Automatic driving vehicle lane selection optimization method and system, medium and equipment

Abstract

The invention provides an automatic driving vehicle lane selection optimization method and system, a medium and equipment, and belongs to the technical field of automatic drive.The method comprises the steps that image data around a vehicle, vehicle motion state data, road information data and traffic flow data are obtained; based on the collected image data, a deep learning image recognition algorithm is used for recognizing trucks around the truck, and the motion state of the trucks is continuously tracked and analyzed; according to the motion state of the truck, in combination with motion state data of vehicles around the truck, road information data and traffic flow data, predicting a driving track of the truck in a future time period; and according to the motion state of the truck and the predicted driving track of the truck in the future time period, the influence of the truck when different lanes are selected is analyzed, and the lane with the minimum influence on the truck is determined as a target lane. According to the method, the situation that the driving efficiency is reduced due to the fact that the automatic driving vehicle follows the truck or runs close to the truck can be avoided.

Description

Technical Field

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

[0002] With the continuous development of autonomous driving technology, lane selection strategies in autonomous driving mode are crucial for driving efficiency and safety. Current autonomous driving lane selection typically relies on fixed lane planning or choosing lanes with shorter convoys. However, in real-world road environments, special vehicles like large trucks, due to their large size, slow acceleration, and poor maneuverability, can significantly impact the driving of surrounding vehicles. In manual driving scenarios, drivers often prioritize avoiding lanes with large trucks, even if the convoy is short. However, current autonomous driving systems generally do not adequately consider this critical factor when selecting lanes, resulting in situations where the driving efficiency and safety of autonomous vehicles fail to reach their optimal levels. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art, and proposes a method, system, medium and device for optimizing lane selection of autonomous vehicles.

[0004] In a first aspect, embodiments of the present invention provide a lane selection optimization method for autonomous vehicles, comprising the following steps:

[0005] S100: Acquire image data, vehicle motion status data, road information data, and traffic flow data surrounding the vehicle;

[0006] S200: Based on the collected image data, the vehicle uses a deep learning image recognition algorithm to identify the large trucks around it and continuously track and analyze the movement status of the large trucks.

[0007] S300: Based on the movement status of the truck, combined with the movement status data of surrounding vehicles, road information data, and traffic flow data, predict the truck's trajectory in the future.

[0008] S400: Based on the movement status of the truck and the predicted driving trajectory of the truck in the future, analyze the impact of selecting different lanes on the truck and determine the lane that will have the least impact on this vehicle as the target lane.

[0009] Furthermore, in step S200, the deep learning image recognition algorithm is the YOLOV8 model.

[0010] Furthermore, the steps for identifying large trucks based on the YOLOv8 model are as follows:

[0011] S210. Preprocessing: Preprocess the acquired image data;

[0012] S220, Object Detection Inference: Input the preprocessed image into the trained YOLOv8 model, perform forward propagation, and output the bounding boxes, confidence scores, and class probabilities of all detected vehicles;

[0013] S230, Large Truck Classification: Filter bounding boxes from the output whose category probability is greater than the threshold, and select the large truck category; for each candidate box, extract the bounding box coordinates, including the center point x and y coordinates, width w, and height h;

[0014] S240. Post-processing: Apply non-maximum suppression to remove overlapping bounding boxes, calculate the IoU between all boxes, and retain boxes with IoU below the threshold.

[0015] S250, Output Results: Output the final truck detection results, including its location and confidence level in the image.

[0016] Furthermore, the continuous tracking and analysis of the movement status of large trucks includes the following steps:

[0017] S250, Target Tracking Initialization: When a truck is detected for the first time, a tracker is initialized;

[0018] S260, State Prediction and Update: Use Kalman filtering to predict the position of the truck in the next frame. The state vector includes position, velocity, and acceleration.

[0019] S270, Continuous parameter estimation: Calculates velocity, acceleration, and direction in real time from the tracked trajectory and smooths them to reduce noise;

[0020] S280, Trajectory Analysis: Analyze historical trajectory points, fit a motion model, and use it to predict future trajectories.

[0021] Further, step S300 includes:

[0022] S310. Based on the movement status of the truck, combined with the movement status data of surrounding vehicles, road information data and traffic flow data, use the intent recognition rule base to determine the short-term driving intent of the truck.

[0023] S320: Based on the identified short-term driving intentions of the truck, predict the future position of the truck using a kinematic extrapolation model.

[0024] Furthermore, the intent recognition rule base is as follows: if a large truck is in the center of the lane and has not activated its turn signal, the intent is to keep the lane; if a large truck continues to approach the lane line and activates its turn signal, the intent is to change lanes; if a large truck is in a dedicated left-turn or right-turn lane, the intent is to turn left or right at the intersection; if the intersection ahead is red and there is a long time remaining, the intent is to slow down and stop.

[0025] Furthermore, the kinematic extrapolation model is as follows:

[0026] The formula for calculating the future position of a large truck when maintaining lane position is as follows:

[0027]

[0028] in, Let be the predicted position vector at time t. For the optimal position estimate at time t-1, Let be the optimal velocity vector at time t-1. Let be the optimal acceleration vector at time t-1. For time intervals;

[0029] When cornering or changing lanes, the vehicle's trajectory should conform to the lane centerline or standard turning path provided in the high-precision map.

[0030] Furthermore, in step S400, the influence of vehicle density, average vehicle speed, path matching degree with destination, and traffic light status in different lanes is analyzed to determine the lane with the least comprehensive influence on the vehicle as the target lane.

[0031] Furthermore, the comprehensive impact scores of large trucks, vehicle density within the lane, average vehicle speed, route matching to the destination, and traffic light status were analyzed when different lanes were selected. The lane with the lowest comprehensive impact score was determined as the target lane. The formula for calculating the comprehensive impact score is as follows:

[0032]

[0033] in, To comprehensively assess the impact score, The weight of the impact factor for large trucks, For the impact factor of large trucks, The weights of the factors influencing vehicle density within the lane. The influencing factor of vehicle density within the lane. The weight of the influence factor of average vehicle speed. The factor affecting the average speed of vehicles. The weights of the factors influencing the path matching degree with the destination are determined. The influencing factor of the route matching degree with the destination. The weights of the factors affecting traffic signal status. This refers to the factors affecting the status of traffic lights.

[0034] Secondly, embodiments of the present invention provide an autonomous vehicle lane selection optimization system, comprising:

[0035] The acquisition module is used to acquire image data, vehicle motion status data, road information data, and traffic flow data around the vehicle.

[0036] The identification module is used to identify large trucks around the vehicle based on the collected image data and using deep learning image recognition algorithms, and to continuously track and analyze the movement status of the large trucks.

[0037] The prediction module is used to predict the truck's trajectory in the future based on the truck's motion status, combined with data on the motion status of surrounding vehicles, road information, and traffic flow.

[0038] The determination module is used to analyze the impact of selecting different lanes on the truck based on the truck's motion status and predicted driving trajectory in the future, and to determine the lane that will have the least impact on the vehicle as the target lane.

[0039] Thirdly, embodiments of the present invention provide an electronic device, including:

[0040] One or more processors;

[0041] Memory, used to store one or more programs;

[0042] When the one or more programs are executed by the one or more processors, the one or more processors perform the method as described above.

[0043] Fourthly, embodiments of the present invention provide a computer-readable medium storing a computer program, which, when executed by a processor, implements the steps of the method described above.

[0044] The present invention provides a lane selection optimization method, system, medium, and device for autonomous vehicles. This method acquires image data, vehicle motion state data, road information data, and traffic flow data surrounding the vehicle. Based on the acquired image data, a deep learning image recognition algorithm is used to identify large trucks around the vehicle, and the motion state of these trucks is continuously tracked and analyzed. Based on the trucks' motion state, combined with the surrounding vehicle motion state data, road information data, and traffic flow data, the vehicle's trajectory in the future is predicted. Based on the trucks' motion state and the predicted trajectory, the impact of different lane selections on the trucks is analyzed, and the lane least affected by the trucks is determined as the target lane. This method incorporates truck factors as an important reference into the autonomous driving lane selection decision, effectively avoiding reduced driving efficiency caused by following or driving close to large trucks. Attached Figure Description

[0045] Figure 1 A flowchart illustrating a lane selection optimization method for autonomous vehicles provided in an embodiment of the present invention;

[0046] Figure 2 A structural block diagram of an autonomous vehicle lane selection optimization system provided in an embodiment of the present invention;

[0047] Figure 3 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0048] To enable those skilled in the art to better understand the technical solutions of the present invention, exemplary embodiments of the present invention are described below in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0049] Where there is no conflict, the various embodiments of the present invention and the features thereof may be combined with each other.

[0050] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0051] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Terms such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0052] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having the meaning consistent with their meaning in the context of the relevant art and the invention, and will not be interpreted as having an idealized or overly formal meaning unless expressly so defined herein.

[0053] This invention provides a method for optimizing lane selection for autonomous vehicles. Figure 1 This is a flowchart illustrating a lane selection optimization method for autonomous vehicles provided in an embodiment of the present invention. The method includes the following steps:

[0054] S100: Acquire image data, vehicle motion status data, road information data, and traffic flow data surrounding the vehicle.

[0055] Specifically, the vehicle utilizes its onboard cameras to collect real-time image data of the surrounding environment. It also uses onboard millimeter-wave radar and lidar sensors to collect real-time data on the position, speed, and other vehicle motion status of other vehicles in the vicinity. Through vehicle-to-everything (V2X) technology, it obtains real-time traffic flow data from surrounding roads, as well as shared driving intentions data from other vehicles, including whether they are using left or right turn signals, or not. Static information data comes from high-precision maps and consists of pre-stored, infrequently changing basic road information, including lane geometry and topology data and road infrastructure and rule data. Lane geometry and topology data includes lane line types (such as solid lines, dashed lines, double solid lines, etc.), lane width, curvature / curve radius, and slope. Road infrastructure and rule data includes traffic signs (such as speed limit signs, no truck signs, turning restriction signs, etc.), traffic light positions and phase sequences (i.e., the remaining time of the red light at the intersection ahead), intersection types and lane guidance (such as crossroads, T-junctions, ramps, whether the lane is for straight-ahead, left-turn, or right-turn only), and special areas (such as bus stops, toll stations, checkpoints, and highway entrance / exit ramps).

[0056] S200: Based on the collected image data, it uses a deep learning image recognition algorithm to identify the large trucks around the vehicle and continuously tracks and analyzes the movement status of the large trucks.

[0057] In one embodiment, the deep learning image recognition algorithm is the YOLOv8 model.

[0058] Specifically, the YOLOv8 model includes: an input layer that receives RGB images, typically resized to 640×640 pixels and normalized; a backbone network using a CSPDarknet53 structure for feature extraction, containing multiple convolutional layers and a CSP (Cross Stage Partial) module, outputting multi-scale feature maps; a neck using a PANet (Path Aggregation Network) structure for feature fusion, combining deep and shallow features to enhance small object detection capabilities; and a head that outputs detection results, including bounding box coordinates (x, y, w, h), confidence score (indicating whether an object is inside the box), and class probability (e.g., large truck, ordinary vehicle, etc.). YOLOv8 uses an anchor-free mechanism to directly predict bounding boxes. The YOLOv8 loss function consists of several parts, including bounding box loss, classification loss, and confidence loss. The loss function formula is as follows:

[0059]

[0060] Where Lbox is the bounding box loss, using CIoU (Complete Intersection over Union) loss; Lcls is the classification loss, using binary cross-entropy (BCE) loss, used to determine the category (e.g., large truck); Ldfl is the distribution focus loss, used to optimize bounding box regression and improve accuracy; λ box , λ cls and λ dfl These are weighting coefficients, which are usually adjusted based on the dataset;

[0061] The formula for calculating the bounding box loss (Lbox) is as follows:

[0062]

[0063] Where IoU is the intersection-union ratio of the predicted bounding box and the ground truth bounding box, ρ is the Euclidean distance, and b is the distance between the predicted and ground truth bounding boxes. pred and b gt α and β are the center points of the predicted bounding box and the ground truth bounding box, respectively; c is the diagonal length of the minimum bounding rectangle; α is the weighting coefficient; and k is the aspect ratio consistency measure.

[0064] The classification loss Lcls is calculated using the following formula:

[0065]

[0066] Among them, y i It's a real tag, p i It is a predicted probability.

[0067] In one embodiment, the steps for identifying large trucks based on the YOLOv8 model are as follows:

[0068] S210. Preprocessing: Preprocess the acquired image data, including: image denoising, contrast enhancement, size adjustment (e.g., 640×640) and pixel normalization (scaling pixel values ​​to the range of [0,1]).

[0069] S220, Object Detection Inference: Input the preprocessed image into the trained YOLOv8 model, perform forward propagation, and output the bounding boxes, confidence scores, and class probabilities of all detected vehicles; the YOLOv8 model has learned the truck category during training, using COCO or a custom dataset containing truck labels;

[0070] S230, Large Truck Classification: Filter bounding boxes from the output whose category probability is greater than a threshold (e.g., 0.5) and select the large truck category; for each candidate box, extract the bounding box coordinates, including the center point x and y coordinates, width w, and height h;

[0071] S240. Post-processing: Apply nonmaximum suppression (NMS) to remove overlapping bounding boxes, calculate the IoU between all boxes, and retain boxes with IoU below a threshold (e.g., 0.5) to ensure that each truck is detected only once.

[0072] S250, Output Results: Output the final truck detection results, including its location and confidence level in the image.

[0073] In one embodiment, the motion state of the truck includes position, direction of travel, speed, and acceleration.

[0074] The method for obtaining the position, direction of travel, speed, and acceleration of a large truck from image data recognition results is as follows:

[0075] Location acquisition: Starting from the center point coordinates (x, y) of the detection bounding box of the truck, and combining the camera calibration parameters (intrinsic and extrinsic parameters), the image coordinates are converted into position coordinates (X, Y, Z) in the world coordinate system through inverse perspective transformation. Assuming the ground is flat and coordinate Z=0, the 2D position coordinates (X, Y) of the truck are obtained.

[0076] Direction of travel acquisition: The direction of travel vector is calculated by observing the positional changes of the truck in consecutive frames (e.g., two adjacent frames). The formula for calculating the direction angle θ is as follows:

[0077]

[0078] Among them, X t Y t It is the 2D position coordinate of the current frame, X t-1 Y t-1 These are the 2D position coordinates of the previous frame;

[0079] Velocity acquisition: Calculate the instantaneous velocity v by dividing the position change by the time interval Δt (e.g., 1 / 30 of a second).

[0080]

[0081] Acceleration acquisition: Acceleration 'a' is calculated by dividing the velocity change by the time interval.

[0082]

[0083] Among them, v t It is the velocity of the current frame, v t-1 It's the speed of the previous frame.

[0084] In one embodiment, continuously tracking and analyzing the motion status of a large truck includes the following steps:

[0085] S250, Target Tracking Initialization: When a large truck is detected for the first time, initialize a tracker (e.g., using the DeepSORT algorithm, which combines Kalman filtering and appearance features for data association).

[0086] S260, State Prediction and Update: Use Kalman filtering to predict the position of the truck in the next frame. The state vector includes position, velocity, and acceleration.

[0087] The Kalman filter formula is as follows:

[0088]

[0089] in, The predicted state at the current moment. This is an estimate of the optimal state at the previous time step. Let the prediction error covariance be at the current moment. The covariance of the estimation error at the previous time step. Here is the state transition matrix. This is the transpose of the state transition matrix. This is the process noise matrix.

[0090] S270, Continuous parameter estimation: Calculates velocity, acceleration, and direction in real time from the tracked trajectory and smooths them to reduce noise;

[0091] S280. Trajectory Analysis: Analyze historical trajectory points and fit motion models (such as linear or curvilinear models) to predict future trajectories.

[0092] S300: Based on the movement status of the truck, combined with the movement status data of surrounding vehicles, road information data, and traffic flow data, predict the truck's trajectory in the future, such as determining whether the truck is about to enter the lane, or whether it may turn or stop at the intersection ahead.

[0093] In one embodiment, step S300 includes:

[0094] S310. Based on the movement status of the truck, combined with the movement status data of surrounding vehicles, road information data and traffic flow data, use the intent recognition rule base to determine the short-term driving intent of the truck.

[0095] Specifically, the intent recognition rule base is as follows: if a large truck is in the center of the lane and has not used its turn signal, the intent is to keep the lane; if a large truck continues to approach the lane line and uses its turn signal, the intent is to change lanes; if a large truck is in a dedicated left or right turn lane, the intent is to turn left or right at the intersection; if the intersection ahead is red and there is a long time remaining, the intent is to slow down and stop.

[0096] S320: Based on the identified short-term driving intentions of the truck, predict the future position of the truck using a kinematic extrapolation model.

[0097] Specifically, the kinematic extrapolation model is as follows:

[0098] The formula for calculating the future position of a large truck when maintaining lane position is as follows:

[0099]

[0100] in, Let be the predicted position vector at time t. For the optimal position estimate at time t-1, Let be the optimal velocity vector at time t-1. Let be the optimal acceleration vector at time t-1. For time intervals;

[0101] When cornering or changing lanes, the vehicle's trajectory should conform to the lane centerline or standard turning path provided in the high-precision map.

[0102] Specifically, the prediction process in step S300 is as follows: The motion state of the truck is acquired, and combined with the motion state data of surrounding vehicles, road information data, and traffic flow data, the most suitable driving intention is matched using an intent recognition rule base. If the intention is lane keeping, the trajectory is extrapolated forward along the center line of the current lane. If the intention is lane changing, an S-shaped trajectory smoothly connecting the current lane and the target lane is generated. If the intention is turning at an intersection, an arc-shaped trajectory along the turning lane is generated based on map data. If the intention is decelerating and stopping, a trajectory decelerating to the stop line at the intersection is generated. A future trajectory point sequence representing the most likely driving path of the truck, along with the predicted speed change trend, is output.

[0103] S400: Based on the movement status of the truck and the predicted driving trajectory of the truck in the future, analyze the impact of selecting different lanes on the truck and determine the lane that will have the least impact on this vehicle as the target lane.

[0104] In one embodiment, the effects of different lane selections on vehicle density, average vehicle speed, path matching with destination, and traffic light status are also analyzed to determine the lane with the least overall impact on the vehicle as the target lane.

[0105] Specifically, the analysis considers the combined impact scores of factors such as the number of trucks, vehicle density within the lane, average vehicle speed, route matching to the destination, and traffic light status when selecting different lanes. The lane with the lowest combined impact score is then selected as the target lane. The formula for calculating the combined impact score is as follows:

[0106]

[0107] in, To comprehensively assess the impact score, The weight of the impact factor for large trucks, For the impact factor of large trucks, The weights of the factors influencing vehicle density within the lane. The influencing factor of vehicle density within the lane. The weight of the influence factor of average vehicle speed. The factor affecting the average speed of vehicles. The weights of the factors influencing the path matching degree with the destination are determined. The influencing factor of the route matching degree with the destination. The weights of the factors affecting traffic signal status. This refers to the factors affecting the status of traffic lights.

[0108] The sum of all weights is 1. The weight of the impact factor of large trucks can be reasonably adjusted based on actual road tests and data analysis, and is generally set between 0.3 and 0.5 to highlight its importance.

[0109] The impact factor for large trucks is assigned different scores based on the distance between the truck and the vehicle, the relative speed, and the lane position. The closer the truck, the greater the relative speed difference, and the truck is directly in front of the vehicle in the direction of travel, the higher the impact score. For example, if a large truck is within 50 meters in front of the vehicle and traveling in the same direction, its impact score is set to a higher value; if the truck is in an adjacent lane and farther away, the impact score is correspondingly lower.

[0110] The formula for calculating the impact factor of heavy trucks is as follows:

[0111]

[0112] in, The weights of the distance influence factor, Distance is the influencing factor. The weights of the relative speed influence factor, The relative speed influence factor, The weights of the lane position influencing factors, This refers to the lane position influencing factor.

[0113] The sum of all weights is 1, which is preferred. . The range is between 0 and 100.

[0114] Distance Influence Factor The following is confirmed:

[0115]

[0116] in, This is the straight-line distance between this vehicle and the large truck, in meters.

[0117] Relative speed influence factor The following is confirmed:

[0118]

[0119] in, The relative speed between the truck and this vehicle, in meters per second.

[0120] Lane position influence factor The following is confirmed:

[0121]

[0122] Scoring rules are also established for other factors such as lane density, average vehicle speed, route fit to destination, and traffic light status. For example, the lower the lane density and the higher the average speed, the higher the lane's score; the higher the route fit to destination, the higher the score; and regarding traffic lights, if a red light is approaching and the red light duration is long, the lane's score will decrease.

[0123] Lane vehicle density influencing factor The following is confirmed:

[0124]

[0125] in, Lane vehicle density, unit: vehicles / meter, is calculated by dividing the number of vehicles in the lane by the lane length.

[0126] Vehicle average speed influence factor The following is confirmed:

[0127]

[0128] in, Average vehicle speed, in kilometers per hour.

[0129] Influence factor on path matching with destination The following is confirmed:

[0130]

[0131] in, Path matching degree with destination, unit: dimensionless, between 0 and 1, 1 indicates a perfect match.

[0132] Traffic signal light status influencing factors The following is confirmed:

[0133]

[0134] Where t0 is the remaining time of the red light ahead, in seconds.

[0135] This invention acquires image data, vehicle motion state data, road information data, and traffic flow data surrounding the vehicle. Based on the acquired image data, a deep learning image recognition algorithm is used to identify large trucks around the vehicle, and the motion state of the trucks is continuously tracked and analyzed. Based on the trucks' motion state, combined with the surrounding vehicle motion state data, road information data, and traffic flow data, the vehicle's trajectory in the future is predicted. Based on the trucks' motion state and the predicted trajectory, the impact of trucks on different lane selections is analyzed, and the lane least affected by the vehicle is determined as the target lane. This invention incorporates truck factors as an important reference into the autonomous driving lane selection decision, effectively avoiding the reduction in driving efficiency caused by autonomous vehicles following or driving close to large trucks.

[0136] The present invention also provides a lane selection optimization system for autonomous vehicles. Figure 2 A structural block diagram of an autonomous vehicle lane selection optimization system provided in an embodiment of the present invention is shown. The system includes:

[0137] The acquisition module 11 is used to acquire image data, vehicle motion status data, road information data and traffic flow data around the vehicle;

[0138] The identification module 12 is used to identify the large trucks around the vehicle based on the collected image data using a deep learning image recognition algorithm, and to continuously track and analyze the movement status of the large trucks.

[0139] The prediction module 13 is used to predict the driving trajectory of the truck in the future period based on the truck's motion status, combined with the motion status data of surrounding vehicles, road information data and traffic flow data.

[0140] The determination module 14 is used to analyze the impact of selecting different lanes on the truck based on the truck's motion state and the predicted driving trajectory of the truck in the future, and to determine the lane that will have the least impact on the vehicle as the target lane.

[0141] This invention also provides an electronic device. Figure 3 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Figure 3 As shown, an embodiment of the present invention provides an electronic device including: one or more processors 101, a memory 102, and one or more I / O interfaces 103. The memory 102 stores one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement any of the lane selection optimization methods for autonomous vehicles described in the above embodiments; the one or more I / O interfaces 103 are connected between the processors and the memory, configured to enable information interaction between the processors and the memory.

[0142] The processor 101 is a device with data processing capabilities, including but not limited to a central processing unit (CPU); the memory 102 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), and flash memory (FLASH); the I / O interface (read / write interface) 103 is connected between the processor 101 and the memory 102, and can realize information interaction between the processor 101 and the memory 102, including but not limited to a data bus (Bus).

[0143] In some embodiments, the processor 101, memory 102, and I / O interface 103 are interconnected via bus 104, and thus connected to other components of the computing device.

[0144] In some embodiments, the one or more processors 101 include a field-programmable gate array.

[0145] This invention also provides a computer-readable medium. The computer-readable medium stores a computer program, which, when executed by a processor, implements the steps in any of the lane selection optimization methods for autonomous vehicles described in the above embodiments. The computer-readable storage medium may be volatile or non-volatile.

[0146] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0147] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0148] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0149] The computer program instructions used to perform the operations of this invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions. This electronic circuitry can execute the computer-readable program instructions to implement various aspects of the invention.

[0150] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0151] Various aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0152] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0153] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0154] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0155] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in conjunction with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of the invention as set forth in the appended claims.

Claims

1. A lane selection optimization method for autonomous vehicles, characterized in that, Includes the following steps: S100: Acquire image data, vehicle motion status data, road information data, and traffic flow data surrounding the vehicle; S200: Based on the collected image data, the vehicle uses a deep learning image recognition algorithm to identify the large trucks around it and continuously track and analyze the movement status of the large trucks. S300: Based on the movement status of the truck, combined with the movement status data of surrounding vehicles, road information data, and traffic flow data, predict the truck's trajectory in the future. S400: Based on the movement status of the truck and the predicted driving trajectory of the truck in the future, analyze the impact of selecting different lanes on the truck and determine the lane that will have the least impact on this vehicle as the target lane.

2. The method according to claim 1, characterized in that, In step S200, the deep learning image recognition algorithm is the YOLOV8 model.

3. The method according to claim 2, characterized in that, The steps for identifying large trucks based on the YOLOv8 model are as follows: S210. Preprocessing: Preprocess the acquired image data; S220, Object Detection Inference: Input the preprocessed image into the trained YOLOv8 model, perform forward propagation, and output the bounding boxes, confidence scores, and class probabilities of all detected vehicles; S230, Large Truck Classification: Filter bounding boxes from the output whose category probability is greater than the threshold, and select the large truck category; for each candidate box, extract the bounding box coordinates, including the center point x and y coordinates, width w, and height h; S240. Post-processing: Apply non-maximum suppression to remove overlapping bounding boxes, calculate the IoU between all boxes, and retain boxes with IoU below the threshold. S250, Output Results: Output the final truck detection results, including its location and confidence level in the image.

4. The method according to claim 3, characterized in that, Continuous tracking and analysis of the movement status of large trucks includes the following steps: S250, Target Tracking Initialization: When a truck is detected for the first time, a tracker is initialized; S260, State Prediction and Update: Use Kalman filtering to predict the position of the truck in the next frame. The state vector includes position, velocity, and acceleration. S270, Continuous parameter estimation: Calculates velocity, acceleration, and direction in real time from the tracked trajectory and smooths them to reduce noise; S280, Trajectory Analysis: Analyze historical trajectory points, fit a motion model, and use it to predict future trajectories.

5. The method according to claim 1, characterized in that, Step S300 includes: S310. Based on the movement status of the truck, combined with the movement status data of surrounding vehicles, road information data and traffic flow data, use the intent recognition rule base to determine the short-term driving intent of the truck. S320: Based on the identified short-term driving intentions of the truck, predict the future position of the truck using a kinematic extrapolation model.

6. The method according to claim 5, characterized in that, The intent recognition rule base is as follows: If the truck is in the center of the lane and has not used its turn signal, the intent is to keep the lane; if the truck continues to approach the lane line and uses its turn signal, the intent is to change lanes; if the truck is in the left or right turn lane, the intent is to turn left or right at the intersection; if the intersection ahead is red and there is a long time remaining, the intent is to slow down and stop.

7. The method according to claim 6, characterized in that, The kinematic extrapolation model is as follows: The formula for calculating the future position of a large truck when maintaining lane position is as follows: in, Let be the predicted position vector at time t. For the optimal position estimate at time t-1, Let be the optimal velocity vector at time t-1. Let be the optimal acceleration vector at time t-1. For time intervals; When cornering or changing lanes, the vehicle's trajectory should conform to the lane centerline or standard turning path provided in the high-precision map.

8. The method according to claim 1, characterized in that, In step S400, the influence of vehicle density, average vehicle speed, path matching degree with destination, and traffic light status in different lanes is analyzed to determine the lane with the least comprehensive influence on the vehicle as the target lane.

9. The method according to claim 8, characterized in that, The comprehensive impact score of different lane selections is analyzed based on factors such as the number of trucks, vehicle density within the lane, average vehicle speed, route matching to the destination, and traffic light status. The lane with the lowest comprehensive impact score is then selected as the target lane. The formula for calculating the comprehensive impact score is as follows: in, To comprehensively assess the impact score, The weight of the impact factor for large trucks, For the impact factor of large trucks, The weights of the factors influencing vehicle density within the lane. The influencing factor of vehicle density within the lane. The weight of the influence factor of average vehicle speed. The factor affecting the average speed of vehicles. The weights of the factors influencing the path matching degree with the destination are determined. The influencing factor of the route matching degree with the destination. The weights of the factors affecting traffic signal status. This refers to the factors affecting the status of traffic lights.

10. A lane selection optimization system for autonomous vehicles, characterized in that, include: The acquisition module is used to acquire image data, vehicle motion status data, road information data, and traffic flow data around the vehicle. The identification module is used to identify large trucks around the vehicle based on the collected image data and using deep learning image recognition algorithms, and to continuously track and analyze the movement status of the large trucks. The prediction module is used to predict the truck's trajectory in the future based on the truck's motion status, combined with data on the motion status of surrounding vehicles, road information, and traffic flow. The determination module is used to analyze the impact of selecting different lanes on the truck based on the truck's motion status and predicted driving trajectory in the future, and to determine the lane that will have the least impact on the vehicle as the target lane.

11. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 10.

12. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 10.

Images