An autonomous vehicle dynamic path planning system and method

Abstract

This invention relates to the field of autonomous vehicle management and provides a dynamic path planning system and method for autonomous vehicles. The global scheduling module allocates scheduled arrival time windows to vehicles and performs global shaping on traffic flowing into key nodes. The scheduling optimization module calculates a macroscopic path for each vehicle based on time window constraints. The regional collaborative control module, acting as a distributed node, has a collaborative maneuvering unit that locks the requesting vehicle and its associated vehicle set when a vehicle needs to change lanes or turn; the calculation module calculates the optimal vehicle speed, safe space window, and anti-disturbance compensation control quantity, and outputs multi-vehicle collaborative instructions; the cluster coordination unit initiates a consistency control algorithm for the upstream vehicle cluster to smooth out fluctuations when abnormal deceleration waves are detected. This invention achieves a transformation from passive single-vehicle pathfinding to proactive system collaborative scheduling, effectively solving the problems of vehicle convergence congestion, diversion conflicts, and phantom traffic jams in high-density scenarios, thus improving traffic efficiency and stability.

Description

Technical Field

[0001] This invention relates to the field of autonomous vehicle management and provides a dynamic path planning system and method for autonomous vehicles. Background Technology

[0002] In recent years, with the rapid development of autonomous driving technology, its application in many unmanned management scenarios has become increasingly mature. Especially in the logistics and transportation sector, unmanned flatbed trucks, with their high degree of automation and flexibility, have been widely used in large-scale warehouse management and material sorting and transfer within smart factories. These systems typically use automated equipment such as robotic arms to load goods onto flatbed trucks, which then autonomously navigate within or between factory buildings along routes planned in real-time by a preset or control system. This achieves fully unmanned and intelligent delivery and storage of goods from a centralized receiving point to various decentralized storage or processing areas. This unmanned warehouse-to-warehouse cargo transfer model significantly improves logistics efficiency and reduces labor costs, making it an important component of modern smart logistics systems.

[0003] Within a typical factory or large logistics hub, the logistics network usually consists of several main roads carrying the majority of traffic and connecting key nodes, such as centralized receiving areas and major warehouses in various distribution areas, as well as numerous branch roads connecting specific work points, such as sorting stations, shelving areas, and production lines, forming a complex transportation network. Existing dynamic path planning systems aim to centrally schedule and control fleets. Through global path planning algorithms, driving route tasks are assigned to each autonomous flatbed truck, enabling a large number of vehicles to operate in a coordinated manner within the system even when goods arrive in concentrated quantities and transportation demand surges.

[0004] However, existing dynamic route planning systems still exhibit significant scheduling bottlenecks and operational efficiency issues in practical applications, especially when large logistics transportation and distribution transfer stations face peak traffic periods such as the Spring Festival and Double Eleven shopping festivals. These issues are specifically reflected in the following aspects:

[0005] 1. Firstly, in the centralized cargo receiving area, when the system mobilizes a large number of driverless flatbed trucks to handle large volumes of incoming goods, it easily leads to concentrated vehicle arrivals on the main roads near the receiving point, causing vehicle congestion. Due to the lack of fine-grained control over the arrival rhythm and near-field spatial distribution, vehicles often queue on the main roads waiting to enter the loading positions, resulting in excessively high vehicle density on some sections of the main roads. This not only increases the risk of collisions between vehicles, but more importantly, it causes severe traffic jams, drastically reducing vehicle speed, decreasing overall throughput efficiency, and slowing down cargo dispersal, creating a vicious cycle.

[0006] 2. After being loaded, vehicles leave the centralized receiving area and proceed via the main road to different destination cargo holds or sorting areas. During this diversion process, vehicles often need to choose or change lanes on the main road according to their destination direction and turn into branch roads at appropriate intersections. Existing planning strategies typically focus on the optimal path for a single vehicle or simple conflict avoidance, making it difficult to efficiently coordinate the diversion operations of a large number of vehicles simultaneously in dense traffic flow. For example, vehicles needing to turn left into a branch road may be continuously blocked by the straight-going traffic on the main road, forcing them to wait for a long time to turn; while vehicles with different destination directions weaving and changing lanes will disrupt the smooth flow of traffic on the main road. This waiting and conflict caused by uncoordinated diversion planning will quickly spread to the rear, exacerbating congestion on the main road.

[0007] 3. When vehicles prepare to turn into a side road, they often need to cross lanes (e.g., from the outermost lane to the innermost lane to make a left turn). On main roads with high traffic volume, this lane-crossing behavior requires the target lane to provide sufficient gaps for insertion. Existing systems have limited ability to collaboratively create such gaps, causing turning vehicles to often wait passively. This not only increases the vehicle's own travel time, but the waiting behavior also temporarily occupies the current lane, hindering the flow of traffic in the same direction behind, becoming a new cause of congestion.

[0008] In the prior art:

[0009] The following technical solution is disclosed in Chinese invention patent application number 202411395689.2, published on December 10, 2024, entitled "Vehicle Automated Lane Change Control Method, Device, Equipment and Storage Medium":

[0010] When a vehicle needs to change lanes, the system assesses the safety conditions of the surrounding lane environment in real time and uses algorithms such as B-spline curves to plan a smooth lane-change trajectory. Precise lateral control guides the vehicle to complete the lane change. This solution primarily addresses the trajectory optimization and execution smoothness issues during lane changes, representing a technological advancement in vehicle local motion control.

[0011] However, in peak-hour scenarios at large logistics hubs, this solution fails to address the challenges of system-level scheduling. First, the technology cannot prevent or alleviate large-scale congestion near cargo receiving points because it only responds to single-vehicle lane-changing commands and does not involve global planning and control of upstream traffic flow departure rhythms and convergence paths. Second, in high-density traffic on main roads, the success of vehicle diversion or turning using this technology depends entirely on the availability of sufficient natural gaps in the target lane. Under saturated traffic conditions, the system will still passively wait due to unmet safety requirements, further increasing delays. This waiting behavior itself becomes a new bottleneck, exacerbating congestion further. Furthermore, for turns requiring continuous crossings of multiple lanes, this single-vehicle decision-making model faces a higher risk of failure when searching for continuous gaps.

[0012] The following technical solution is disclosed in Chinese invention patent application number 202411782257.7, published on March 7, 2025, entitled "Method and Device for Collision Risk Assessment Based on Lane and Surrounding Vehicle Motion Prediction":

[0013] The system predicts the future trajectories of surrounding vehicles, particularly inferring the probability distribution of their lane-changing intentions based on lateral speed, and combines this with road geometry information to generate multiple local candidate paths for the vehicle. It then quantitatively assesses the collision risk of each path. The ultimate goal is to provide a more accurate risk assessment for the vehicle's path selection, while simultaneously achieving hazard avoidance.

[0014] However, this technical solution also fails to address the congestion problem stemming from the lack of coordinated scheduling in multi-vehicle systems. First, the core function of this patent is "assessment" and "early warning," not "scheduling." The system cannot regulate the convergence flow of vehicles towards the cargo receiving point at its source, thus failing to prevent concentrated vehicle congestion on the main road at the receiving point. Second, in main road diversion scenarios, this solution can select the lane-changing path with the lowest collision risk for each vehicle at the current moment, but this is a passive and adaptive choice. When the overall traffic density is extremely high and the risk of all candidate paths remains high, vehicles will still be stuck waiting. The system cannot proactively coordinate vehicles in the target lane to create safe merging gaps, thus failing to fundamentally improve diversion efficiency or break the congestion chain caused by waiting.

[0015] In summary, existing dynamic path planning systems for autonomous vehicles have significant shortcomings when dealing with concentrated, high-density transportation tasks during peak periods at large logistics hubs. These shortcomings include traffic flow management at vehicle receiving points and minimizing the interference of vehicle lane-crossing and turning behaviors on the main traffic flow. These problems are interconnected and exacerbate each other, ultimately leading to decreased overall system efficiency, cargo backlog, and even increased safety risks, thus limiting the stability and reliability of autonomous logistics systems under extreme peak pressure.

[0016] Therefore, there is an urgent need for a more effective way to deal with the above scenarios and achieve dynamic, collaborative, and efficient route planning for vehicles. Summary of the Invention

[0017] To address the aforementioned shortcomings, the present invention aims to provide a dynamic path planning system and method for autonomous vehicles, thereby solving the problems mentioned in the background art. The system includes:

[0018] The global scheduling module is responsible for integrating information from the entire road network and all transportation tasks, and planning the flow of vehicles merging into key nodes;

[0019] The scheduling optimization module is responsible for calculating the specific macro-path and formulating the specific time plan for each vehicle.

[0020] The regional collaborative control module is a distributed computing node deployed at key nodes of the road network. The regional collaborative control module includes a collaborative maneuvering unit for identifying requesting vehicles and their associated vehicles, and a cluster coordination unit for real-time monitoring of the average speed of each local traffic flow segment within the managed road section. The collaborative maneuvering unit selects the target vehicle and, through subsequent solutions, creates a safe gap for the requesting vehicle in the target lane. When the cluster coordination unit detects an abnormal deceleration segment, it initiates a consistency control algorithm for the affected upstream vehicle cluster.

[0021] The regional collaborative control module includes a real-time collaborative computing module.

[0022] The real-time collaborative computing module receives vehicle "set" information, vehicle status, road geometry, and traffic flow information sent by the above-mentioned units, and sequentially solves for the optimal vehicle speed, safe space window, and compensation control quantity to suppress traffic wave propagation required for collaborative maneuvering, and finally outputs multi-vehicle collaborative motion commands.

[0023] Furthermore, the real-time collaborative computing module includes: an optimal vehicle speed planning unit, a safety window synthesis unit, and a disturbance suppression planning unit;

[0024] The optimal speed planning unit is responsible for calculating the optimal speed for the requesting vehicle when performing a turning or merging maneuver.

[0025] The safety window synthesis unit is responsible for calculating the minimum space clearance required for the requesting vehicle to safely enter the target lane, and determining the target clearance size that should be achieved through cooperative control;

[0026] The disturbance suppression planning unit, using a car-following model, is responsible for calculating the disturbances caused to the traffic flow behind by the execution of coordination commands;

[0027] The real-time collaborative computing module is activated upon receiving a turning or lane-changing request from the regional collaborative control module; it first calculates the optimal vehicle speed. Later based on Solving for the optimal safe passage window size Subsequently, the disturbance suppression planning unit assesses the downstream disturbances that the instruction may cause and calculates the compensation acceleration instruction.

[0028] Furthermore, the calculation process of the optimal vehicle speed planning unit is as follows:

[0029] Based on the radius of curvature of the target curve The radius of curvature R of the target turning path and the estimated value of the current road surface adhesion coefficient. Calculate the physical speed limit that will prevent the vehicle from skidding. ;

[0030] ;

[0031] Do not exceed the safe speed limit And not lower than the system's minimum allowable speed Given the premise, find the optimal vehicle speed :

[0032] ;

[0033] in, Calculate the average speed of the target lane The requested vehicle's current speed The weighted average; These are weighting coefficients; Values ​​range from 0.5 to 2.

[0034] Furthermore, the steady-state safety distance calculation of the safety window synthesis unit includes the following process: calculating the minimum safety distance. :

[0035] ;

[0036] in, It is the stationary safe distance set by the administrator according to the standard. It is the expected headway. It's the speed difference between the two cars. The two cars approach zero at the moment they ideally merge; and These are the settings for maximum acceleration and deceleration;

[0037] Calculate the optimal safe passage window size :

[0038] ;

[0039] in, This is the natural traffic flow interval; To coordinate the incremental distance generated by the acceleration of the vehicle in front and the deceleration of the vehicle behind; satisfy .

[0040] Furthermore, the car-following model applies to the affected set of vehicles. Each car in The compensated acceleration was calculated for all of them. ;

[0041] ;

[0042] in, It is a vehicle The current speed, It is the desired speed for this road segment set by the administrator in the system. It is a vehicle The communication neighbor set of This is the average speed of this road section; These are all control gains set by the administrator in the system;

[0043] Finally, the system processes the original collaborative instructions. With compensation acceleration Weighted fusion is performed to generate the optimal compensation acceleration command set that is finally issued to each associated vehicle. .

[0044] Furthermore, the global scheduling module includes:

[0045] The traffic condition perception unit is responsible for continuously receiving condition data from vehicles and roadside areas across the entire road network, constructing and updating the dynamic traffic density distribution map and global status of the entire road network in real time.

[0046] The scheduling and management unit is responsible for modeling key nodes such as the cargo receiving area as a queuing system with fixed service capabilities, and executing reservation scheduling based on the current service load, the set of vehicles in transit, and task priorities.

[0047] Furthermore, the scheduling optimization module specifically includes:

[0048] The route planning unit calculates the macro-level optimized route for each vehicle based on the origin and destination of the transportation task, the real-time road network status, and the reservation time window constraints issued by the reservation and scheduling management unit.

[0049] The time window constraint unit ensures that the estimated arrival time of each macro-path output by the path planning unit falls within the specified reservation time window by adjusting the segment weights in the path planning or directly setting the upper and lower bounds of the path travel time.

[0050] An autonomous vehicle dynamic path planning method, based on an autonomous vehicle dynamic path planning system, includes the following steps:

[0051] S1. The system receives global status data and performs global path planning at a macro level. The system continuously receives status data from all autonomous vehicles and roadside equipment, constructs and updates the dynamic traffic flow status map of the entire network in real time. For key service nodes such as cargo receiving areas, the system models them as queuing systems with fixed capacity. Then, it calculates the path from the origin to the destination for each vehicle.

[0052] S2. When a vehicle needs to perform maneuvers that may cause conflict, such as lane changing or turning, a local coordination request is triggered during the journey, which includes generating a coordination maneuver request. Subsequently, the system locks all associated vehicles participating in the coordination based on the location and intent of the requesting vehicle. At the same time, it obtains the status information of all vehicles in the vehicle set, as well as the geometric parameter information of the roads where all vehicles are located.

[0053] S3. Real-time collaborative calculation process of control commands: includes the following sequentially calculated sub-steps;

[0054] S3.1. Based on the data obtained in step S2.2, perform optimal speed planning; calculate the optimal safe speed for the requested vehicle to perform the maneuver. ;

[0055] S3.2 Determine the minimum space gap required for safe insertion through calculation. and the incremental gaps that need to be created ;

[0056] S3.3 The disturbance calculation unit evaluates and suppresses "ghost traffic jams" that may be caused by cooperative maneuvers; it outputs the optimal compensation acceleration command set to counteract the disturbance. ;

[0057] S4. The system issues a collaborative control command, requesting the vehicle to execute the command accordingly;

[0058] S4.1, The regional collaborative control module will use the data obtained in step S3.1... The compensation command obtained in step S3.3 to achieve The required original collaborative adjustment instructions are merged to generate a final collaborative control instruction set for all associated vehicles, which is then distributed to each vehicle.

[0059] S4.2 Each vehicle executes the instruction set.

[0060] Furthermore, the specific calculation process in step S3.1 is as follows:

[0061] Input: Average speed of the target lane Request the vehicle's current speed Curvature radius of curve Road surface adhesion coefficient ;

[0062] The calculation unit determines the safe speed limit based on vehicle dynamics. ;

[0063] Subsequently, within the set range Internal, solve :

[0064] .

[0065] Furthermore, the specific calculation process in step S3.2 is as follows:

[0066] Input: Optimal vehicle speed The set maximum acceleration deceleration Headway ;

[0067] The computing unit calculates the minimum safe distance required for steady-state car-following. ;

[0068] Subsequently, based on the current natural traffic flow gaps Determine the target window size that needs to be created collaboratively. ,Require Target increment .

[0069] Therefore, the beneficial effects of the present invention are as follows:

[0070] 1. Effectively solved the systemic congestion problem caused by vehicles arriving at cargo receiving points in an disorderly and concentrated manner.

[0071] Existing technologies only focus on the behavior of individual vehicles while they are in motion, lacking the ability to regulate the source of traffic flow convergence. This invention implements a "global traffic balancing" logic through a global scheduling module, assigning scheduled arrival time windows to vehicles heading to key nodes and using these windows as constraints for route planning. This transforms the vehicle arrival process from random and explosive to uniform and controlled, eliminating congestion and queue overflows on the main roads of receiving points at the source, ensuring continuous smooth flow and high throughput at the logistics entry point.

[0072] 2. It has changed the situation where vehicles passively wait when diverting or turning, and has achieved conflict-free collaborative passage under high-density traffic flow.

[0073] Existing technologies fail when traffic flow is saturated because they cannot find safe gaps, causing vehicles to wait and leading to congestion. This invention implements a "cooperative synthesis of local gaps" logic through a regional cooperative control module. When a vehicle needs to change lanes or turn, the system treats the requesting vehicle and associated vehicles as a single unit, calculates and issues cooperative speed adjustment commands through a real-time cooperative computing module, and proactively synthesizes a safe passage window in the target lane. This allows vehicles to smoothly complete maneuvers without waiting, fundamentally eliminating traffic flow interruptions and delays caused by waiting.

[0074] 3. It actively suppresses the generation and propagation of deceleration waves in traffic flow, effectively preventing "phantom traffic jams" and improving the overall stability of the system.

[0075] This invention introduces distributed speed consistency control logic as a continuously running background service. By monitoring road segment speeds through a cluster speed coordination unit, once an abnormal deceleration wave is detected, a consistency control algorithm is applied to the upstream vehicle cluster to synchronize vehicle speeds with the cluster's average speed and the target speed.

[0076] 4. Existing technologies are limited to optimizing individual vehicle routes or avoiding instantaneous risks. This invention, through a three-tiered "center-edge-vehicle" architecture, organically combines three major logics: global flow shaping, collaborative creation of local gaps, and active suppression of traffic waves, achieving unified and optimized scheduling of vehicle group movement. This enables the system to maximize overall throughput, significantly shorten and make more predictable the average vehicle travel time when dealing with high-density transportation tasks, while reducing accident risks through collaborative operations, thus comprehensively overcoming the bottlenecks of existing systems. Detailed Implementation

[0077] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0078] The purpose of this invention is to provide a dynamic path planning system and method for autonomous vehicles, mainly for collaborative path planning of autonomous vehicles in high-density, closed scenarios. Through the control system and dispatching method, dynamic, collaborative and efficient route planning of vehicles can be achieved.

[0079] In summary, this invention is based on a three-tiered collaborative control architecture of "center-edge-vehicle," enabling real-time perception of the traffic flow status across the entire network. Through unified analysis and calculation, the system actively regulates and transforms the traditional single-vehicle path planning problem into a multi-vehicle collaborative motion planning problem. Its core purpose is to solve three problems at large logistics transfer stations during peak hours: systemic congestion caused by disorderly vehicle convergence, diversion conflicts, and the propagation of motion disturbances.

[0080] It should be understood that this invention relies on two core control logics: "global flow balancing" and "cooperative synthesis of local gaps".

[0081] For "global traffic balancing", the system first shapes the traffic flow at key nodes where vehicles converge, such as cargo receiving areas, at the global level to avoid overload and congestion. Then, at the local level, it actively adjusts the movement status of relevant vehicles to dynamically create safe and efficient passage gaps for maneuvering behaviors that may cause conflicts, such as lane changes and turns. This ensures the overall traffic capacity of the road network and guarantees the efficient transportation of global traffic flow.

[0082] It is important to explain that the "global traffic balancing" logic aims to manage the vehicle arrival process. The system models key service nodes such as loading and unloading areas as queuing systems with a fixed service rate. By assigning a "scheduled arrival time window" to each vehicle planning to go to that node, and binding its departure time or route planning to the time window, the random and explosive arrival flow is shaped into a uniform and controlled arrival process, preventing queue overflow and main road congestion from the source.

[0083] Based on the above-mentioned "global traffic balance" control of traffic sources, vehicles will still face local conflicts when traveling in the road network. It is important to explain that the "cooperative synthesis of local gaps" logic is used to resolve these conflicts. When a vehicle needs to change lanes or turn, the system no longer passively waits for naturally occurring traffic gaps, but instead creates the required gaps. The problem of creating the required gaps is defined as a multi-vehicle cooperative optimal control problem with the objective of minimizing overall disturbance.

[0084] For "cooperative synthesis of local gaps," a unified motion model is established within the system, incorporating the requesting vehicle and its associated vehicles. This transforms the original problem into an optimization problem with safety constraints, resulting in a set of cooperative speed adjustment commands for multiple vehicles. These commands are issued synchronously and executed by multiple vehicles simultaneously, enabling the cooperative synthesis of a safe passage size window that accommodates the requesting vehicle's lane change at a predetermined time and location.

[0085] More specifically, during this process, the associated vehicle and the requesting vehicle are initially in different lanes. At this point, the two adjacent vehicle bodies within the traffic safety size window are defined as the "front associated vehicle body" and the "rear associated vehicle body," respectively. The gap between the front associated vehicle body and each of the several vehicle bodies in front of it is called the front associated gap; the same applies to the rear associated vehicle body.

[0086] For the rear-connected vehicle body, the rear-connected vehicle body and several vehicles extending backward will decelerate and reduce the forward-connected gap when creating local gaps. At this time, under the influence of the gap control of this system, the traffic flow is prone to spontaneously generated and backward-propagating deceleration fluctuations, i.e., "phantom traffic jams". In this case, the system applies "distributed speed consistency control" logic to suppress it.

[0087] For "distributed speed consistency control," the system monitors the average speed of vehicles on each road segment in real time. Once an abnormal deceleration wave is detected, a consistency control algorithm is applied to the vehicle cluster upstream of the affected road segment. This algorithm enhances the acceleration process of a group of vehicles in the latter part of the deceleration segment. During this process, the target speed of the target vehicle in this segment depends not only on the vehicle in front but also on the average speed of the cluster and the expected speed of the road segment. Thus, by injecting a damping-like effect into the system, speed fluctuations are actively smoothed out, and traffic flow stability is restored.

[0088] Based on the above overall logic, from the perspective of the final effect, this invention, through the synergistic effect of the three layers of control logic, achieves logical optimization from "a single vehicle finding its way within existing traffic flow" to "the system dynamically optimizing the traffic flow structure." Thus, under the unified scheduling of the system, efficient paths are opened, ensuring maximum overall system throughput and reliable travel time under high-density transportation tasks.

[0089] At the system level, this invention provides a dynamic path planning system for autonomous vehicles, including: a global scheduling module, a scheduling optimization module, a regional cooperative control module, a trajectory execution module, a vehicle-road cooperative communication network, and a high-precision fusion positioning module. The regional cooperative control module includes a real-time cooperative computation module. As the core algorithm for the two main logics of "cooperative synthesis of local gaps" and "distributed speed consistency control," the real-time cooperative computation module is responsible for completing the calculations required for local cooperative maneuvers within a short time.

[0090] The global scheduling module, serving as the system's central decision-making core, is deployed in the cloud or on a central server. This module is responsible for integrating information from the entire road network and all transportation tasks. Its core function is to implement the aforementioned "global traffic balancing" logic and plan the traffic flow of vehicles merging into key nodes.

[0091] To achieve this function, the global scheduling module specifically includes a traffic status perception unit and a scheduling management unit.

[0092] The traffic condition perception unit continuously receives condition data from vehicles and roadside areas across the entire road network, constructs and updates the dynamic traffic density distribution map and global condition table of the entire road network in real time, and provides real-time situational awareness for traffic control.

[0093] The scheduling management unit models key nodes such as the cargo receiving area as a queuing system with fixed service capacity. Based on the current service load, the set of vehicles en route, and task priorities, it performs reservation scheduling. Specifically, the scheduling process involves assigning a reservation arrival time window to each vehicle scheduled to arrive at such a node, ensuring that the expected number of arriving vehicles at any given time does not exceed the node's service capacity. This reshapes the traffic flow at the source, preventing overload congestion caused by disorderly convergence. The generated reservation scheduling table is then sent to the scheduling optimization module.

[0094] The scheduling optimization module receives the reservation schedule from the global scheduling module and, in conjunction with real-time traffic conditions, calculates the specific macro-path and formulates a specific time plan for each vehicle. The core objective of the scheduling optimization module is to optimize the overall driving efficiency of all vehicles on the network while satisfying global constraints such as reservation time windows.

[0095] To achieve this function, the scheduling optimization module specifically includes a path planning unit and a time window constraint unit.

[0096] The route planning unit calculates an optimized route for each vehicle at a macro level based on the origin and destination of the transportation task, the real-time road network status, and the reservation time window constraints issued by the reservation scheduling management unit. During the route planning process, on-time arrival at the reservation node is prioritized as a high-priority optimization objective.

[0097] The time window constraint unit ensures that the estimated arrival time of each macro-path output by the path planning unit falls within the specified reservation time window by adjusting the segment weights in the path planning or directly setting the upper and lower bounds of the path travel time. Finally, the scheduling optimization module issues macro-path instructions with spatiotemporal constraints to the relevant vehicles and the corresponding regional collaborative control modules.

[0098] The regional collaborative control module consists of distributed computing nodes deployed at key nodes of the road network. Key nodes include arterial road entrances, diversion zones, curves, and other changing nodes in the road network. Each distributed computing node is responsible for a geographical area, receiving macro-path instructions and short-term vehicle status within the area, and executing the aforementioned "collaborative synthesis of local gaps" and "distributed speed consistency control" logic.

[0099] To achieve this function, the regional collaborative control module specifically includes a collaborative maneuver unit and a cluster coordination unit.

[0100] The Cooperative Maneuver Management Unit is the core of the "Cooperative Synthesis Local Gap" logic. When a vehicle within its jurisdiction issues a lane change or turning maneuver request, this unit identifies the requesting vehicle and its associated vehicles, locks the vehicle codes, and uploads them as a "set." It's worth noting that associated vehicles include the preceding associated vehicle body, the following associated vehicle body, and the extended traffic flows in both directions; the number of extended traffic flows can be pre-specified by the system administrator.

[0101] The purpose of the cooperative maneuver management unit is to select a target vehicle (either actively or passively based on other system functions, such as selecting a vehicle in an accident lane due to an unforeseen event for cooperative lane changing) and create a safe gap for the requesting vehicle in the target lane. Through subsequent calculations, a set of cooperative control commands for all associated vehicles is generated (e.g., slight acceleration of the preceding associated vehicle and slight deceleration of the following associated vehicle), aiming to synthesize a passage window conforming to safe dimensions at a predetermined time and space point with minimal overall disturbance. These commands are synchronously issued through the vehicle-to-infrastructure (V2I) communication network.

[0102] The cluster coordination unit is responsible for monitoring the average speed of each local traffic flow segment within its jurisdiction in real time, based on the "distributed speed consistency control" logic. Once an abnormal deceleration segment is detected due to the synthesis of coordination gaps or other reasons, the consistency control algorithm is initiated for the affected upstream vehicle cluster. The cluster coordination unit's logic is similar to that of the collaborative mobility management unit, which selects target vehicles to provide a sample set for subsequent calculations, actively smooths out speed fluctuations across the entire segment, and prevents the propagation of "phantom traffic jams."

[0103] Understandably, at dense traffic convergence points, the number of lane change tasks to be processed is enormous. Therefore, the tasks generated by the collaborative mobility management unit and the cluster coordination unit are distributed and processed by the real-time collaborative computing module in the distributed computing nodes of multiple key road network nodes in the regional collaborative control module.

[0104] The real-time collaborative computing module receives vehicle "set" information, vehicle status, road geometry, and traffic flow information from the aforementioned units. Through the following calculation process, it sequentially solves for the optimal vehicle speed, safe space window, and compensation control quantities to suppress traffic wave propagation required for collaborative maneuvering. Finally, it outputs a set of precise multi-vehicle collaborative motion commands to ensure a safe, smooth maneuvering process with minimal interference to the main traffic flow. The real-time collaborative computing module consists of three core units: an optimal vehicle speed planning unit, a safe window synthesis unit, and a disturbance suppression planning unit.

[0105] The optimal speed planning unit is responsible for calculating the optimal speed for the requesting vehicle when performing a turning or merging maneuver. This speed needs to be balanced between the vehicle's dynamic limits and smooth integration with the target traffic flow.

[0106] Input: The current state of the requested vehicle, the radius of curvature R of the target turning path, the estimated value of the road adhesion coefficient μ, and the velocity vector of the associated vehicle. The calculation of the safe speed limit includes the following steps:

[0107] Based on the radius of curvature of the target curve and the current estimated value of road surface adhesion coefficient Calculate the physical speed limit that will prevent the vehicle from skidding. .

[0108]

[0109] Where g is the acceleration due to gravity; R is the radius of curvature of the target curve; and μ is the road surface adhesion coefficient. The calculation of the safe speed limit ensures that the required centripetal force does not exceed the maximum static friction force that the road surface can provide.

[0110] Within physical limits, that is, within the maximum safe speed limit. And not lower than the system's minimum allowable speed Given the set values, find the optimal vehicle speed. To achieve a smooth inflow.

[0111]

[0112] in, Calculate the average speed of the target lane The requested vehicle's current speed The weighted average.

[0113] These are weighting coefficients. From the original balance coefficient With time step Decide; The larger the value, the more the calculation results tend to maintain the current speed. The more gradual the change; The smaller the value, the more the calculation result is biased towards matching the target traffic flow speed. .

[0114] Among them, the function Ensure that the calculation results are not lower than the minimum operating speed specified by the system. If the weighted average result is less than Then the value is .function Ensure the final result does not exceed the limits based on physical conditions (radius of curvature of the curve). Road surface adhesion coefficient The calculated safe speed limit If the value after lower bound constraint is greater than The final value is .

[0115] In practical applications, Choose a moderate positive value (e.g., 0.5~2) to achieve a balance between the two objectives.

[0116] The safety window synthesis unit, also known as the safety window calculation unit, is responsible for calculating the minimum space clearance required for a requesting vehicle to safely enter the target lane; simultaneously, it determines the target clearance size to be achieved through coordinated control. The safety window calculation unit comprehensively considers both the safe following distance and the additional space created by active coordination when calculating the steady-state safe distance. The steady-state safe distance calculation includes the following processes:

[0117] Calculate the minimum safe distance :

[0118]

[0119] in, It is the stationary safe distance set by the administrator according to the standard. It is the expected headway. It is the speed difference between the two vehicles (which approaches zero at the moment of ideal merging). and These represent the maximum acceleration and deceleration, respectively. This formula expresses the minimum space required to prevent a rear-end collision under constant speed and stable car-following conditions.

[0120] Calculate the optimal safe passage window size :

[0121]

[0122] in, This is the natural traffic flow interval; To coordinate the incremental distance generated by the acceleration of the vehicle in front and the deceleration of the vehicle behind.

[0123] The system operates during natural traffic flow intervals. The optimal safe passage window is calculated based on the following: Subsequently, the system coordinated adjustments This makes the above formula satisfy This minimizes the total adjustment range for related vehicles.

[0124] The disturbance suppression planning unit, also known as the disturbance calculation unit, is responsible for calculating the disturbances that may be caused to the traffic flow behind due to the execution of coordination commands (in this case, deceleration commands); that is, to solve the problem of "phantom traffic jams" caused by the adjustment distance of vehicles associated with the previous calculation.

[0125] The disturbance calculation unit simulates the propagation of the deceleration behavior in the convoy behind the vehicle after the related vehicle executes a deceleration command through the disturbance prediction model.

[0126] The disturbance prediction model employs the existing "car-following model," which, simply put, is used to simulate the deceleration command executed by the following related vehicles. Subsequently, this deceleration behavior propagated within the following convoy. Predicted time... Subsequently, the disturbance will affect the set of vehicles. and the expected speed reduction of each vehicle For the affected vehicle set Each car in The compensated acceleration was calculated for all of them. To smooth the velocity gradient and suppress disturbance amplification.

[0127] Calculate the compensated acceleration :

[0128]

[0129] in, It is a vehicle The current speed, It is the desired speed for this road segment set by the administrator in the system. It is a vehicle The communication neighbors (i.e., the front and rear vehicles). This is the average speed of this road section. These are all control gains set by the administrator in the system. The first option drives the vehicle back to the desired speed; the second option makes the speeds of neighboring vehicles converge (diffusion damping); the third option makes the individual vehicle speed consistent with the overall average speed (global damping).

[0130] In an ideal state, we have:

[0131]

[0132]

[0133] Finally, the system processes the original collaborative instructions. With compensation acceleration Weighted fusion is performed to generate the optimal compensation acceleration command set that is finally issued to each associated vehicle. This maintains the overall stability and high efficiency of traffic flow on the main line.

[0134] Therefore, the regional collaborative control module coordinates road vehicles through a fixed collaborative workflow.

[0135] The collaborative workflow of the modules is as follows: When the regional collaborative control module establishes a turning or lane-changing request, the real-time collaborative calculation module is activated. The optimal speed planning unit first calculates the optimal speed. ; based on Sure These two outputs serve as key parameters input to the Cooperative Maneuver Management Unit, which generates the initial speed adjustment command to create the window. Subsequently, the Disturbance Suppression Planning Unit assesses the downstream disturbances that this command may trigger and calculates the compensation acceleration command. Finally, all commands are coordinated and synchronously sent to the trajectory execution modules of the relevant vehicles.

[0136] It is worth noting that, based on the above, the real-time collaborative computing module upgrades the system from control that relies on rules and experience thresholds to precise control based on model prediction and optimization calculation. This is the key technical means that distinguishes this technical solution from existing technologies and enables efficient, smooth, and congestion-free collaboration.

[0137] The trajectory execution module is deployed on each autonomous vehicle and is responsible for receiving and executing various instructions from upper-level modules. It receives, parses, and coordinates various path instructions from the scheduling optimization module, as well as cooperative control instructions from the cooperative maneuver management unit and speed coordination instructions from the cluster speed coordination unit.

[0138] When executing vehicle motion control commands, each vehicle combines its own sensor information with its own drive equipment, braking and steering systems to achieve high-precision control of the vehicle trajectory, ensuring that macro-path tracking, micro-coordinated maneuvering and speed coordination can be executed accurately and smoothly.

[0139] The vehicle-to-infrastructure (V2I) communication network serves as the data transmission hub connecting all the aforementioned modules. Based on C-V2X or similar low-latency, high-reliability communication technologies, the network ensures millisecond-level state sharing, intent transmission, and control command synchronization between vehicles (V2V), vehicles and roadside (V2I), and roadside and center (I2C). This communication network is existing technology and will not be elaborated upon here.

[0140] The high-precision positioning module provides the system with a unified and accurate spatiotemporal reference. This module integrates GNSS, roadside-assisted positioning, vehicle-mounted inertial navigation, and environmental feature matching technologies to provide each autonomous vehicle with high-precision real-time pose information. This information is fundamental for traffic condition perception, cooperative motion planning, and precise trajectory tracking.

[0141] Therefore, based on the collaboration of all the aforementioned modules, this system constitutes a complete and closed-loop processing system. The global scheduling module and the scheduling optimization module work together to avoid systemic congestion at the source and along the path; the regional collaborative control module resolves conflicts and smooths out disturbances locally; and the trajectory execution module ensures the accurate implementation of various optimization commands. Through this collaborative architecture, a fundamental shift can be achieved from traditional independent pathfinding by individual vehicles to proactive system optimization and coordinated movement of vehicle groups, ultimately ensuring efficient, stable, and safe operation in high-density, closed scenarios at the system level.

[0142] Based on an autonomous vehicle dynamic path planning system, the present invention provides an autonomous vehicle dynamic path planning method that is specifically implemented according to the following steps:

[0143] S1. The system receives global status data and performs global path planning at a macro level.

[0144] This step is executed collaboratively by the global scheduling module and the scheduling optimization module, aiming to reshape traffic at the source and avoid systemic congestion. Specifically, it includes the following sub-steps:

[0145] S1.1 The system's traffic state perception unit continuously receives state data from all autonomous vehicles and roadside equipment, including vehicle location, speed, destination, and road occupancy, and constructs and updates the dynamic traffic flow state map of the entire network in real time.

[0146] S1.2 For key service nodes such as cargo receiving areas, the scheduling management unit models them as queuing systems with fixed service capacity. Based on the current node load, the set of vehicles en route, and task priorities, a reservation scheduling algorithm is run to assign a reservation arrival time window to each vehicle scheduled to travel to the node, ensuring that the expected arrival rate is consistently lower than the node service rate, thereby achieving traffic flow shaping.

[0147] S1.3 The path planning unit calculates the path from the starting point to the destination for each vehicle.

[0148] S2. When a vehicle needs to perform maneuvers that may cause conflict, such as lane changes or turns, a local coordination request is triggered. This step is handled by the coordinated maneuver unit. Specifically, it includes the following sub-steps:

[0149] S2.1 Cooperative Maneuver Request Generation. Based on its macro-path instructions, when a vehicle approaches a preset location where a lane change is required (or is passively designated by the system based on abnormal situations), it sends a cooperative maneuver request to the cooperative maneuver unit in its area through the vehicle-road cooperative communication network. The request includes information such as the maneuver type and the target lane.

[0150] S2.2 The system locates all associated vehicles participating in the coordination based on the location and intent of the requesting vehicle. All associated vehicles specifically include the requesting vehicle, the preceding and following associated vehicles in the target lane, and several affected vehicles extending in the directions before and after it according to the system strategy. Simultaneously, the system acquires the real-time status of all vehicles in the vehicle group, including speed, position, and acceleration, as well as geometric parameters such as the radius of curvature R of the road curves where all vehicles are located.

[0151] S3. Real-time collaborative calculation process of control commands:

[0152] This step is executed by the real-time collaborative computing module within the regional collaborative control module, and includes three sequential calculation sub-steps.

[0153] S3.1. Based on the data obtained in step S2.2, perform optimal vehicle speed planning. Calculate the optimal safe speed for the requested vehicle to perform the maneuver. .

[0154] Input: Average speed of the target lane Request the vehicle's current speed Curvature radius of curve Road surface adhesion coefficient .

[0155] The calculation unit determines the safe speed limit based on vehicle dynamics. .

[0156] Subsequently, within the set range Internally, the solution is obtained through a weighted optimization formula. :

[0157] ;

[0158] S3.2 Determine the minimum space gap required for safe insertion through calculation.

[0159] Input: Optimal vehicle speed Vehicle dynamics parameters (set maximum acceleration) deceleration Headway ).

[0160] The computing unit calculates the minimum safe distance required for steady-state car-following. (This is a simplified representation, ignoring the speed difference term).

[0161] Subsequently, based on the current natural traffic flow gaps Determine the target window size that needs to be created collaboratively. ,Require Target increment .

[0162] The computing unit outputs the target collaborative window size. and the incremental gaps that need to be created This is a space target for coordinated control.

[0163] S3.3 The disturbance calculation unit assesses and suppresses "ghost traffic jams" that may be caused by cooperative maneuvers.

[0164] The system inputs the current state of the associated vehicle set, and step S3.2 is for creating... The generated initial coordination command (which is subsequently associated with the vehicle body needing to decelerate) The system uses a car-following model to simulate the initial deceleration command. Propagation within the following traffic flow, predicting the set of affected vehicles. and its speed reduction .

[0165] The calculation unit performs compensation control calculations and... Each vehicle Calculate the compensation acceleration Smooth the velocity gradient:

[0166] ;

[0167] in, The desired speed for the road segment, The average speed of the current road segment. To control the gain.

[0168] The final system output is the optimal compensation acceleration command set used to counteract disturbances. .

[0169] S4. The system issues a collaborative control command, requesting the vehicle to execute the command accordingly;

[0170] S4.1, Command Fusion and Issuance. The regional collaborative control module will issue the commands obtained in step S3.1. The compensation command obtained in step S3.3 to achieve The required initial coordination adjustment commands are merged to generate a final coordination control command set for all associated vehicles. This command set is then distributed to each vehicle in a time-synchronized manner via the vehicle-to-infrastructure (V2I) communication network.

[0171] S4.2 Execution process of the instruction set by each vehicle. After receiving the instruction, the trajectory execution module of each vehicle parses it, and the instruction coordination unit then controls the drive, braking, and steering systems to precisely track and execute the assigned speed or acceleration instructions, thereby collaboratively synthesizing a safety window at a preset time point. This allows the requested vehicle to complete the maneuver smoothly.

[0172] S5. Distributed speed consistency control. The system employs active damping control to address the inherent instability of traffic flow, executed by the cluster speed coordination unit.

[0173] S5.1 Abnormal deceleration wave detection. The system monitors the average speed of each road segment in real time. When a certain road segment is detected When the rate of descent exceeds the threshold within a short period of time and the vehicle density is high, it is determined to be the starting point of an abnormal deceleration wave ("ghost traffic jam").

[0174] S5.2, Consistency Control Command Generation. The system takes the affected vehicle cluster upstream of the deceleration wave as the control object and applies the distributed consistency control algorithm in S3.3 to calculate a set of coordinated acceleration commands aimed at improving the average speed of the cluster and smoothing the speed gradient.

[0175] S5.3 The generated coordination commands are uniformly distributed to vehicles within the cluster for execution. By correlating vehicle speeds with the cluster's average speed and desired speed, a "damping" effect is created, actively smoothing out deceleration waves and restoring smooth traffic flow to the road segment. This process is independent of specific coordinated maneuvers and continues to run as a background service to maintain global flow stability.

[0176] Through the connection and iteration of steps S1 to S5 above, this invention achieves closed-loop management from global traffic predictive shaping to proactive resolution of local conflicts and traffic fluctuation suppression control, ultimately achieving efficient, stable and safe dynamic path planning under high-density traffic flow at the system level.

[0177] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0178] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0179] This application also provides a storage medium storing an autonomous vehicle dynamic path planning program, which, when executed by a processor, implements the steps of the autonomous vehicle dynamic path planning method described in any one of the above descriptions.

[0180] The storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0181] The aforementioned storage medium may be included in the dynamic path planning device for autonomous vehicles; or it may exist independently and not be installed in the dynamic path planning device for autonomous vehicles.

[0182] The aforementioned computer-readable storage medium carries one or more programs that, when executed by an autonomous vehicle dynamic path planning device, cause the autonomous vehicle dynamic path planning device to: acquire lane information in response to an automatic lane change command; plan a lane change path based on the lane information; and control the vehicle to complete the automatic lane change based on the lane change path and the vehicle's real-time position.

[0183] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can 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 remote computers, the remote computer can 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 can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0184] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0185] The storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described dynamic path planning method for autonomous vehicles, thereby solving the technical problems raised in the background art. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the dynamic path planning method for autonomous vehicles provided in the above embodiments, and will not be repeated here.

[0186] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the content of this application specification, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A dynamic path planning system for autonomous vehicles, characterized in that, include: The global scheduling module is responsible for integrating information from the entire road network and all transportation tasks, and planning the flow of vehicles merging into key nodes; The scheduling optimization module is responsible for calculating the specific macro-path and formulating the specific time plan for each vehicle. The regional collaborative control module is a distributed computing node deployed at key nodes of the road network. The regional collaborative control module includes a collaborative maneuvering unit for identifying requesting vehicles and their associated vehicles, and a cluster coordination unit for real-time monitoring of the average speed of each local traffic flow segment within the managed road section. The collaborative maneuvering unit selects the target vehicle and, through subsequent solutions, creates a safe gap for the requesting vehicle in the target lane. When the cluster coordination unit detects an abnormal deceleration segment, it initiates a consistency control algorithm for the affected upstream vehicle cluster. The regional collaborative control module includes a real-time collaborative computing module. The real-time collaborative computing module receives vehicle "set" information, vehicle status, road geometry and traffic flow information sent by the collaborative maneuvering unit, and sequentially solves for the optimal vehicle speed, safe space window and compensation control quantity to suppress traffic wave propagation required for collaborative maneuvering, and finally outputs multi-vehicle collaborative motion command. The calculation process for the optimal vehicle speed is as follows: Based on the radius of curvature R of the target turning path and the estimated value of the current road surface adhesion coefficient. Calculate the physical speed limit that will prevent the vehicle from skidding. ; ； Do not exceed the safe speed limit And not lower than the system's minimum allowable speed Given the premise, find the optimal vehicle speed. : ； in, Calculate the average speed of the target lane The requested vehicle's current speed The weighted average; These are weighting coefficients; Values ​​range from 0.5 to 2; Steady-state safety distance calculation includes the following process: calculating the minimum safety distance. : ； in, It is the stationary safe distance set by the administrator according to the standard. The following vehicle is at the optimal speed. The time it takes for the vehicle to reach the position of the front of the vehicle in front; It's the speed difference between the two vehicles; To achieve the optimal vehicle speed; and These are the settings for maximum acceleration and deceleration, respectively. Calculate the optimal safe passage window size : ； in, This is the natural traffic flow interval; To coordinate the incremental distance generated by the acceleration of the vehicle in front and the deceleration of the vehicle behind; satisfy ; For the affected vehicle set Each car in The compensated acceleration was calculated for all of them. ; ； in, It is a vehicle The current speed, It is the desired speed for this road segment set by the administrator in the system. It is a vehicle The communication neighbor set of This is the average speed of this road section; These are all control gains set by the administrator in the system; Finally, the system processes the original collaborative instructions. With compensation acceleration Weighted fusion is performed to generate the optimal compensation acceleration command set that is finally issued to each associated vehicle. .

2. The dynamic path planning system for autonomous vehicles according to claim 1, characterized in that, The real-time collaborative computing module includes: an optimal vehicle speed planning unit, a safety window synthesis unit, and a disturbance suppression planning unit; The optimal speed planning unit is responsible for calculating the optimal speed for the requesting vehicle when performing a turning or merging maneuver. The safety window synthesis unit is responsible for calculating the minimum space clearance required for the requesting vehicle to safely enter the target lane, and determining the target clearance size that should be achieved through cooperative control; The disturbance suppression planning unit, using a car-following model, is responsible for calculating the disturbances caused to the traffic flow behind by the execution of coordination commands; The real-time collaborative computing module is activated upon receiving a turning or lane-changing request from the regional collaborative control module; it first calculates the optimal vehicle speed. Later based on Solving for the optimal safe passage window size Subsequently, the disturbance suppression planning unit assesses the downstream disturbances that the instruction may cause and calculates the compensation acceleration instruction.

3. The dynamic path planning system for autonomous vehicles according to claim 1, characterized in that, The global scheduling module includes: The traffic condition perception unit is responsible for continuously receiving condition data from vehicles and roadside areas across the entire road network, constructing and updating the dynamic traffic density distribution map and global status of the entire road network in real time. The scheduling management unit is responsible for modeling key nodes as queuing systems with fixed service capabilities, and executing reservation scheduling based on the current service load, the set of vehicles in transit, and task priorities.

4. The dynamic path planning system for autonomous vehicles according to claim 1, characterized in that, The scheduling optimization module specifically includes: The route planning unit calculates the macro-level optimized route for each vehicle based on the origin and destination of the transportation task, the real-time road network status, and the reservation time window constraints issued by the reservation and scheduling management unit. The time window constraint unit ensures that the estimated arrival time of each macro-path output by the path planning unit falls within the specified reservation time window by adjusting the segment weights in the path planning or directly setting the upper and lower bounds of the path travel time.

5. A dynamic path planning method for autonomous vehicles, characterized in that, The autonomous vehicle dynamic path planning system according to any one of claims 1-4 includes the following steps: S1. The system receives global status data and performs global path planning at a macro level. The system continuously receives status data from all autonomous vehicles and roadside equipment, constructs and updates the dynamic traffic flow status map of the entire network in real time. For key service nodes, the system models them as queuing systems with fixed capacity. Then, it calculates the path from the origin to the destination for each vehicle. S2. When a vehicle needs to perform a maneuver that may cause a conflict, a local coordination request is triggered during the journey, which includes the generation of a coordination maneuver request. Subsequently, the system locks down all associated vehicles participating in the coordination based on the location and intent of the requesting vehicle. At the same time, it obtains the status information of all vehicles in the vehicle set, as well as the geometric parameter information of the roads where all vehicles are located. S3. Real-time collaborative calculation process of control commands: includes the following sequentially calculated sub-steps; S3.

1. Based on the acquired data, perform optimal speed planning; calculate the optimal safe speed for the requested vehicle to perform the maneuver. ; S3.2 Determine the minimum space gap required for safe insertion through calculation. and the incremental gaps that need to be created ; S3.3 The disturbance calculation unit evaluates and suppresses "ghost traffic jams" that may be caused by cooperative maneuvers; it outputs the optimal compensation acceleration command set to counteract the disturbance. ; S4. The system issues a collaborative control command, requesting the vehicle to execute the command accordingly; S4.1, The regional collaborative control module will use the data obtained in step S3.1... The compensation command obtained in step S3.3 to achieve The required original collaborative adjustment instructions are merged to generate a final collaborative control instruction set for all associated vehicles, which is then distributed to each vehicle. S4.2 Each vehicle executes the instruction set.

6. The dynamic path planning method for autonomous vehicles according to claim 5, characterized in that, The specific calculation process in step S3.1 is as follows: Input: Average speed of the target lane Request the vehicle's current speed Curvature radius of curve Road surface adhesion coefficient ; The calculation unit determines the upper limit of safe speed based on vehicle dynamics. ; Subsequently, within the set range Internal, solve : 。 7. The dynamic path planning method for autonomous vehicles according to claim 5, characterized in that, The specific calculation process in step S3.2 is as follows: Input: Optimal vehicle speed The set maximum acceleration deceleration The following vehicle is at the optimal speed Time when it reaches the position of the front of the car in front ; The computing unit calculates the minimum safe distance required for steady-state car-following. ; Subsequently, based on the current natural traffic flow gaps Determine the target window size that needs to be created collaboratively. ,Require Target increment .

Images