Multi-link redundancy guarantee method for intelligent driving vehicle in extreme scene of highway

Abstract

The invention relates to the technical field of intelligent traffic and automatic driving, and discloses a multi-link redundancy guarantee method for an intelligent driving vehicle in a highway extreme scene, which comprises the following steps: acquiring vehicle-mounted sensor data, roadside equipment data and cloud traffic data corresponding to a preset highway section, constructing a global traffic situation map according to the vehicle-mounted sensor data, the roadside equipment data and the cloud traffic data; according to the global traffic situation map, evaluating the risk level of each highway section, and determining the takeover control strategy of the intelligent driving vehicle in each highway section according to the risk level; under the condition that the takeover control strategy is a formation cooperation strategy, generating a formation cooperation instruction of intelligent driving vehicles in a corresponding highway section, and sinking the formation cooperation instruction to a roadside edge node; and sending the formation cooperation instruction to the intelligent driving vehicle based on a multi-link redundancy architecture.

Description

Technical Field

[0001] This application relates to the technical fields of intelligent transportation and autonomous driving, and particularly to a multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways. Background Art

[0002] With the rapid development of intelligent transportation systems and autonomous driving technologies, highways, as an important carrier for inter-city transportation, have put forward higher requirements for intelligent management and the safe and efficient operation of driverless autonomous vehicles. Especially in extreme or complex traffic scenarios such as road closures due to construction, accident handling, and traffic congestion, how to achieve remote takeover and collaborative scheduling of intelligent driving vehicles has become a key issue in improving traffic efficiency and ensuring road safety. For this reason, multi-source information fusion and control strategies that integrate in-vehicle perception, roadside collaboration, and cloud scheduling have gradually become a research hotspot.

[0003] However, there are still many deficiencies in the existing technologies when dealing with extreme scenarios on highways. First, the traditional single-vehicle intelligent perception method is limited by the line of sight and environmental interference, and it is prone to perception blind spots in complex weather or sudden situations. Second, some collaborative control methods based on V2X (vehicle-to-everything) rely on a single communication link and lack a redundancy guarantee mechanism in the face of network fluctuations or interruptions, affecting the reliability of command transmission. In addition, most current takeover strategies adopt a unified mode and lack the ability to dynamically adjust according to different risk-level road sections, making it difficult to meet the hierarchical response requirements in complex traffic situations. Summary of the Invention

[0004] In view of this, an embodiment of this application provides a multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways.

[0005] According to one aspect of this application, a multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways is provided, including: Obtain in-vehicle sensor data, roadside device data, and cloud traffic data corresponding to a preset highway section, and construct a global traffic situation map according to the in-vehicle sensor data, the roadside device data, and the cloud traffic data; Evaluate the risk level of each highway section according to the global traffic situation map, and determine the takeover control strategy of intelligent driving vehicles in each highway section according to the risk level; When the takeover control strategy is a formation collaboration strategy, generate a formation collaboration instruction for intelligent driving vehicles in the corresponding highway section, and sink the formation collaboration instruction to a roadside edge node; Send the formation collaboration instruction to the intelligent driving vehicle based on a multi-link redundancy architecture.

[0006] According to another aspect of the present application, there is provided a storage medium having a computer program stored thereon, and when the program is executed by a processor, it implements the above-mentioned multi-link redundancy guarantee method for intelligent driving vehicles in extreme highway scenarios.

[0007] According to yet another aspect of the present application, there is provided a computer device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor. When the processor executes the program, it implements the above-mentioned multi-link redundancy guarantee method for intelligent driving vehicles in extreme highway scenarios.

[0008] By means of the above technical solutions, the multi-link redundancy guarantee method for intelligent driving vehicles in extreme highway scenarios provided by the embodiments of the present application constructs a global traffic situation map through multi-source data fusion, dynamically adjusts the takeover strategy based on risk grading, realizes local processing of instructions by using roadside edge computing, and uses multi-link redundant communication to ensure reliable transmission of instructions. The beneficial effects are as follows: It transforms the traditional "passive response" into "active collaboration". Through global perception, hierarchical decision-making, and redundant execution, it not only solves the perception and decision-making limitations of single-vehicle intelligence in complex scenarios, but also overcomes the reliability bottleneck of a single communication link, improves the traffic control efficiency in extreme scenarios, reduces the risk of secondary accidents (by reducing human misoperations through formation collaboration), and enhances the fault tolerance of the system to network anomalies and equipment failures, ultimately achieving a double improvement in highway traffic efficiency and safety.

[0009] The above description is only an overview of the technical solutions of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of the present application more obvious and understandable, the specific embodiments of the present application are specifically given below. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The schematic embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings: Figure 1 A flowchart showing a multi-link redundancy guarantee method for intelligent driving vehicles in extreme highway scenarios provided by an embodiment of the present application is shown; Figure 2 A flowchart showing another multi-link redundancy guarantee method for intelligent driving vehicles in extreme highway scenarios provided by an embodiment of the present application is shown; Figure 3 A schematic diagram showing the communication relationship among the cloud, roadside edge nodes, and intelligent driving vehicles provided by an embodiment of the present application is shown. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0011] The present application will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments may be combined with each other.

[0012] In this embodiment, a multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways is provided. As Figure 1 shown, the method includes: Step 101: Obtain in-vehicle sensor data, roadside device data, and cloud traffic data corresponding to a preset highway section, and construct a global traffic situation map according to the in-vehicle sensor data, the roadside device data, and the cloud traffic data.

[0013] Step 102: Evaluate the risk levels of each highway section according to the global traffic situation map, and determine the takeover control strategies for intelligent driving vehicles within each highway section according to the risk levels.

[0014] Step 103: When the takeover control strategy is a formation cooperation strategy, generate a formation cooperation instruction for intelligent driving vehicles within the corresponding highway section, and sink the formation cooperation instruction to the roadside edge node.

[0015] Step 104: Based on the multi-link redundancy architecture, send the formation cooperation instruction to the intelligent driving vehicle.

[0016] In the embodiment of the present application, first, by integrating in-vehicle sensors (such as cameras, radars, lidars) of intelligent driving vehicles (driverless autonomous vehicles), roadside devices (such as roadside cameras, meteorological monitoring stations), and cloud traffic data (historical congestion rules, construction information, accident databases), a global traffic situation map covering the entire highway section is constructed. For example, when the visibility is reduced due to heavy rain on a certain section, the in-vehicle sensors can capture real-time rainfall and road surface slipperiness data, the roadside devices provide the visibility monitoring value of the area, and the cloud supplements the accident rate data under the same type of historical weather. The three are combined to form a comprehensive situation map including environmental risks, vehicle states, and road conditions. Thus, it breaks through the line-of-sight and computing power limitations of traditional single-vehicle perception, eliminates perception blind spots (such as data missing in occluded areas or bad weather) through multi-source data complementarity, provides a global perspective for subsequent decision-making, and avoids misjudgments caused by local information.

[0017] Secondly, based on the global traffic situation map, the risk levels of each highway section are evaluated, and corresponding takeover control strategies are formulated accordingly, so as to flexibly respond to challenges under different road conditions, ensure more cautious and effective control measures in high-risk areas, and improve the safety and adaptability of autonomous driving vehicles. It realizes the transformation from "unified takeover" to "graded response", avoids resource waste (such as no need for excessive intervention in low-risk sections), and at the same time improves the control accuracy in high-risk scenarios, balancing the traffic efficiency and safety requirements.

[0018] Next, specifically for the situation that requires formation coordination, the embodiments of the present application design specific formation coordination instructions, and send these instructions to the roadside edge nodes for localization processing through the roadside edge computing nodes. Utilizing the advantages of edge computing, the latency is reduced and the response speed is improved, enabling intelligent driving vehicles to complete necessary adjustment and coordination actions in a short time, and enhancing the flexibility and coordination of the entire system.

[0019] Finally, a multi-link redundancy architecture with parallel transmission of three links, namely 5G cellular network, C-V2X direct communication, and dedicated short-range communication (DSRC), is adopted. Each link carries the same instructions but isolates interference through different frequency bands / protocols. For example, when the 5G signal is blocked in a tunnel, the system automatically switches to the C-V2X or DSRC link; if the latency of a certain link exceeds the threshold due to congestion, the system selects the instruction that arrives first through a multi-link voting mechanism for execution. The vulnerability problem of a single communication link is solved, and the reachability of instructions is ensured through link redundancy and dynamic switching, avoiding the risk of vehicle out-of-control caused by communication interruption.

[0020] By applying the technical solution of this embodiment, a global traffic situation map is constructed through multi-source data fusion, the takeover strategy is dynamically adjusted based on risk grading, the instruction localization processing is realized by using roadside edge computing, and multi-link redundant communication is adopted to ensure the reliable transmission of instructions. The embodiments of the present application transform the traditional "passive response" into "active coordination". Through global perception, hierarchical decision-making, and redundant execution, it not only solves the perception and decision-making limitations of single-vehicle intelligence in complex scenarios, but also overcomes the reliability bottleneck of a single communication link, improves the traffic control efficiency in extreme scenarios, reduces the risk of secondary accidents (by reducing human misoperations through formation coordination), and enhances the fault tolerance of the system to network anomalies and equipment failures, ultimately achieving a double improvement in highway traffic efficiency and safety.

[0021] Furthermore, as a refinement and extension of the specific implementation manner of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, another method for multi-link redundancy guarantee of intelligent driving vehicles in extreme scenarios on highways is provided, as Figure 2 shown. This method includes: Step 201: Obtain in-vehicle sensor data, roadside device data, and cloud traffic data corresponding to a preset highway section; perform spatio-temporal alignment on the in-vehicle sensor data, the roadside device data, and the cloud traffic data, generate local dynamic obstacle information based on the spatio-temporally aligned in-vehicle sensor data, generate static obstacle information based on the spatio-temporally aligned roadside device data, and generate global traffic information based on the spatio-temporally aligned cloud traffic data; perform weighted fusion on the local dynamic obstacle information, the static obstacle information, and the global traffic information to obtain traffic fusion data; generate a global traffic situation map according to the traffic fusion data, where the global traffic situation map includes the road network structure, traffic flow status, and obstacle distribution of each highway section.

[0022] In the embodiments of the present application, for the time and space differences of in-vehicle sensor data, roadside device data, and cloud traffic data, a timestamp synchronization algorithm (such as the NTP protocol) and a coordinate conversion technology (such as mapping vehicle GPS coordinates to a high-precision map coordinate system) are used to achieve spatio-temporal alignment of multi-source data, and eliminate the spatio-temporal misalignment of multi-source data caused by acquisition frequency, transmission delay, or device location. For example, an in-vehicle radar detects an obstacle 50 meters ahead, while a roadside camera reports the same obstacle at 60 meters due to different viewing angles. Spatio-temporal alignment uses a spatial coordinate conversion and time interpolation algorithm to unify the two data to the same spatio-temporal reference. Further, based on the spatio-temporally aligned in-vehicle sensor data, a target tracking algorithm (such as Kalman filtering) is used to identify moving obstacles (such as vehicles, pedestrians), and predict their movement trajectories within the next 3 seconds, generating a dynamic obstacle layer including position, speed, and acceleration, to obtain local dynamic obstacle information. Roadside device data is used to identify road fixed facilities (such as construction cones, guardrails) and abnormal stationary objects (such as accident vehicles), and combined with high-precision maps to mark the types and positions of static obstacles, to obtain static obstacle information. Using historical traffic flow data (such as average vehicle speed, traffic volume), event information (such as construction plans), and weather forecasts and other cloud traffic data obtained from the cloud, a lane-level traffic status map covering the entire section (such as congestion index, visibility level) is generated to obtain global traffic information. Then, a dynamic weight allocation model is designed to adjust the fusion weights according to the data source reliability (such as high real-time performance of in-vehicle sensors but vulnerable to occlusion, limited coverage of roadside devices but strong stability) and scenario requirements (such as giving priority to cloud historical data during congestion, relying on in-vehicle perception during sudden accidents). For example, in rainy weather, the weight of roadside weather station data is increased to 60%, while the weight of in-vehicle cameras is reduced to 30% due to raindrop interference. The fused data is input into a 3D map engine, overlaying the road network structure (such as lane lines, ramp positions), traffic flow status (color gradient represents congestion degree), and obstacle distribution (icons mark types and distances), to generate a global traffic situation map with spatio-temporal continuity.

[0023] Step 202: For each highway section in the global traffic situation map, extract the current closed road length L of the highway section according to the road network structure of the highway section now and the maximum closed road length L max . Extract the current vehicle density ρ of the highway section according to the traffic flow state of the highway section now and the maximum vehicle density ρ max . Extract the current number of accident points N of the highway section according to the obstacle distribution of the highway section now and the maximum possible number of accident points N max .

[0024] Step 203: Obtain the current visibility corresponding to each highway section ; According to the current closed road length, maximum closed road length, current vehicle density, maximum vehicle density, current number of accident points, maximum possible number of accident points, and current visibility corresponding to each highway section , determine the risk metric RI of each highway section respectively; Determine the risk level of each highway section according to the interval to which the risk metric of each highway section belongs.

[0025] Among them, the risk metric calculation formula is: , respectively represent the first risk weight, second risk weight, third risk weight, and fourth risk weight represents the natural constant.

[0026] In the embodiments of the present application, the risk level assessment is based on multi-dimensional data of the global traffic situation map, and the risk metrics of each road section are dynamically calculated through quantitative indicators. The specific steps are as follows: Parse the road section closure information from the road network structure of the global situation map, extract the current closure length (such as the lane length closed due to construction or accident) and the maximum closure length of the road section (such as the maximum allowable closure length designed), and calculate the closure ratio (current closure length / maximum closure length) to reflect the degree of restriction of the road traffic capacity. Parse the current vehicle density data (the number of vehicles in a unit length road section) from the traffic flow state of the global traffic situation map, and combine it with the historical maximum vehicle density (such as the vehicle density when the road section is most congested) to calculate the congestion index (current vehicle density / maximum vehicle density) to quantify the traffic flow saturation. Analyze the current number of accident points (such as the locations of accidents or breakdown vehicles that have occurred) and the maximum possible number of accident points (such as potential risk points predicted based on historical accident rates) from the obstacle distribution of the global situation map, and calculate the accident risk coefficient (current number of accident points / maximum possible number of accident points) to evaluate the probability of secondary accidents. And obtain the current visibility (such as the visible distance under haze or heavy rain) through roadside equipment or meteorological data as the environmental risk correction factor. Finally, the above parameters are normalized and fused into a risk metric. For example, using the formula: , respectively represent the first risk weight, the second risk weight, the third risk weight, and the fourth risk weight, represents the natural constant. Further, according to the risk metric threshold (such as R < 0.3 for low risk, 0.3 ≤ R < 0.6 for medium risk, R ≥ 0.6 for high risk), the road sections are graded by risk, providing a decision basis for subsequent takeover strategies. The embodiments of the present application achieve refined risk quantification of highway road sections through multi-dimensional parameter fusion and weighting. Compared with the traditional evaluation method that relies on a single indicator (such as only vehicle density), this solution comprehensively considers road structure, traffic flow, obstacles, and environmental factors, avoiding misjudgment caused by fluctuations in a single parameter. For example, in the case of extremely low visibility but acceptable vehicle density, the traditional method may underestimate the risk, while this solution improves the risk metric through the visibility correction factor, triggering a more stringent control strategy.

[0027] Step 204: When the risk level is the low risk level, determine that the takeover control strategy for the intelligent driving vehicle in the highway road section is the autonomous planning strategy; when the risk level is the medium risk level, determine that the takeover control strategy for the intelligent driving vehicle in the highway road section is the formation cooperation strategy; when the risk level is the high risk level, determine that the takeover control strategy for the intelligent driving vehicle in the highway road section is the second-level takeover strategy.

[0028] Among them, the autonomous planning strategy instructs the intelligent driving vehicle to autonomously perform driving planning and send driving planning information to the remote safety officer terminal; the second-level takeover strategy instructs the intelligent driving vehicle to be immediately taken over by the remote safety officer terminal.

[0029] In the embodiments of the present application, on low-risk road sections (such as normal visibility, vehicle density lower than the threshold, and no road closures), the intelligent driving vehicle is allowed to make autonomous decisions based on in-vehicle sensors and high-precision maps, including lane keeping, speed adjustment, and overtaking planning. At the same time, the vehicle needs to upload the planned trajectory (such as the driving path in the next 10 seconds) to the remote safety officer terminal in real time via the 5G network, and the safety officer can intervene and correct it at any time. On medium-risk road sections (such as construction areas or minor accident areas), the vehicle needs to join a formation coordinated by roadside edge nodes, maintaining a fixed distance (such as a 10-meter vehicle distance) and a synchronous speed (such as 80 km / h). The formation instructions are synchronized to all member vehicles to ensure the stability of the queue. On high-risk road sections (such as visibility less than 50 meters or detected hazardous material leakage), the vehicle immediately triggers remote takeover, and the safety officer directly controls the steering wheel, accelerator, and brake. The takeover response time needs to be less than 1 second. At the same time, the vehicle automatically turns on the hazard lights and plays a takeover prompt tone. For example, when the visibility suddenly drops to 30 meters due to a fog bank, the vehicle sensor triggers takeover within 0.5 seconds after detecting the risk. The safety officer takes over the vehicle in the cloud cockpit, drives slowly at a speed of 20 km / h, and broadcasts the takeover status to surrounding vehicles through roadside equipment. Through risk classification, precise allocation of control resources is achieved, avoiding the efficiency loss of the traditional "one-size-fits-all" mode. The autonomous planning in low-risk scenarios not only reduces the burden on remote safety officers but also retains the manual intervention channel through real-time trajectory reporting, balancing automation and safety; the formation coordination in medium-risk scenarios uses swarm intelligence to replace single-vehicle decision-making, significantly reducing the operation error rate in complex environments; the second-level takeover in high-risk scenarios shortens the reaction time in extreme scenarios through the closed-loop link of "vehicle trigger - system response - manual control", improves the accident avoidance ability, enhances the system's adaptability to emergencies, and ensures the continuous operation of traffic flow within the safety boundary.

[0030] Step 205: Determine candidate lead vehicles based on the remote takeover capabilities of each intelligent driving vehicle within the corresponding highway section. Determine the lead vehicle according to the driving parameters of each candidate lead vehicle, and use the remaining intelligent driving vehicles within the corresponding highway section as following vehicles. Number each following vehicle according to its position, and determine the preceding vehicle corresponding to each following vehicle according to the number of each following vehicle. Among them, the preceding vehicle of the first following vehicle is the lead vehicle, and the preceding vehicle of the remaining following vehicles is the following vehicle with the previous number; the remote safety officer terminal generates a lead control instruction for the lead vehicle as the formation cooperation instruction of the lead vehicle, and sends the formation cooperation instruction of the lead vehicle to the lead vehicle and sinks it to the roadside edge node; based on the lead control instruction and the numbers of each following vehicle, the roadside edge node calculates the following control instructions of each following vehicle one by one starting from the first following vehicle to obtain the formation cooperation instructions of each following vehicle.

[0031] In the embodiment of the present application, the implementation of the formation cooperation strategy is achieved through dynamic election of the lead vehicle, hierarchical sinking of instructions, and distributed calculation. The specific process is as follows: First, screen candidate lead vehicles according to the vehicle remote takeover capabilities (such as in-vehicle computing power, communication delay, historical takeover success rate), and then determine the optimal lead vehicle in combination with real-time driving parameters (such as vehicle position, vehicle speed stability, distance from the obstacle ahead). For example, preferentially select an intelligent driving vehicle that has the ability to directly communicate with the remote safety officer terminal, has a good current communication status, and is located in the front position of the vehicle fleet as the lead vehicle. Then, number the vehicles from the back to the front according to their positions (for example, vehicles A, B, and C are numbered 1, 2, and 3 in sequence). The preceding vehicle of the first following vehicle (number 1) is the lead vehicle, and the subsequent vehicles use the vehicle with the previous number as the preceding vehicle. This logic ensures that the formation forms a chain structure, facilitating the gradual transmission of instructions. As Figure 3As shown in the figure, it is a schematic diagram of the communication relationship among the cloud, roadside edge nodes, and intelligent driving vehicles. Further, the remote safety officer terminal (cloud) generates a lead vehicle instruction (such as a target speed of 80 km / h) based on the global situation map, and this instruction is sent to both the lead vehicle and the roadside edge node simultaneously. The roadside edge node stores the lead instruction as a formation reference parameter and avoids instruction loss caused by cloud interruption through local caching. For example, in a tunnel scenario, even if the 5G signal is interrupted, the edge node can still maintain the formation for more than 30 seconds based on the locally stored instruction. After that, the roadside edge node adopts a distributed computing framework to generate control instructions step by step in the order of the following vehicle numbers. For example, when the lead vehicle accelerates, the edge node first calculates the vehicle speed and steering angle that the following vehicle numbered 1 needs to match, and then derives the instruction for the following vehicle numbered 2 based on the real-time state of the following vehicle numbered 1, and so on. In addition, the following vehicle instruction can also include compensation parameters (such as wind resistance correction, prediction of the acceleration of the preceding vehicle) to eliminate the cumulative error of chain transmission. For example, when the lead vehicle brakes suddenly, the roadside edge node generates a deceleration instruction for the following vehicle 0.3 seconds in advance through a prediction model to avoid rear-end collisions. The embodiment of the present application, through the cooperation of the dynamic formation architecture and edge computing, helps to improve the cooperative control ability in medium-risk scenarios. The dynamic lead vehicle election mechanism avoids the disintegration risk of the fixed formation in case of vehicle failure, ensuring that the formation is always dominated by the optimal vehicle; the hierarchical sinking of instructions and local caching enables the formation control to get rid of the dependence on a single communication link, and the local cooperation can still be maintained even if the cloud is interrupted; in addition, the chained preceding vehicle binding logic simplifies the formation topology management and avoids instruction conflicts in complex scenarios. For example, when changing lanes in a construction area, the vehicle can quickly reorganize the queue based on the number of the preceding vehicle, reducing the need for manual intervention.

[0032] Step 206: Based on each communication link corresponding to the multi-link redundancy architecture, perform a communication link score on each following vehicle, and determine the target communication link of each following vehicle according to the communication link with the highest communication link score corresponding to each following vehicle; divide the data related to the preset key parameters in the formation cooperation instruction corresponding to each following vehicle into a preset number of key parameter data blocks and pack the non-key parameters into other parameter data packets, and send the key parameter data blocks and other parameter data packets to each following vehicle based on the target communication link of each following vehicle; each following vehicle performs integrity verification on the received key parameter data blocks, and restores the formation cooperation instruction of the following vehicle based on the complete data blocks and other parameter data packets. Among them, if the number of continuously received incomplete data blocks reaches the emergency trigger number, the local degradation takeover strategy of the following vehicle is triggered, and the local degradation takeover strategy instructs the following vehicle to stop at the nearest available stop position.

[0033] In the embodiments of the present application, according to the real-time communication performance between each communication link provided in the multi-link redundancy architecture and each following vehicle (such as the signal strength of the 5G cellular network, the latency of C-V2X direct communication, and the packet loss rate of DSRC), the communication link score corresponding to each following vehicle is dynamically calculated. For example, in a tunnel scenario, when the 5G signal of the vehicle attenuates, the system automatically reduces its score and preferentially selects the C-V2X link with stronger anti-interference ability. Then, the link with the highest score is assigned to each following vehicle as the target communication link. For example, the vehicle at the front of the convoy uses C-V2X because it is close to the roadside equipment, and the vehicle at the rear switches to 5G for supplementary coverage because it is far from the base station. When data is transmitted, the key parameters in the formation coordination instruction (such as the target vehicle speed and steering angle) are divided into multiple data blocks, and redundant check information is added using forward error correction coding (such as Reed-Solomon code); the non-critical parameters (such as the in-vehicle temperature and entertainment system status) are packed into a single data packet. The following vehicle performs cyclic redundancy check (CRC) on the received data block. If data corruption is detected, it attempts to repair the data through the redundant check information. For example, if 10% of the content of a data block is lost due to interference, the system uses the error correction code to restore the complete data. If the data blocks received continuously for 3 times cannot be repaired (reaching the emergency trigger count), the local degradation strategy is automatically triggered, and the vehicle gradually decelerates to stop at the nearest emergency lane, and activates the hazard lights to communicate with the roadside equipment. Thus, through multi-link redundancy and data priority management, the reliability and fault tolerance of instruction transmission are improved, network resource allocation is optimized, and while ensuring the stability of the formation, the traffic efficiency and safety redundancy in high-risk scenarios on highways are enhanced.

[0034] In the embodiments of the present application, optionally, after triggering the local degradation takeover strategy of the following vehicle, it further includes: the following vehicle sends a local degradation takeover signal to the roadside edge node, and the roadside edge node determines the departing following vehicle based on the local degradation takeover signal and updates the following vehicles, and re-determines the numbers of each following vehicle; the roadside edge node monitors the recovery of each communication link corresponding to the departing following vehicle, and when the communication link score of any communication link of the departing following vehicle recovers to the preset rejoining score, the departing following vehicle is followed up and the numbers of each following vehicle are re-determined again.

[0035] In this embodiment, the subsequent process after triggering the local downgrade takeover policy is realized through dynamic formation reorganization and link recovery monitoring. The specific process is as follows: After the follower vehicle triggers the downgrade policy, it can send a structured signal to the roadside edge node through V2X direct communication, including the vehicle ID, downgrade reason (such as continuous loss of data blocks), current position and status (such as having decelerated to a safe vehicle speed). After receiving the signal, the roadside edge node marks the vehicle as "off-team status", indicating that the vehicle is an off-team follower vehicle, and removes it from the formation list. For example, if vehicle No. 3 leaves the team due to communication interruption, the edge node will move the numbers of vehicle No. 4 and subsequent vehicles forward (the original No. 4 becomes 3, and the original No. 5 becomes 4), and at the same time update the formation topology. The edge node recalculates the control parameters of each vehicle according to the new numbers (such as the inter-vehicle distance is adjusted from 15 meters to 12 meters to adapt to the shortened formation length), and issues an update instruction through the target communication link. In addition, the roadside edge node continuously monitors the communication link scores of the off-team vehicle (such as 5G latency, C-V2X packet loss rate). When any link score exceeds the preset threshold (such as the latency is lower than 50 ms), that is, when it returns to the preset rejoin score, the rejoin process is triggered. The edge node sends a rejoin instruction to the off-team vehicle, including the new number, target position and speed synchronization parameters. For example, after vehicle No. 3 resumes communication, the edge node inserts it at the end of the current formation, adjusts the number to 5, and requires it to accelerate to 80 km / h to match the vehicle speed of the fleet. Through the dynamic formation management and automatic recovery mechanism in this embodiment of the application, the adaptability of the system to vehicle departure and rejoining is improved. The downgrade signal sending and off-team processing ensure that the formation can be quickly reorganized in case of vehicle failure, avoiding the disintegration of the formation caused by single-point failure (in a traditional fixed formation, the whole formation needs to stop when one vehicle leaves the team); the dynamic number adjustment keeps the formation in a compact structure all the time, reducing the control chaos caused by number vacancies (such as a vehicle misidentifying the number of the preceding vehicle and causing an emergency brake); the link recovery monitoring and automatic rejoining pass through the "monitoring-triggering-synchronization" closed loop, shortening the rejoining time after the vehicle leaves the team and improving the formation throughput. In addition, the speed synchronization and position insertion mechanisms during the rejoining process avoid the risk of rear-end collision and ensure that the formation can still maintain stable operation during dynamic changes.

[0036] In the embodiment of the present application, optionally, key parameter data blocks and other parameter data packets are sent to each follower vehicle based on the target communication link of each follower vehicle, including: if the communication link score of the target communication link is greater than the preset high-performance link score, key parameter data blocks and other parameter data packets are sent to each follower vehicle based on the target communication link of each follower vehicle; if the communication link score of the target communication link is less than or equal to the preset high-performance link score, a key parameter database is sent to each follower vehicle based on the target communication link of each follower vehicle, and other parameter data packets are sent to each follower vehicle based on the communication link with the second highest communication link score corresponding to each follower vehicle.

[0037] In an embodiment of the present application, optionally, after restoring the formation cooperation instruction of the following vehicle based on the complete data block, the method further includes: each following vehicle performs a consistency calculation on the leading vehicle to determine the following difference between each following vehicle and the leading vehicle, and adjusts the following parameters of the formation cooperation instruction of the following vehicle based on the following difference; where the consistency calculation formula is: , where represents the following difference, respectively represent the desired formation spacing and the desired formation spacing change rate, respectively represent the actual spacing between the i-th following vehicle and the leading vehicle, the actual spacing change rate between the i-th following vehicle and the leading vehicle, respectively represent the spacing control gain and the spacing change rate control gain.

[0038] In this embodiment, after restoring the formation cooperation instruction of the following vehicle based on the complete data block, the following vehicle realizes the dynamic stability of the vehicle through the consistency calculation and parameter adjustment mechanism. The specific process is as follows: The following vehicle uses on-vehicle sensors (such as radar and camera) to collect the motion state of the leading vehicle in real time, including position (latitude and longitude coordinates), speed (vector direction and magnitude), acceleration (longitudinal and lateral), and heading angle (vehicle driving direction). The distributed consistency algorithm is used to calculate the following difference in combination with the state difference between the vehicle itself and the leading vehicle, and based on the calculation result of the following difference, the system dynamically adjusts the throttle opening, brake pressure, and steering angle compensation value. In addition, after the parameters of the following vehicle are adjusted, its state change can be transmitted to the subsequent vehicle through V2X communication to trigger cascade adjustment. Further, the roadside edge node can also monitor the overall formation parameters (such as average vehicle distance and speed dispersion) in real time. When detecting abnormal fluctuations (such as three vehicles braking simultaneously), a global speed adjustment instruction is sent to the leading vehicle to restore the formation stability by adjusting the driving parameters of the leading vehicle. Thereby improving the formation's ability to respond to emergencies (such as stability in emergency braking and lane-changing scenarios), optimizing the formation passing efficiency (reducing speed fluctuations caused by frequent adjustments), and enhancing the system's adaptability to complex traffic situations (such as queue maintenance during detours in construction areas).

[0039] The embodiments of the present application further provide a computer device, which may specifically be a personal computer, a server, a network device, etc. The computer device includes a bus, a processor, a memory, and a communication interface, and may further include an input / output interface and a display device. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store location information. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the steps in the method embodiments are implemented.

[0040] Those skilled in the art can understand that the structure of the above computer device is only a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components, or combine certain components, or have different component arrangements.

[0041] In one embodiment, a computer-readable storage medium is provided. The computer-readable storage medium may be non-volatile or volatile, and stores a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0042] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0043] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties.

[0044] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0045] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0046] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways, characterized in that, Including: Obtain in-vehicle sensor data, roadside equipment data, and cloud traffic data corresponding to a preset highway section, and construct a global traffic situation map according to the in-vehicle sensor data, the roadside equipment data, and the cloud traffic data; Evaluate the risk levels of each highway section according to the global traffic situation map, and determine the takeover control strategies for intelligent driving vehicles within each highway section according to the risk levels; When the takeover control strategy is a formation cooperation strategy, generate a formation cooperation instruction for the intelligent driving vehicles within the corresponding highway section, and sink the formation cooperation instruction to the roadside edge node; Send the formation cooperation instruction to the intelligent driving vehicle based on a multi-link redundancy architecture.

2. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways according to claim 1, characterized in that Construct a global traffic situation map according to the in-vehicle sensor data, the roadside equipment data, and the cloud traffic data, including: Perform spatio-temporal alignment on the in-vehicle sensor data, the roadside equipment data, and the cloud traffic data, generate local dynamic obstacle information according to the spatio-temporally aligned in-vehicle sensor data, generate static obstacle information according to the spatio-temporally aligned roadside equipment data, and generate global traffic information according to the spatio-temporally aligned cloud traffic data; Perform weighted fusion on the local dynamic obstacle information, the static obstacle information, and the global traffic information to obtain traffic fusion data; Generate a global traffic situation map according to the traffic fusion data, where the global traffic situation map includes the road network structure, traffic flow status, and obstacle distribution of each highway section.

3. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme highway scenarios according to claim 2, characterized in that, Evaluate the risk levels of each highway section according to the global traffic situation map, including: For each highway section in the global traffic situation map, extract the current closed road length L of the highway section according to the road network structure of the highway section now and the maximum closed road length L max , extract the current vehicle density ρ of the highway section according to the traffic flow state of the highway section now and the maximum vehicle density ρ max , extract the current number of accident points N of the highway section according to the obstacle distribution of the highway section now and the maximum possible number of accident points N max ; Obtain the current visibility corresponding to each highway section ; According to the current road closure length, maximum road closure length, current vehicle density, maximum vehicle density, current number of accident points, maximum possible number of accident points, and current visibility corresponding to each highway section , the risk metric RI of each highway section is determined respectively, where the risk metric calculation formula is: , respectively represent the first risk weight, the second risk weight, the third risk weight, and the fourth risk weight, represents the natural constant; Respectively determine the risk levels of each highway section according to the intervals to which the risk metrics of each highway section belong.

4. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways according to claim 1, wherein, Determine the takeover control strategies for intelligent driving vehicles within each highway section according to the risk levels, including: When the risk level is a low risk level, determine that the takeover control strategy for the intelligent driving vehicle within the highway section is an autonomous planning strategy; When the risk level is a medium risk level, determine that the takeover control strategy for the intelligent driving vehicle within the highway section is a formation cooperation strategy; When the risk level is a high risk level, determine that the takeover control strategy for the intelligent driving vehicle within the highway section is a second-level takeover strategy.

5. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme highway scenarios according to claim 4, wherein The autonomous planning strategy instructs the intelligent driving vehicle to perform autonomous driving planning and send driving planning information to the remote safety officer terminal; The second-level takeover strategy instructs the intelligent driving vehicle to be immediately taken over by the remote safety officer terminal.

6. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways according to claim 1, characterized in that When the takeover control strategy is a formation cooperation strategy, generate a formation cooperation instruction for the intelligent driving vehicles within the corresponding highway section, and sink the formation cooperation instruction to the roadside edge node, including: Determine candidate lead vehicles according to the remote takeover capabilities of each intelligent driving vehicle in the corresponding highway section, determine the lead vehicle according to the driving parameters of each candidate lead vehicle, use the remaining intelligent driving vehicles in the corresponding highway section as following vehicles, number each following vehicle according to its position, and determine the preceding vehicle corresponding to each following vehicle according to the number of each following vehicle. Among them, the preceding vehicle of the first following vehicle is the lead vehicle, and the preceding vehicle of the remaining following vehicles is the following vehicle with the previous number; The remote safety officer terminal generates a lead control instruction for the lead vehicle as the formation cooperation instruction of the lead vehicle, and sends the formation cooperation instruction of the lead vehicle to the lead vehicle and sinks it to the roadside edge node; Based on the lead control instruction and the numbers of each following vehicle, the roadside edge node calculates the following control instructions for each following vehicle one by one starting from the first following vehicle to obtain the formation cooperation instructions for each following vehicle.

7. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways according to claim 6, wherein, Based on the multi-link redundancy architecture, sending the formation cooperation instruction to the intelligent driving vehicle includes: Based on each communication link corresponding to the multi-link redundancy architecture, score the communication links of each following vehicle, and determine the target communication link of each following vehicle according to the communication link with the highest communication link score corresponding to each following vehicle; Divide the data related to the preset key parameters in the formation cooperation instruction corresponding to each following vehicle into a preset number of key parameter data blocks and pack the non-key parameters into other parameter data packets, and send the key parameter data blocks and other parameter data packets to each following vehicle based on the target communication link of each following vehicle; Each following vehicle performs integrity verification on the received key parameter data blocks, and restores the formation cooperation instruction of the following vehicle based on the complete data blocks and other parameter data packets. Among them, if the number of continuously received incomplete data blocks reaches the emergency trigger number, trigger the local downgrade takeover strategy of the following vehicle, and the local downgrade takeover strategy instructs the following vehicle to stop at the nearest available stop position.

8. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways according to claim 7, wherein, After triggering the local downgrade takeover strategy of the following vehicle, it further includes: The following vehicle sends a local downgrade takeover signal to the roadside edge node, and the roadside edge node determines the departing following vehicle based on the local downgrade takeover signal and performs following vehicle update, and re-determines the numbers of each following vehicle; The roadside edge node monitors the recovery of each communication link corresponding to the departing following vehicle, and when the communication link score of any communication link of the departing following vehicle recovers to the preset rejoin score, performs following recovery on the departing following vehicle and re-determines the numbers of each following vehicle again.

9. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways according to claim 7, characterized in that Sending the key parameter data blocks and other parameter data packets to each following vehicle based on the target communication link of each following vehicle includes: If the communication link score of the target communication link is greater than the preset high-performance link score, send the key parameter data blocks and other parameter data packets to each following vehicle based on the target communication link of each following vehicle; If the communication link score of the target communication link is less than or equal to the preset high-performance link score, send the key parameter database to each following vehicle based on the target communication link of each following vehicle, and send other parameter data packets to each following vehicle based on the communication link with the second highest communication link score corresponding to each following vehicle.

10. The multi-link redundancy guarantee method for intelligent driving vehicles in extreme scenarios on highways according to claim 7, wherein, After restoring the formation cooperation instruction of the following vehicle based on the complete data block, it further includes: Each following vehicle performs a consistency calculation on the leading vehicle to determine the following difference between each following vehicle and the leading vehicle, and adjusts the following parameters of the formation cooperation instruction of the following vehicle based on the following difference; wherein, the consistency calculation formula is: , where represents the following difference, respectively represent the desired formation spacing and the desired formation spacing change rate, respectively represent the actual spacing between the i-th following vehicle and the preceding vehicle, the actual spacing change rate between the i-th following vehicle and the preceding vehicle, respectively represent the spacing control gain and the spacing change rate control gain.

Images