A vehicle control system

Abstract

The present invention discloses a vehicle control system, including the following modules: a multi-source data acquisition module for the vehicle active safety system, which is used to generate a standardized data set for the vehicle active safety system; a vertical architecture data distribution construction module for the vehicle active safety system, which is used to construct a vertical architecture data distribution model for the vehicle active safety system based on the standardized data set for the vehicle active safety system; a fractional-order seagull optimization global path search module, which is used to form a set of preliminarily optimized data transmission paths for the vehicle active safety system; a wolf pack search local path optimization module, which is used to generate the finally optimized data transmission paths for the vehicle active safety system; and a data distribution execution module for the vehicle active safety system, which is used to execute the data transmission tasks for the vehicle active safety system. The present invention adopts a vertical architecture, directly establishing the shortest path from the data acquisition node to the execution module, avoiding the data transmission delay caused by multi-level transfer and significantly reducing the intermediate links of data transmission.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle control systems, and particularly to a vehicle control system. Background Art

[0002] With the development of intelligent connected vehicle technology, the vehicle active safety system plays an increasingly important role in the modern traffic environment. The vehicle active safety system collects and processes data in real time through sensors, communication modules, and control units to achieve the perception, judgment, and decision-making of the vehicle on the external environment, thereby improving driving safety.

[0003] Currently, traditional vehicle active safety systems usually adopt a layered data transmission architecture, that is, data is transmitted from the perception layer to the decision-making layer and the execution layer step by step after multi-level processing. Although the layered architecture helps with the modular design of the system, its hierarchical structure results in a long data transmission path, and information is transferred between multiple nodes, increasing communication latency. Especially in emergency situations, such as sudden collision warning or automatic emergency braking, the high latency of the layered architecture may lead to a lag in safety decisions and affect the active safety performance of the vehicle.

[0004] On the other hand, existing data distribution path optimization methods lack intelligence and real-time performance and are difficult to adapt to the complex and changing in-vehicle network environment. Currently, common path optimization methods are mostly based on static rules or simple heuristic search algorithms and are difficult to dynamically adapt to the multi-source data distribution and real-time transmission requirements in the vehicle active safety system. With the popularization of intelligent connected vehicles, the in-vehicle network environment becomes more complex, and the network load will change dynamically with factors such as road conditions, weather, and vehicle density. Existing static optimization methods are prone to data congestion under high load conditions, causing the system to be unable to transmit important information in a timely manner at critical moments. In addition, some path optimization methods only focus on the shortest path search and ignore factors such as network congestion and node load balancing, resulting in data backlog on some high-load paths and reducing the overall transmission efficiency.

[0005] To solve the problems existing in the prior art, there is an urgent need for a new data distribution method for vehicle active safety systems, which replaces the traditional layered architecture with a flat or vertical architecture to reduce data transmission latency, and at the same time combines intelligent optimization algorithms to improve the real-time optimization ability of the data transmission path, enabling the system to always maintain efficient and stable data distribution capabilities in a complex dynamic network environment. Summary of the Invention

[0006] An object of the present invention is to provide a vehicle control system. The present invention adopts a vertical architecture, directly establishing the shortest path from the data acquisition node to the execution module, avoiding data transmission latency caused by multiple-level transfers, and significantly reducing the intermediate links of data transmission by optimizing the data distribution topology structure.

[0007] A vehicle control system according to an embodiment of the present invention includes the following modules:

[0008] A multi-source data acquisition module for the vehicle active safety system, which is used to acquire multi-source data involved in the operation of the vehicle active safety system and generate a standardized data set for the vehicle active safety system.

[0009] A vertical architecture data distribution construction module for the vehicle active safety system, which is used to construct a vertical architecture data distribution model for the vehicle active safety system based on the standardized data set of the vehicle active safety system;

[0010] A fractional-order seagull optimization global path search module, which is used to globally search and optimize the data transmission path based on the vertical architecture data distribution model of the vehicle active safety system by using the fractional-order seagull optimization algorithm, and form a set of preliminarily optimized data transmission paths for the vehicle active safety system;

[0011] A wolf pack search local path optimization module, which is used to locally optimize the preliminarily optimized data transmission paths of the vehicle active safety system by using the wolf pack search algorithm and generate the finally optimized data transmission paths of the vehicle active safety system;

[0012] A data distribution execution module for the vehicle active safety system, which is used to execute the data transmission task of the vehicle active safety system according to the finally optimized data transmission paths of the vehicle active safety system.

[0013] A vehicle control method is applied to a vehicle control system, including the following steps:

[0014] S1. Collect multi-source data of the vehicle active safety system, and preprocess the collected multi-source data of the vehicle active safety system to generate a standardized data set for the vehicle active safety system;

[0015] S2. Construct a vertical architecture data distribution model for the vehicle active safety system based on the standardized data set of the vehicle active safety system, and the vertical architecture data distribution model of the vehicle active safety system adopts a flattened data transmission structure;

[0016] S3. Use the fractional-order seagull optimization algorithm to globally search and optimize each multi-source data transmission path in the vertical architecture data distribution model of the vehicle active safety system to form a set of preliminarily optimized data transmission paths for the vehicle active safety system;

[0017] S4. Use the wolf pack search algorithm to locally search and optimize the set of preliminarily optimized data transmission paths of the vehicle active safety system, refine the data transmission paths of the vehicle active safety system, and generate the finally optimized data transmission paths of the vehicle active safety system;

[0018] S5. Implement real-time data transmission during the operation of the vehicle active safety system based on the finally optimized data transmission path of the vehicle active safety system, dynamically collect the network status of the vehicle active safety system and the operation status information of the vehicle active safety system, and perform real-time monitoring and dynamic adjustment on the data transmission path of the vehicle active safety system.

[0019] Optionally, the S1 includes the following steps:

[0020] S11. Collect multi-source data of the vehicle active safety system. The multi-source data of the vehicle active safety system includes data from in-vehicle sensors, environmental perception devices, vehicle control modules, and network communication modules. Each multi-source data sample of the vehicle active safety system is represented as:

[0021] ;

[0022] Wherein, is the th multi-source data sample of the vehicle active safety system, represents the timestamp of this multi-source data sample of the vehicle active safety system, represents the numerical information of this multi-source data sample of the vehicle active safety system, is the device identifier of the source of the multi-source data of the vehicle active safety system, is the data category label;

[0023] S12. Construct a dataset of the vehicle active safety system based on the collected multi-source data of the vehicle active safety system. Set a sampling time window, sort the collected multi-source data of the vehicle active safety system in chronological order to form an original dataset of the vehicle active safety system ;

[0024] S13. Clean the data of the original dataset of the vehicle active safety system, remove invalid data, duplicate data, and abnormal data, and obtain a cleaned dataset of the vehicle active safety system;

[0025] S14. Perform data standardization processing on the cleaned dataset of the vehicle active safety system, and use the normalization method to convert the data to the interval , and obtain a standardized dataset of the vehicle active safety system;

[0026] S15. Remove redundant information from the standardized dataset of the vehicle active safety system, calculate the feature correlation matrix of the multi-source data of the vehicle active safety system, and form a final standardized dataset of the vehicle active safety system .

[0027] Optionally, the calculation of the feature correlation matrix of the multi-source data of the vehicle active safety system is:

[0028] ;

[0029] wherein, represents the correlation coefficient between the -th dimensional feature and the -th dimensional feature, is the total number of data samples, and respectively represent the means of the k-th data sample on the -th dimensional feature and the -th dimensional feature, and and respectively represent the standardized values of the k-th data sample on the

[0030] ;

[0031] wherein, is the standardized data set of the vehicle active safety system, is the multi-source data information of the vehicle active safety system after redundancy removal, is the -th multi-source data sample of the vehicle active safety system finally.

[0032] Optionally, the S2 includes the following steps:

[0033] S21. Based on the standardized data set of the vehicle active safety system, define a mapping function to convert each standardized data sample of the vehicle active safety system into a corresponding data distribution node, where is the set of all data distribution nodes; is the mapping function from the data sample to the node;

[0034] S22. According to the data distribution node set construct a direct connection set between nodes, and define a similarity evaluation function :

[0035] ;

[0036] wherein, represents the set of positive real numbers, and are the timestamps of the corresponding data samples respectively, and is a positive real - valued weight coefficient. When the similarity evaluation function is satisfied, it is considered that there is a direct transmission ability between node and node , and a direct connection is established to obtain the set of direct connections between nodes :

[0037] ;

[0038] wherein, is a preset similarity threshold;

[0039] S23. For each direct connection , a transmission weight function is defined to quantify the delay or cost of data transmission:

[0040] ;

[0041] wherein, represents the data transmission delay from node to node , represents the data transmission cost on this direct connection, and are the weight coefficients corresponding to the delay and cost respectively;

[0042] S24. Using the results of steps S21 to S23, the vertical architecture data distribution model of the vehicle active safety system is constructed as a graph model:

[0043] ;

[0044] wherein, represents the vertical architecture data distribution model of the vehicle active safety system. The vertical architecture data distribution model of the vehicle active safety system adopts a flattened data transmission structure. The node set represents all data acquisition and control execution units, and the direct connection set and its transmission weight function enable data to directly reach the control execution module from the acquisition node.

[0045] Optionally, the said S3 includes the following steps:

[0046] S31. Initialize the fractional - order seagull population set based on the vertical architecture data distribution model M of the vehicle active safety system. Each fractional - order seagull individual is represented as:

[0047] ;

[0048] Among them, represents the th direct connection, the starting point of the path is the acquisition node, and the end point is the control execution module. represents the population size;

[0049] S32. Define the load balancing fitness function for the multi-source data transmission path of the vehicle active safety system :

[0050] ;

[0051] Among them, is the data transmission weight of the direct connection , k is the number of direct connections in the data transmission path of the vehicle active safety system, is the data transmission load of the th node in the path, is the average value of the data transmission loads of the nodes on the path, is the load balancing weight coefficient, enabling the path to consider both the shortest transmission and load balancing performance;

[0052] S33. Construct the fractional-order seagull position update for the network state adaptation of the vehicle active safety system as:

[0053] ;

[0054] Among them, and are the fractional-order seagull individual path positions at the th and th iterations respectively, and are the position update step size and the global optimal guidance coefficient, is the fractional-order order of the vehicle active safety system network state adaptation dynamic fraction, is the gamma function, is the vehicle active safety system real-time network load dynamic regulation factor;

[0055] S34. Define the update formula for the global optimal data transmission path of the vehicle active safety system:

[0056] ;

[0057] Among them, is the optimal path position in the th iteration process;

[0058] S35. Based on the vehicle active safety system network state dynamic feedback mechanism, adjust the fractional-order order and the network load dynamic regulation factor in real time:

[0059] ;

[0060] ;

[0061] wherein, and are the minimum and maximum boundaries of the fractional order respectively, and are the minimum and maximum boundaries of the dynamic adjustment factor of the network load respectively, is the current real-time network load value of the vehicle active safety system, and are the historical minimum and maximum values of the network load, is the current real-time network delay value of the vehicle active safety system, and are the historical minimum and maximum values of the network delay;

[0062] S36. Repeat steps S33 to S35 until the iteration stop condition is reached, and obtain the initially optimized set of data transmission paths of the vehicle active safety system :

[0063] ;

[0064] wherein, represents the optimal solution of the data transmission path of the vehicle active safety system after adaptive dynamic fractional order optimization update of the th fractional order seagull individual.

[0065] Optionally, the S4 includes the following steps:

[0066] S41. Initialize the set of wolf pack individuals based on the initially optimized set of data transmission paths of the vehicle active safety system , and each wolf pack individual is represented as:

[0067] ;

[0068] wherein, represents the initial search position of the th wolf pack individual, is the size of the wolf pack population, is the rd direct connection in the initially optimized data transmission path of the vehicle active safety system;

[0069] S42. Define the refined fitness function of the vehicle active safety system path :

[0070] ;

[0071] Among them, represents the data transmission weight of the direct connection , is the dynamic stability coefficient of the direct connection in the real-time network environment, is the real-time network load imbalance coefficient of the direct connection , represents the real-time data transmission bandwidth of the direct connection , is a positive real number weight coefficient;

[0072] S43. Construct a dynamic memory tracking behavior position update mechanism for the wolf pack to implement refined search and optimization of the data transmission path of the vehicle active safety system:

[0073] ;

[0074] Among them, and are the positions of the wolf pack individuals at the th and th iterations respectively, is the position of the wolf pack leader at the current iteration, representing the real-time local optimal path, is the memory position of the historical best path of the wolf pack individual, is the individual dynamic memory weight coefficient, , are the adjustment coefficients of the current optimal path guidance and historical path memory respectively;

[0075] S44. Define a real-time update mechanism for the wolf pack leader position to achieve local optimization guidance of the data transmission path of the vehicle active safety system in the dynamic network environment:

[0076] ;

[0077] Among them, is the real-time optimal path position determined based on the refined fitness function in the th iteration;

[0078] S45. Construct an adaptive wolf pack intelligent reconnaissance mechanism for the network environment of the vehicle active safety system and define the update formula for the wolf pack reconnaissance behavior:

[0079] ;

[0080] Among them, represents the real-time optimal path position after reconnaissance, , They are the random exploration coefficient and the optimal guidance coefficient of the reconnaissance behavior respectively. is the dynamic random perturbation coefficient of the reconnaissance behavior, which is adjusted according to the real-time network fluctuation amplitude of the vehicle active safety system. is the dynamic sensitivity coefficient of the wolf pack individual's reconnaissance, which changes dynamically with the path stability and load fluctuation. is to randomly select the position of the wolf pack individual.

[0081] S46. Repeat steps S43 to S45 until the local optimization iteration stop condition is reached, and generate the finally optimized data transmission path of the vehicle active safety system:

[0082] ;

[0083] Among them, is the optimal path position of the finally optimized data transmission of the vehicle active safety system.

[0084] The beneficial effects of the present invention are as follows:

[0085] 1. The present invention adopts a vertical architecture, directly establishing the shortest path from the data acquisition node to the execution module, avoiding the data transmission delay caused by multi-level transit. By optimizing the data distribution topology structure, the intermediate links of data transmission are significantly reduced, and the data transmission time in case of emergency is shortened to less than 60% of the traditional layered architecture, effectively improving the reaction speed of the active safety system.

[0086] 2. In the process of optimizing the data distribution path, the present invention combines the fractional-order seagull optimization algorithm, utilizes the memory characteristics and long-range correlation of fractional-order calculus, and can efficiently search for the global optimal data transmission path in a complex dynamic network environment. Through the dynamic fractional-order seagull flight model, the optimization algorithm takes into account both the local search ability and the global search ability. In a complex dynamic vehicle-mounted network environment, it can improve the adaptive optimization ability of the data transmission path, enhance the adaptability of path planning, and still maintain high path stability in case of sudden changes in network load.

[0087] 3. The present invention combines the improved wolf pack search algorithm to locally optimize the preliminarily optimized path, focuses on the real-time load balancing problem of the data transmission path, and ensures that the data transmission path has a low network congestion rate while being the optimal short path by introducing a dynamic memory tracking mechanism and an adaptive reconnaissance mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] The drawings are used to provide further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, but do not constitute a limitation to the present invention. In the drawings:

[0089] Figure 1A flowchart of a vehicle control system proposed by the present invention. Detailed implementation manners

[0090] Now, the present invention will be further described in detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0091] Refer to Figure 1 , a vehicle control system, including the following modules:

[0092] A multi-source data acquisition module for the vehicle active safety system, which is used to acquire multi-source data involved in the operation of the vehicle active safety system and generate a standardized data set for the vehicle active safety system.

[0093] A vertical architecture data distribution construction module for the vehicle active safety system, which is used to construct a vertical architecture data distribution model for the vehicle active safety system based on the standardized data set of the vehicle active safety system;

[0094] A fractional-order seagull optimization global path search module, which is used to globally search and optimize the data transmission path based on the vertical architecture data distribution model of the vehicle active safety system by using the fractional-order seagull optimization algorithm, and form a set of initially optimized data transmission paths for the vehicle active safety system;

[0095] A wolf pack search local path optimization module, which is used to locally optimize the initially optimized data transmission path of the vehicle active safety system by using the wolf pack search algorithm, and generate a finally optimized data transmission path for the vehicle active safety system;

[0096] A data distribution execution module for the vehicle active safety system, which is used to execute the data transmission task of the vehicle active safety system according to the finally optimized data transmission path of the vehicle active safety system.

[0097] A vehicle control method, which is applied to a vehicle control system, including the following steps:

[0098] S1. Collect multi-source data of the vehicle active safety system, and preprocess the collected multi-source data of the vehicle active safety system to generate a standardized data set for the vehicle active safety system;

[0099] S2. Construct a vertical architecture data distribution model for the vehicle active safety system based on the standardized data set of the vehicle active safety system, and the vertical architecture data distribution model of the vehicle active safety system adopts a flat data transmission structure;

[0100] S3. Use the fractional-order seagull optimization algorithm to globally search and optimize the multi-source data transmission paths of each vehicle active safety system in the vertical architecture data distribution model of the vehicle active safety system, forming a set of initially optimized vehicle active safety system data transmission paths;

[0101] S4. Use the wolf pack search algorithm to locally search and optimize the set of initially optimized vehicle active safety system data transmission paths, refine the vehicle active safety system data transmission paths, and generate the finally optimized vehicle active safety system data transmission paths;

[0102] S5. Implement real-time data transmission during the operation of the vehicle active safety system based on the finally optimized vehicle active safety system data transmission paths, dynamically collect the network status of the vehicle active safety system and the operation status information of the vehicle active safety system, and perform real-time monitoring and dynamic adjustment of the vehicle active safety system data transmission paths.

[0103] In this embodiment, S1 includes the following steps:

[0104] S11. Collect multi-source data of the vehicle active safety system. The multi-source data of the vehicle active safety system includes data from in-vehicle sensors, environmental perception devices, vehicle control modules, and network communication modules. Each multi-source data sample of the vehicle active safety system is represented as:

[0105] ;

[0106] Among them, is the th multi-source data sample of the vehicle active safety system, represents the timestamp of this multi-source data sample of the vehicle active safety system, represents the numerical information of this multi-source data sample of the vehicle active safety system, is the device identifier of the source of the multi-source data of the vehicle active safety system, is the data category label;

[0107] S12. Construct a vehicle active safety system dataset based on the collected multi-source data of the vehicle active safety system, set a sampling time window, sort the collected multi-source data of the vehicle active safety system in chronological order, and form an original vehicle active safety system dataset ;

[0108] S13. Clean the data of the original vehicle active safety system dataset, remove invalid data, duplicate data, and abnormal data, and obtain the cleaned vehicle active safety system dataset;

[0109] S14. Perform data standardization processing on the cleaned vehicle active safety system dataset, and use the normalization method to convert the data to the interval , a standardized dataset of the vehicle active safety system is obtained;

[0110] S15. Remove redundant information from the standardized dataset of the vehicle active safety system, and calculate the feature correlation matrix of the multi-source data of the vehicle active safety system , and form the final standardized dataset of the vehicle active safety system .

[0111] In this embodiment, calculating the feature correlation matrix of the multi-source data of the vehicle active safety system is:

[0112] ;

[0113] Among them, represents the correlation coefficient between the th dimension feature and the th dimension feature, is the total number of data samples, and respectively represent the means of the kth data sample on the ith dimension feature and the jth dimension feature, and respectively represent the standardized values of the kth data sample on the ith dimension feature and the jth dimension feature. If the correlation coefficient exceeds the set redundancy threshold , then remove the high-redundancy dimension data to form the final standardized dataset of the vehicle active safety system:

[0114] ;

[0115] Among them, is the standardized dataset of the vehicle active safety system, is the multi-source data information of the vehicle active safety system after redundancy removal, is the final th multi-source data sample of the vehicle active safety system.

[0116] In this embodiment, S2 includes the following steps:

[0117] S21. Based on the standardized dataset of the vehicle active safety system, define the mapping function , and convert each standardized data sample of the vehicle active safety system into the corresponding data distribution node, where is the set of all data distribution nodes; is the mapping function from the data sample to the node;

[0118] S22. According to the data distribution node set Construct a set of direct connections between nodes , define a similarity evaluation function :

[0119] ;

[0120] Among them, represents the set of positive real numbers, and are the timestamps of the corresponding data samples respectively, and are positive real number weight coefficients. When the similarity evaluation function , it is considered that there is a direct transmission ability between node and node , and a direct connection is established to obtain the set of direct connections between nodes :

[0121] ;

[0122] Among them, is a preset similarity threshold;

[0123] S23. For each direct connection , define a transmission weight function to quantify the delay or cost of data transmission:

[0124] ;

[0125] Among them, represents the data transmission delay from node to node , represents the data transmission cost on this direct connection, and are the weight coefficients of the corresponding delay and cost respectively;

[0126] S24. Using the results of steps S21 to S23, construct the vehicle active safety system vertical architecture data distribution model as a graph model:

[0127] ;

[0128] Among them, represents the vehicle active safety system vertical architecture data distribution model. The vehicle active safety system vertical architecture data distribution model adopts a flattened data transmission structure. The node set represents all data acquisition and control execution units. The direct connection set and its transmission weight function enable data to reach the control execution module directly from the acquisition node.

[0129] This embodiment proposes a brand-new vertical architecture data distribution model, which breaks through the problems of excessive levels, long response time, and high computational redundancy in the traditional layered data architecture during data transmission. The vertical architecture data distribution model of the vehicle active safety system is based on the principle of constructing the shortest path from the data acquisition end to the execution end, combined with the data distribution topology optimization strategy, to ensure that the data distribution process in the vehicle active safety system does not pass through long hierarchical transmissions, but directly reaches the decision-making module and the execution unit, thereby realizing a low-latency and highly stable data transmission path. In addition, the introduction of the data node similarity evaluation mechanism and the data transmission weight calculation strategy can make the data flow more intelligent according to different network environments and actual situations, providing efficient data support for the subsequent path optimization algorithm.

[0130] In this embodiment, S3 includes the following steps:

[0131] S31. Initialize the fractional-order seagull population set based on the vertical architecture data distribution model M of the vehicle active safety system , and each fractional-order seagull individual is represented as:

[0132] ;

[0133] Among them, represents the th direct connection, with the path starting point being the acquisition node and the end point being the control execution module, represents the population size;

[0134] S32. Define the load balancing fitness function of the multi-source data transmission path of the vehicle active safety system :

[0135] ;

[0136] Among them, is the data transmission weight of the direct connection , k is the number of direct connections in the data transmission path of the vehicle active safety system, is the data transmission load of the th node in the path, is the average value of the data transmission loads of the nodes on the path, is the load balancing weight coefficient, enabling the path to take into account both the shortest transmission and the load balancing performance at the same time;

[0137] S33. Construct the fractional-order seagull position update that adapts to the network state of the vehicle active safety system as:

[0138] ;

[0139] Among them, and are respectively the th and th iteration fractional-order seagull individual path positions, and are the position update step size and the global optimal guiding coefficient, is the adaptive dynamic fractional-order order of the vehicle active safety system network state, is the gamma function, is the dynamic adjustment factor of the vehicle active safety system real-time network load;

[0140] S34. Define the global optimal vehicle active safety system data transmission path update formula:

[0141] ;

[0142] Among them, is the optimal path position in the th iteration process;

[0143] S35. Based on the vehicle active safety system network state dynamic feedback mechanism, adjust the dynamic fractional-order order and the network load dynamic adjustment factor in real time:

[0144] ;

[0145] ;

[0146] Among them, and are respectively the minimum and maximum boundaries of the fractional-order order, and are respectively the minimum and maximum boundaries of the network load dynamic adjustment factor, is the current real-time network load value of the vehicle active safety system, and are the historical minimum and maximum values of the network load, is the current real-time network delay value of the vehicle active safety system, and are the historical minimum and maximum values of the network delay;

[0147] S36. Repeat steps S33 to S35 until the iteration stop condition is reached, and obtain the initially optimized vehicle active safety system data transmission path set :

[0148] ;

[0149] Among them, represents the optimal solution of the data transmission path of the vehicle active safety system after the -th fractional-order seagull individual is updated by adaptive dynamic fractional-order optimization.

[0150] In this embodiment, a globally optimized data transmission path of the vehicle active safety system is constructed. The fractional-order seagull optimization algorithm is used for global search and optimization to find the globally optimal data transmission path set. Compared with the traditional optimization algorithm, the fractional-order seagull optimization of the present invention introduces the memory characteristics and long-range correlation of fractional-order calculus, enabling the algorithm to have both global search ability and local convergence ability during the search process, and avoiding the problem of falling into local optimum. Through the dynamic adaptive fractional-order adjustment mechanism, the fractional-order seagull optimization algorithm can adjust the optimization strategy in real time according to the changes in the vehicle networking environment (bandwidth fluctuation, data congestion, node failure), ensuring that the data distribution path always remains optimal, improving the global optimization ability, stability and dynamic adaptability of the data transmission path, and providing more reliable data support for the vehicle active safety system.

[0151] In this embodiment, S4 includes the following steps:

[0152] S41. Initialize the wolf pack individual set based on the initially optimized data transmission path set of the vehicle active safety system Each wolf pack individual is represented as: ,

[0153] ;

[0154] Among them, represents the initial search position of the -th wolf pack individual, is the wolf pack population size, is the -th direct connection in the initially optimized data transmission path of the vehicle active safety system;

[0155] S42. Define the refined fitness function of the vehicle active safety system path :

[0156] ;

[0157] Among them, represents the data transmission weight of the direct connection , is the dynamic stability coefficient of the direct connection in the real-time network environment, is the real-time network load imbalance coefficient of the direct connection , represents the real-time data transmission bandwidth of the direct connection . is a positive real - valued weight coefficient;

[0158] S43. Construct a mechanism for updating the positions of the wolf - pack's dynamic memory tracking behavior to implement a refined search and optimization of the data transmission path of the vehicle active safety system:

[0159] ;

[0160] Among them, and are the positions of the wolf - pack individuals at the th and th iterations respectively, is the position of the wolf - pack leader in the current iteration, representing the real - time local optimal path, is the memory position of the historical best path of the wolf - pack individual, is the individual dynamic memory weight coefficient, , are the adjustment coefficients of the current optimal path guidance and historical path memory respectively;

[0161] S44. Define a mechanism for real - time updating of the wolf - pack leader position to achieve local optimization guidance of the data transmission path of the vehicle active safety system in a dynamic network environment:

[0162] ;

[0163] Among them, is the real - time optimal path position determined based on the refined fitness function in the th iteration;

[0164] S45. Construct an adaptive wolf - pack intelligent reconnaissance mechanism for the network environment of the vehicle active safety system and define an update formula for the wolf - pack reconnaissance behavior:

[0165] ;

[0166] Among them, represents the real - time optimal path position after reconnaissance, , are the random exploration coefficient and the optimal guidance coefficient of the reconnaissance behavior respectively, is the dynamic random perturbation coefficient of the reconnaissance behavior, which is adjusted according to the real - time network fluctuation amplitude of the vehicle active safety system, is the dynamic sensitivity coefficient of the wolf - pack individual's reconnaissance, which changes dynamically with the path stability and load fluctuation, is a randomly selected position of the wolf - pack individual;

[0167] S46. Repeat steps S43 to S45 until the local optimization iteration stop condition is reached to generate the finally optimized data transmission path of the vehicle active safety system:

[0168] ;

[0169] Among them, is the optimal path position for the final data transmission of the vehicle active safety system.

[0170] Based on S3, in this embodiment, S4 further introduces an improved wolf pack search algorithm to locally optimize the data transmission path. Thus, on the premise of ensuring the global optimum, the local adaptability and load balancing ability of data transmission are further improved. The vehicle networking environment of the vehicle active safety system is highly dynamic, the data flow changes at any time, and the network congestion situation cannot be fully predicted. Therefore, relying solely on the globally optimized path planning is difficult to adapt to complex environments. Through the dynamic memory tracking mechanism and adaptive reconnaissance behavior update mechanism of the wolf pack search, the data transmission path can be adjusted at any time, reducing local congestion and improving the balance of the data stream without increasing the overall path calculation overhead. This enables the data transmission of the vehicle active safety system to still maintain low latency, low congestion, and high throughput in a highly dynamic environment, ensuring that critical safety information can be transmitted to the execution module in the first time and improving the real-time response ability of the active safety system.

[0171] Example 1:

[0172] On March 15, 2024, the test of the present invention was carried out at a certain intelligent networked vehicle test base. The test scenarios included 15 safety events such as emergency braking response, avoidance of sudden obstacles, and optimization of the lane keeping system. The goal was to verify the improvement effects of the present invention in terms of data distribution architecture optimization, data transmission path optimization, and real-time response speed.

[0173] During the test, the experimenters deployed two sets of systems on 80 test vehicles. One group adopted the traditional layered data distribution architecture, and the other group adopted the flattened data distribution architecture of the present invention combined with the fractional-order seagull optimization and wolf pack search algorithms. All test vehicles were equipped with high-precision lidar, millimeter-wave radar, front-view camera, ultrasonic sensor, and V2X communication module, and were connected to the 5G-V2X dedicated network of the test base, with a data transmission speed of 1 Gbps.

[0174] Case 1: High-speed emergency braking response test:

[0175] At 13:37:24 on March 15, 2024, on the G60 highway section of the test road, a test vehicle (number T-A12) was traveling at a speed of 120 km / h. When a vehicle suddenly cut in and occupied the lane 180 meters ahead, the system needed to complete data collection, path optimization, and braking instruction transmission in the shortest time to ensure safe braking.

[0176] Response situation of the traditional method:

[0177] At 13:37:24.015, the vehicle ahead suddenly changed lanes. The millimeter-wave radar and camera of T-A12 detected an anomaly, and the data was transmitted to the perception layer.

[0178] At 13:37:24.028, the perception layer transmitted the data to the fusion layer for comprehensive data analysis.

[0179] At 13:37:24.041, the fusion layer transmitted the data to the decision-making layer to calculate the braking strategy.

[0180] At 13:37:24.057, the decision-making layer sent a braking instruction to the execution layer. After receiving the signal, the braking system executed the braking action.

[0181] Total delay: 42 ms, the braking distance of the vehicle was 4.1 meters, and it almost rear-ended the vehicle ahead.

[0182] Response of the method of the present invention:

[0183] At 13:37:24.015, when the vehicle ahead changed lanes, the millimeter-wave radar and camera data were directly transmitted to the decision-making layer without hierarchical transfer.

[0184] At 13:37:24.021, the fractional-order seagull optimization algorithm dynamically calculated the optimal data transmission path and optimized the transmission stability in combination with the wolf pack search algorithm.

[0185] At 13:37:24.029, the data reached the execution layer directly, triggering the early intervention of the braking system.

[0186] Total delay: 14 ms, the braking distance of the vehicle was 1.5 meters, and the rear-end collision accident was successfully avoided.

[0187] Case 2: Avoiding sudden obstacles on urban roads:

[0188] On March 15, 2024, at 14:12:45, on the urban section of the test site, when the test vehicle (number T-C37) was driving at a speed of 50 km / h to an intersection, a pedestrian suddenly entered the lane. The system needed to calculate the optimal lane-changing path in the shortest time and control the vehicle to change lanes to avoid the pedestrian.

[0189] Response of the traditional method:

[0190] At 14:12:45.002, the forward-looking camera detected that the pedestrian entered, and the data was uploaded to the fusion layer.

[0191] At 14:12:45.018, after analysis, the fusion layer transmitted the data to the decision-making layer to calculate the lane-changing path.

[0192] At 14:12:45.037, the decision-making layer sent the instruction to the control module to execute the steering adjustment.

[0193] Total delay: 35 ms. The vehicle lane change delay almost caused a collision in the left lane.

[0194] Response of the method of the present invention:

[0195] At 14:12:45.002, the camera data was directly transmitted to the decision-making layer, omitting the intermediate fusion layer.

[0196] At 14:12:45.008, the fractional-order seagull optimization algorithm calculated the optimal lane change trajectory within 7 ms.

[0197] At 14:12:45.014, the data directly reached the execution layer, and the lane change operation was completed 23 ms in advance.

[0198] Total delay: 14 ms. The vehicle changed lanes smoothly, avoided a collision with a pedestrian, and maintained a safe distance from the vehicle on the left at the same time.

[0199] Case 3: Braking on a wet and slippery road surface at night with low visibility:

[0200] On March 15, 2024, at 21:45:12, the test vehicle (number T-N58) was driving on the urban main road in a low-illumination environment (<50 lux) at a speed of 65 km / h. Due to the wet road surface, the vehicle in front skidded out of control, and the test vehicle had to immediately reduce its speed to ensure safety.

[0201] Response of the traditional method:

[0202] At 21:45:12.032, the camera and millimeter-wave radar detected an abnormality of the vehicle in front.

[0203] At 21:45:12.048, the data was transmitted to the fusion layer for processing to generate a risk warning.

[0204] At 21:45:12.061, the decision-making layer calculated the optimal braking strategy and sent an instruction to the execution layer.

[0205] Total delay: 49 ms. The vehicle completed braking at a distance of 2.8 meters from the out-of-control vehicle and still had a minor scrape.

[0206] Response of the method of the present invention:

[0207] At 21:45:12.032, the sensor data was directly transmitted to the decision-making layer, reducing the fusion calculation time.

[0208] At 21:45:12.039, the fractional-order optimization was combined to dynamically adjust the data transmission path, and the transmission time was reduced to 7 ms.

[0209] At 21:45:12.047, the data directly reached the execution layer, and the braking intervened in advance.

[0210] Total delay: 15 ms. The vehicle brakes suddenly in advance, maintaining a safe distance of 0.8 meters from the vehicle ahead to avoid scratching.

[0211] And the experimenter made data comparison and statistical analysis:

[0212] Test Items Traditional Layered Architecture (Latency ms) Method of the Present Invention (Latency ms) Optimization Rate Actual Effect High-Speed Emergency Braking 42ms 14ms 66.7% Avoid Rear-End Collision Sudden Obstacle Avoidance 35ms 14ms 60.0% Smooth Lane Change Low Visibility Braking at Night 49ms 15ms 69.4% Avoid Scratching

[0213] In the test of this embodiment, the experimenter verified the advantages of the data distribution method of the vertical architecture of the vehicle active safety system of the present invention in terms of data transmission efficiency, path optimization calculation, and emergency response through real road events. The test data shows that:

[0214] 1. The average data transmission delay is reduced by 63.7%, effectively improving the real-time performance of the active safety system;

[0215] 2. After optimizing the data transmission path, the information reaches the execution layer directly, avoiding the problem of hierarchical data lag in the traditional architecture;

[0216] 3. Combining with the intelligent optimization algorithm, the adaptability of data transmission is improved, ensuring that the optimal path transmission can still be maintained in a complex in-vehicle network environment.

[0217] In summary, the technical solution of the present invention effectively improves the emergency response ability of the vehicle active safety system, ensuring that fast and accurate safety protection can still be provided in high-speed, urban roads, and low visibility environments at night.

[0218] The present invention adopts a vertical architecture, directly establishing the shortest path from the data acquisition node to the execution module, avoiding the data transmission delay caused by multi-level transfer. By optimizing the data distribution topology structure, the intermediate links of data transmission are significantly reduced, and the data transmission time in case of emergency is shortened to less than 60% of the traditional layered architecture, effectively improving the reaction speed of the active safety system.

[0219] The present invention combines the fractional-order seagull optimization algorithm in the process of optimizing the data distribution path. By using the memory characteristics and long-range correlation of fractional-order calculus, it can efficiently search for the global optimal data transmission path in a complex dynamic network environment. Through the dynamic fractional-order seagull flight model, the optimization algorithm takes into account both local search ability and global search ability. In a complex dynamic in-vehicle network environment, it can improve the adaptive optimization ability of the data transmission path, improve the adaptability of path planning, and still maintain high path stability in case of sudden change of network load.

[0220] The present invention combines the improved wolf pack search algorithm to locally optimize the preliminarily optimized path, focusing on the real-time load balancing problem of the data transmission path. By introducing a dynamic memory tracking mechanism and an adaptive reconnaissance mechanism, it ensures that while the data transmission path is the optimal short path, it has a low network congestion rate.

[0221] As described above, it is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes equivalent substitutions or changes, and should be covered within the protection scope of the present invention.

Claims

1. A vehicle control system, characterized in that, It includes the following modules: The multi-source data acquisition module of the vehicle active safety system is used to obtain multi-source data involved in the operation of the vehicle active safety system and generate a standardized data set of the vehicle active safety system; The vertical architecture data distribution construction module of the vehicle active safety system is used to construct a vertical architecture data distribution model of the vehicle active safety system based on the standardized data set of the vehicle active safety system; The vertical architecture data distribution construction module of the vehicle active safety system includes: Based on the standardized dataset of vehicle active safety systems , define a mapping function , and convert each standardized data sample of the vehicle active safety system into a corresponding data distribution node, where is the set of all data distribution nodes; is the mapping function from data samples to nodes; According to the data distribution node set Construct the direct connection set between nodes , define the similarity evaluation function : ； Among them, is the i-th multi-source data information in the vehicle active safety system after redundancy removal, is the j-th multi-source data information in the vehicle active safety system after redundancy removal, represents the set of positive real numbers, and are the timestamps of the corresponding data samples respectively, and are positive real weight coefficients. When the similarity evaluation function , it is considered that there is a direct transmission ability between node and node , and a direct connection is established to obtain the set of direct connections between nodes: ； Among them, is a preset similarity threshold; For each direct connection , define a transmission weight function to quantify the latency or cost of data transmission: ； Among them, represents the data transmission delay from node to node ; represents the data transmission cost on this direct connection, represents the data transmission cost on this direct connection, and are the weight coefficients corresponding to the delay and cost respectively; Construct the vertical architecture data distribution model of the vehicle active safety system into a graph model: ； Among them, represents the vertical architecture data distribution model of the vehicle active safety system. The vertical architecture data distribution model of the vehicle active safety system adopts a flat data transmission structure, and the node set represents all data acquisition and control execution units, and the direct connection set and its transmission weight function enable data to directly reach the control execution module from the acquisition node; The fractional-order seagull optimization global path search module is used to globally search and optimize the data transmission path based on the vertical architecture data distribution model of the vehicle active safety system by using the fractional-order seagull optimization algorithm to form a set of preliminarily optimized data transmission paths of the vehicle active safety system; The wolf pack search local path optimization module is used to locally optimize the preliminarily optimized data transmission path of the vehicle active safety system by using the wolf pack search algorithm to generate the finally optimized data transmission path of the vehicle active safety system; The data distribution execution module of the vehicle active safety system is used to execute the data transmission task of the vehicle active safety system according to the finally optimized data transmission path of the vehicle active safety system.

2. A vehicle control method, applied to the vehicle control system according to claim 1, characterized in that It includes the following steps: S1. Collect multi-source data of the vehicle active safety system, and preprocess the collected multi-source data of the vehicle active safety system to generate a standardized data set of the vehicle active safety system; S2. Construct a vertical architecture data distribution model of the vehicle active safety system based on the standardized data set of the vehicle active safety system, and the vertical architecture data distribution model of the vehicle active safety system adopts a flattened data transmission structure; The S2 includes the following steps: S21. Based on the standardized dataset of the vehicle active safety system , define a mapping function , and convert each standardized data sample of the vehicle active safety system into a corresponding data distribution node, where is the set of all data distribution nodes; is the mapping function from the data sample to the node; S22. According to the data distribution node set Construct a direct connection set between nodes , define a similarity evaluation function :[[]]END]] ； Among them, represents the set of positive real numbers, and are the timestamps of the corresponding data samples respectively, and are positive real weight coefficients. When the similarity evaluation function is satisfied, it is considered that there is a direct transmission ability between node and node , and a direct connection is established to obtain the set of direct connections between nodes : ； Among them, is a preset similarity threshold; S23. For each direct connection , define a transmission weight function to quantify the latency or cost of data transmission: ； Among them, represents the data transmission delay from node to node ; represents the data transmission cost on this direct connection, and are the weight coefficients corresponding to the delay and cost respectively;​ S24. Use the results of steps S21 to S23 to construct the vertical architecture data distribution model of the vehicle active safety system into a graph model: ； Among them, represents the vertical architecture data distribution model of the vehicle active safety system. The vertical architecture data distribution model of the vehicle active safety system adopts a flat data transmission structure, and the node set represents all data acquisition and control execution units, and the direct connection set and its transmission weight function enable data to directly reach the control execution module from the acquisition node; S3. Use the fractional-order seagull optimization algorithm to globally search and optimize each multi-source data transmission path in the vertical architecture data distribution model of the vehicle active safety system to form a set of preliminarily optimized data transmission paths of the vehicle active safety system; S4. Use the wolf pack search algorithm to locally search and optimize the set of preliminarily optimized data transmission paths of the vehicle active safety system to refine the data transmission path of the vehicle active safety system and generate the finally optimized data transmission path of the vehicle active safety system; S5. Implement real-time data transmission during the operation of the vehicle active safety system based on the finally optimized data transmission path of the vehicle active safety system, and dynamically collect the network status and operation status information of the vehicle active safety system to monitor and dynamically adjust the data transmission path of the vehicle active safety system in real time.

3. The vehicle control method according to claim 2, characterized in that, The S1 includes the following steps: S11. Collect multi-source data of the vehicle active safety system. The multi-source data of the vehicle active safety system includes data from in-vehicle sensors, environment perception devices, vehicle control modules, and network communication modules. Each multi-source data sample of the vehicle active safety system is represented as: ； Among them, is the th multi-source data sample of the vehicle active safety system, represents the timestamp of the multi-source data sample of the vehicle active safety system, represents the numerical information of the multi-source data sample of the vehicle active safety system, is the device identifier of the multi-source data source of the vehicle active safety system, is the data category label; S12. Construct a vehicle active safety system dataset based on the multi-source data collected from the vehicle active safety system. Set a sampling time window, sort the multi-source data of the vehicle active safety system collected in chronological order to form an original vehicle active safety system dataset ; S13. Clean the original vehicle active safety system dataset, remove invalid data, duplicate data and abnormal data, and obtain the cleaned vehicle active safety system dataset; S14. Perform data standardization processing on the dataset of the vehicle active safety system after cleaning, and use the normalization method to convert the data to the interval , and obtain the standardized dataset of the vehicle active safety system; S15. Remove redundant information from the standardized vehicle active safety system dataset and calculate the feature correlation matrix of multi-source data of the vehicle active safety system , forming the final standardized dataset of the vehicle active safety system .

4. The vehicle control method according to claim 3, wherein, The characteristic correlation matrix for calculating multi-source data of the vehicle active safety system is as follows: ； Among them, represents the correlation coefficient between the -dimensional feature and the -dimensional feature. is the total number of data samples. and respectively represent the means of the k-th data sample on the i-th feature and the j-th feature. and respectively represent the normalized values of the k-th data sample on the i-th feature and the j-th feature. If the correlation coefficient exceeds the set redundancy threshold , the high redundancy dimensional data is removed to form the final standardized data set of the vehicle active safety system: ； Among them, is the standardized dataset of the vehicle active safety system, is the multi-source data information of the vehicle active safety system after redundancy removal, is the final multi-source data sample of the vehicle active safety system.

5. A vehicle control method according to claim 4, wherein The said S3 includes the following steps: S31. Initialize the fractional-order seagull population set based on the data distribution model M of the vertical architecture of the vehicle active safety system , each fractional-order seagull individual is expressed as: ； Among them, represents the th direct connection, with the starting point of the path being the acquisition node and the ending point being the control execution module, represents the population size; S32. Define the load - balancing fitness function for the multi - source data transmission path of the vehicle active safety system : ； Among them, is the data transmission weight for direct connection k is the number of direct connections in the data transmission path of the vehicle active safety system, is the data transmission load of the th node in the path, is the average value of the data transmission loads of the nodes on the path, is the load balancing weight coefficient, enabling the path to consider both the shortest transmission and load balancing performance at the same time; S33. Construct the fractional-order seagull position update for the adaptive network state of the vehicle active safety system as: ； Among them, and are respectively the -th and the -th iteration fractional-order seagull individual path positions, and are the position update step size and the global optimal guiding coefficient, is the adaptive dynamic fractional-order order of the vehicle active safety system network state, is the gamma function, is the real-time network load dynamic regulation factor of the vehicle active safety system; S34. Define the update formula for the globally optimal data transmission path of the vehicle active safety system: ； Among them, is the optimal path position in the th iteration process; S35. Based on the dynamic feedback mechanism of the vehicle active safety system network status, the dynamic fractional order is adjusted in real time and the dynamic adjustment factor of network load : ； ； Among them, and are the minimum and maximum boundaries of the fractional order respectively, and are the minimum and maximum boundaries of the network load dynamic adjustment factor respectively, is the current real-time network load value of the vehicle active safety system, and are the historical minimum and maximum values of the network load, is the current real-time network delay value of the vehicle active safety system, and are the historical minimum and maximum values of the network delay; S36. Repeat steps S33 to S35 until the iteration stop condition is reached, and obtain a set of preliminarily optimized data transmission paths for the vehicle active safety system : ； Among them, represents the optimal solution of the data transmission path of the vehicle active safety system after the update of the adaptive dynamic fractional-order optimization of the fractional-order seagull individual.

6. The vehicle control method according to claim 2, characterized in that, The said S4 includes the following steps: S41. Based on the initially optimized data transmission path set of the vehicle active safety system Initialize the set of wolf pack individuals , and each wolf pack individual is represented as: ； Among them, represents the initial search position of the th individual in the wolf pack, is the size of the wolf pack population, is the th direct connection in the data transmission path of the preliminarily optimized vehicle active safety system; S42. Define the refined fitness function of the vehicle active safety system path : ； Among them, represents the data transmission weight of the direct connection . is the direct connection The dynamic stability coefficient in the real-time network environment is the direct connection The real-time network load imbalance coefficient of represents the direct connection The real-time data transmission bandwidth of is a positive real number weight coefficient; S43. Construct the wolf pack dynamic memory tracking behavior position update mechanism to implement the refined search and optimization of the vehicle active safety system data transmission path: ； Among them, and are respectively the positions of the wolf pack individuals at the -th and the -th iterations. is the position of the wolf pack leader in the current iteration, representing the real-time local optimal path. is the memory position of the historical best path of the wolf pack individuals. is the individual dynamic memory weight coefficient. , are respectively the adjustment coefficients of the current optimal path guidance and the historical path memory. S44. Define the real-time update mechanism for the wolf pack leader position to achieve the local optimization guidance of the vehicle active safety system data transmission path in a dynamic network environment: ； Among them, is the real-time optimal path position determined based on the refined fitness function in the S45. Construct the wolf pack intelligent reconnaissance mechanism adaptable to the vehicle active safety system network environment and define the update formula for the wolf pack reconnaissance behavior: ； Among them, represents the real-time optimal path position after reconnaissance, , are respectively the random exploration coefficient and the optimal guidance coefficient of the reconnaissance behavior, is the dynamic random perturbation coefficient of the reconnaissance behavior, which is adjusted according to the real-time network fluctuation amplitude of the vehicle active safety system, is the dynamic sensitivity coefficient of the wolf pack individual reconnaissance, which changes dynamically with the path stability and load fluctuation, is to randomly select the position of the wolf pack individual; S46. Repeat steps S43 to S45 until the local optimization iteration stop condition is reached, and generate the finally optimized vehicle active safety system data transmission path: ； Among them, is the optimal path position for the final data transmission of the vehicle active safety system.

Images