state, and multiple trajectory sub-intervals of the endpoint states together represent the lane-changing trajectory interval.

5. The method according to claim 4, characterized in that, The process of performing safety rule verification on the lane-changing trajectory section to obtain the verification result includes the following steps: Perform safety rule verification on the trajectory sub-intervals of the current endpoint state; If the trajectory sub-interval of the current destination state does not meet the safety rules, the destination state is updated according to the current traffic state and the target lane. The steps of trajectory planning based on the destination state and the starting state of the main vehicle to obtain the trajectory equation considering the safety factor and the steps of verifying the safety rules of the trajectory sub-interval of the current destination state are repeated until the trajectory sub-interval of the current destination state meets the safety rules. Then, the verification result is determined to be that the lane-changing trajectory interval meets the safety rules. If none of the trajectory sub-intervals in all endpoint states meet the safety rules, then the verification result is determined to be that the lane-changing trajectory interval does not meet the safety rules.

6. A vehicle autonomous lane-changing decision-making and planning system, characterized in that, include: The first module is used to make lane-changing decisions based on the current traffic conditions through a lane-changing decision model, and to obtain the lane-changing intention. The lane change intention is used to determine whether to change lanes to the target lane; The second module is used to perform preliminary trajectory planning based on the current traffic conditions and the target lane when the lane-changing intention is to change lanes to the target lane, and to obtain the lane-changing trajectory range. The third module is used to perform safety rule verification on the lane-changing trajectory section and obtain the verification result. The fourth module is used to continue to stay in the current lane and re-verify the safety rules of the lane-changing trajectory section according to the real-time updated current traffic status if the verification result shows that the lane-changing trajectory section does not meet the safety rules. The fifth module is used to execute the decision to change lanes to the target lane if the verification result shows that the lane-changing trajectory section meets the safety rules. The safety rules include a first constraint rule regarding the displacement of the main lane at the first critical moment when the main vehicle does not collide in its own lane, a second constraint rule regarding the speed and displacement of the target lane at the second critical moment when the main vehicle collides in the target lane, and a third constraint rule regarding the speed and displacement of the target lane at the moment the lane change is completed. The decision intention to change lanes to the target lane includes the following steps: If the current time is before the first critical time, then the first constraint rule, the second constraint rule and the third constraint rule are used to replan the trajectory of the lane-changing trajectory interval updated based on the current traffic state to obtain a dynamic tracking trajectory. If the current time is after the first critical time, the second constraint rule and the third constraint rule are used to replan the trajectory of the lane-changing trajectory interval updated based on the current traffic state to obtain a dynamic tracking trajectory.

7. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 5.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 5.

Images