A multi-sensor fusion positioning method and system based on extended Kalman filter

Abstract

The present invention relates to a multi-sensor fusion positioning method and system based on an extended Kalman filter. The method comprises: R1. When a vehicle is traveling on a road, point cloud data information of the road is acquired in real time using an on-board laser radar, image data information of the road is acquired in real time using an on-board camera, absolute positioning data information of the vehicle is acquired in real time using an on-board integrated navigation system, posture data information of the vehicle is acquired in real time using an on-board IMU, and data information of the vehicle speed and wheel angle is acquired in real time using an on-board speedometer; R2. Based on the point cloud data information and image data information of the road, a fusion algorithm of an Archimedean-optimized neural network model is used to fuse the point cloud and image of the road. The present invention not only provides more robust navigation estimation than using a single sensor by combining complementary or redundant information from different sensors, but also provides stable and reliable positioning information in all operational scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-sensor fusion positioning technology, and in particular to a multi-sensor fusion positioning method and system based on extended Kalman filtering. Background Art

[0002] When operating, unmanned container trucks need to accurately stop at designated locations in the yard and under the quay cranes, where they wait for loading and unloading operations, requiring positioning accuracy at the centimeter level. High-precision positioning can ensure accurate parking of vehicles, improve handling efficiency, and reduce road congestion accidents caused by braking. Existing positioning methods mainly include combined navigation based on observation satellite signals, map matching positioning based on high-precision maps, laser SLAM using lidar, and visual SLAM using cameras. Due to the characteristics of the port environment such as complex road structure, frequent environmental changes, non-fixed operating routes, and mixed operations of manned and unmanned container trucks, the use of a single sensor and positioning source cannot guarantee continuous, stable, and reliable high-precision positioning in all operational scenarios. Therefore, the present invention comprehensively considers the particularity of the port environment and proposes a multi-sensor fusion positioning method to realize the fusion processing of observation data from different sensors, providing accurate and reliable positioning information for downstream modules such as environmental perception, planning and collision avoidance of unmanned container trucks.

[0003] Prior art Chinese patents (Application Number: 202311571021.4, Publication Number: CN 117584989A) disclose a tightly coupled SLAM system and algorithm using a LiDAR / IMU / vehicle kinematic constraints. This system first constructs vehicle kinematic constraints using the IMU, rear axle wheel speed, and front wheel steering angle. By decoupling the displacement and attitude information of the vehicle kinematic constraints, the system constructs displacement and attitude constraints separately, thereby improving the accuracy of the optimization results. Secondly, an adaptive coefficient is introduced based on the number of feature points and steering angle to adjust the weights of the vehicle kinematic constraints in real time. However, this is not fully applicable to port environments, where unmanned container trucks transport cargo in containers between cargo ships and container stacking areas. Within the container stacking area, containers are arranged in an orderly and aligned manner, and in some areas, the number of containers on both sides is exactly the same. For laser SLAM, this structured scenario offers limited feature information, and the lack of sufficient constraints can easily lead to degradation, making it difficult to guarantee the continuity and reliability of positioning information. Summary of the Invention

[0004] In view of the above shortcomings of the existing technology, the present invention provides a multi-sensor fusion positioning method and system based on extended Kalman filtering, which not only provides more robust navigation estimation than using a single sensor by combining complementary or redundant information from different sensors, but also provides stable and reliable positioning information in all operating scenarios.

[0005] In order to achieve the above-mentioned and other related purposes, the present invention provides a technical solution as follows: a multi-sensor fusion positioning method based on extended Kalman filtering, the method comprising:

[0006] R1. When a vehicle is driving on a road, it acquires real-time point cloud data of the road using the onboard lidar, image data of the road using the onboard camera, absolute positioning data of the vehicle using the onboard integrated navigation, posture data of the vehicle using the onboard IMU, and speed and wheel angle data of the vehicle using the onboard speedometer.

[0007] R2. Based on the point cloud data information of the road and the image data information of the road, the point cloud and image of the road are fused using a fusion algorithm of an Archimedean optimized neural network model to obtain fused road status data information;

[0008] R3. Based on the absolute positioning data information of the vehicle, the posture data information of the vehicle, and the data information of the vehicle speed and wheel angle, the vehicle motion state data is fused using a multimodal feature fusion algorithm based on the zebra optimization algorithm to obtain the fused vehicle state data information;

[0009] R4. Based on the fused vehicle state data and the fused road state data, an extended Kalman filter algorithm incorporating a golden sine factor is used to fuse the vehicle and road state data, and a fused state matrix is ​​constructed to obtain fused vehicle and road state matrix data information;

[0010] R5. Based on the fusion state matrix data information of the vehicle and the road, construct the vehicle fusion positioning function W, characterize the vehicle fusion positioning data, and obtain the vehicle fusion positioning data information.

[0011] Furthermore, the fusion positioning function W of the vehicle is:

[0012]

[0013] Among them, x is the fusion state matrix data information of the vehicle and the road, and α1, α2 and α3 are the fusion positioning factors of the vehicle.

[0014] Furthermore, the constraint function f of the fusion positioning factors α1, α2 and α3 of the vehicle is,

[0015]

[0016] Among them, the value range of the constraint function f is (2,3).

[0017] Furthermore, in step R2, the fusion algorithm of the Archimedean optimized neural network model is used to fuse the point cloud and the image of the road, including:

[0018] R21. Based on the point cloud data information of the road, construct a point cloud parameter matrix of the road, based on the image data information of the road, construct an image parameter matrix of the road, and obtain the parameter matrix data information of the point cloud and image of the road;

[0019] R22. Based on the parameter matrix data information of the point cloud and image of the road, construct the Archimedean target optimization function G,

[0020]

[0021] Where y1 is the parameter matrix data information of the road point cloud, y2 is the parameter matrix data information of the road image, β1, β2 and β3 are the target optimization factors of Archimedes, and the point cloud and image parameters of the road are optimized to obtain the optimized point cloud and image parameter data information of the road;

[0022] R23. Input the optimized road point cloud and image parameter data information into the neural network model for training and learning, and determine the feature fusion function H.

[0023]

[0024] Among them, z1 is the point cloud parameter data information of the optimized road, z2 is the image parameter data information of the optimized road, δ1, δ2 and δ3 are the optimization factors of the point cloud and image parameters of the road. The point cloud and image of the road are fused to obtain the state data information of the fused road.

[0025] Furthermore, the constraints of the optimization factors δ1, δ2 and δ3 of the point cloud and image parameters of the road are:

[0026]

[0027] The target optimization factors β1, β2 and β3 of Archimedes are:

[0028]

[0029] Among them, y1 is the parameter matrix data information of the point cloud of the road, and y2 is the parameter matrix data information of the image of the road.

[0030] Furthermore, in step R3, the multimodal feature fusion algorithm based on the zebra optimization algorithm is used to fuse the vehicle's motion state data, including:

[0031] R31. Based on the absolute positioning data information of the vehicle, the posture data information of the vehicle and the data information of the vehicle speed and wheel angle, establish the vehicle motion state relationship function L,

[0032]

[0033] Among them, a1 is the absolute positioning data information of the vehicle, a2 is the posture data information of the vehicle, a3 is the data information of the vehicle speed and wheel angle, γ1 and γ2 are relationship factors, which characterize the relationship between the vehicle's motion state data and obtain the relationship data information of the vehicle's motion state data;

[0034] R32 based on the relationship data information of the vehicle's motion state data, establish the relationship between the vehicle motion state data zebra optimization function Q,

[0035]

[0036] Wherein, b is the relationship data information of the vehicle's motion state data, η1, η2 and η3 are the relationship optimization factors of the vehicle's motion state, and the relationship of the vehicle's motion state data is optimized to obtain the optimized relationship data information of the vehicle's motion state data;

[0037] R33. Based on the optimized relationship data information of the vehicle motion state data, construct a multimodal fusion function P of the vehicle state,

[0038]

[0039] Among them, c is the relational data information of the optimized vehicle motion state data, λ1, λ2 and λ3 are the multimodal fusion feature factors of the vehicle motion state, and the vehicle motion state data is fused to obtain the fused vehicle state data information.

[0040] Furthermore, the constraint function g of the multimodal fusion characteristic factors λ1, λ2 and λ3 of the vehicle motion state is,

[0041]

[0042] Among them, the value range of the constraint function g is (1,2).

[0043] Furthermore, in step R4, the use of the extended Kalman filter algorithm integrated with the golden sine factor to fuse the vehicle and road state data includes:

[0044] R41. Based on the fused vehicle status data information and the fused road status data information, establish the state variable function S of the extended Kalman filter for vehicle fusion positioning,

[0045]

[0046] Among them, r1 is the state data information of the fused vehicle, r2 is the state data information of the fused road, A is the state matrix, and B is the control matrix. The state parameters of the extended Kalman filter of the vehicle fusion positioning are characterized to obtain the state parameter data information of the extended Kalman filter of the vehicle fusion positioning;

[0047] R42. Based on the fused vehicle status data information and the fused road status data information, establish the observation variable function R of the extended Kalman filter for vehicle fusion positioning,

[0048]

[0049] Among them, r1 is the state data information of the fused vehicle, r2 is the state data information of the fused road, and U is the observation matrix. The observation parameters of the extended Kalman filter of the vehicle fusion positioning are characterized to obtain the observation parameter data information of the extended Kalman filter of the vehicle fusion positioning;

[0050] R43 based on the observation parameter data information of the extended Kalman filter of the vehicle fusion positioning and the state parameter data information of the extended Kalman filter of the vehicle fusion positioning, establish a fusion function O,

[0051]

[0052] Among them, g1 is the observation parameter data information of the extended Kalman filter for vehicle fusion positioning, g2 is the state parameter data information of the extended Kalman filter for vehicle fusion positioning, μ1, μ2 and μ3 are the golden sine factors of the extended Kalman filter. The state data of the vehicle and the road are fused, and the fusion state matrix is ​​constructed to obtain the fusion state matrix data information of the vehicle and the road.

[0053] Furthermore, the golden sine factors μ1, μ2 and μ3 of the extended Kalman filter are,

[0054]

[0055] Among them, g1 is the observation parameter data information of the extended Kalman filter for vehicle fusion positioning, and g2 is the state parameter data information of the extended Kalman filter for vehicle fusion positioning.

[0056] In order to achieve the above-mentioned and other related purposes, the present invention also provides a computer-readable storage medium, which stores a computer program programmed or configured to execute any one of the multi-sensor fusion positioning methods based on extended Kalman filtering.

[0057] The present invention has the following positive effects:

[0058] 1. The present invention fuses the point cloud and image of the road by adopting a fusion algorithm based on an Archimedean-optimized neural network model, and fuses the vehicle's motion state data by combining a multimodal feature fusion algorithm based on a zebra optimization algorithm. This not only provides more robust navigation estimation than using a single sensor by combining complementary or redundant information from different sensors, but also provides stable and reliable positioning information in all operating scenarios.

[0059] 2. The present invention fuses the vehicle and road status data by adopting an extended Kalman filter algorithm integrated with a golden sine factor, and characterizes the vehicle's fused positioning data by combining it with the construction of a vehicle fusion positioning function W. The fused output positioning information includes not only global absolute positioning information but also local relative positioning information, which participates in operational processes such as container loading and unloading, container transfer between container areas and quay cranes, automatic parking, and charging, and provides stable and reliable positioning information for full-scenario operations in port environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 Schematic diagram of the method flow of the present invention;

[0061] Figure 2 A schematic flow chart of the fusion algorithm of the Archimedean optimized neural network model of the present invention;

[0062] Figure 3 Schematic diagram of the process of the multimodal feature fusion algorithm based on the zebra optimization algorithm of the present invention;

[0063] Figure 4 Schematic diagram of the flow of the extended Kalman filter algorithm integrated with the golden sine factor of the present invention. DETAILED DESCRIPTION

[0064] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0065] Example 1: Figure 1 As shown, a multi-sensor fusion positioning method based on extended Kalman filtering includes:

[0066] R1. When a vehicle is driving on a road, it acquires real-time point cloud data of the road using the onboard lidar, image data of the road using the onboard camera, absolute positioning data of the vehicle using the onboard integrated navigation, posture data of the vehicle using the onboard IMU, and speed and wheel angle data of the vehicle using the onboard speedometer.

[0067] R2. Based on the point cloud data information of the road and the image data information of the road, the point cloud and image of the road are fused using a fusion algorithm of an Archimedean optimized neural network model to obtain fused road status data information;

[0068] R3. Based on the absolute positioning data information of the vehicle, the posture data information of the vehicle, and the data information of the vehicle speed and wheel angle, the vehicle motion state data is fused using a multimodal feature fusion algorithm based on the zebra optimization algorithm to obtain the fused vehicle state data information;

[0069] R4. Based on the fused vehicle state data and the fused road state data, an extended Kalman filter algorithm incorporating a golden sine factor is used to fuse the vehicle and road state data, and a fused state matrix is ​​constructed to obtain fused vehicle and road state matrix data information;

[0070] R5. Based on the fusion state matrix data information of the vehicle and the road, construct the vehicle fusion positioning function W, characterize the vehicle fusion positioning data, and obtain the vehicle fusion positioning data information.

[0071] In this embodiment, the fusion positioning function W of the vehicle is:

[0072]

[0073] Among them, x is the fusion state matrix data information of the vehicle and the road, and α1, α2 and α3 are the fusion positioning factors of the vehicle.

[0074] In this embodiment, the constraint function f of the fusion positioning factors α1, α2 and α3 of the vehicle is,

[0075]

[0076] Among them, the value range of the constraint function f is (2,3).

[0077] In this embodiment, if Figure 2 As shown, in step R2, the fusion algorithm using the Archimedean optimized neural network model to fuse the point cloud and the image of the road includes:

[0078] R21. Based on the point cloud data information of the road, construct a point cloud parameter matrix of the road, based on the image data information of the road, construct an image parameter matrix of the road, and obtain the parameter matrix data information of the point cloud and image of the road;

[0079] R22. Based on the parameter matrix data information of the point cloud and image of the road, construct the Archimedean target optimization function G,

[0080]

[0081] Where y1 is the parameter matrix data information of the road point cloud, y2 is the parameter matrix data information of the road image, β1, β2 and β3 are the target optimization factors of Archimedes, and the point cloud and image parameters of the road are optimized to obtain the optimized point cloud and image parameter data information of the road;

[0082] R23. Input the optimized road point cloud and image parameter data information into the neural network model for training and learning, and determine the feature fusion function H.

[0083]

[0084] Among them, z1 is the point cloud parameter data information of the optimized road, z2 is the image parameter data information of the optimized road, δ1, δ2 and δ3 are the optimization factors of the point cloud and image parameters of the road. The point cloud and image of the road are fused to obtain the state data information of the fused road.

[0085] In this embodiment, the constraints of the optimization factors δ1, δ2, and δ3 of the point cloud and image parameters of the road are:

[0086]

[0087] The target optimization factors β1, β2 and β3 of Archimedes are:

[0088]

[0089] Among them, y1 is the parameter matrix data information of the point cloud of the road, and y2 is the parameter matrix data information of the image of the road.

[0090] Example 2: Based on the multi-sensor fusion positioning method based on extended Kalman filtering in Example 1, the present invention is further illustrated and described below.

[0091] like Figure 1 As shown, a multi-sensor fusion positioning method based on extended Kalman filtering includes:

[0092] R1. When a vehicle is driving on a road, it acquires real-time point cloud data of the road using the onboard lidar, image data of the road using the onboard camera, absolute positioning data of the vehicle using the onboard integrated navigation, posture data of the vehicle using the onboard IMU, and speed and wheel angle data of the vehicle using the onboard speedometer.

[0093] R2. Based on the point cloud data information of the road and the image data information of the road, the point cloud and image of the road are fused using a fusion algorithm of an Archimedean optimized neural network model to obtain fused road status data information;

[0094] R3. Based on the absolute positioning data information of the vehicle, the posture data information of the vehicle, and the data information of the vehicle speed and wheel angle, the vehicle motion state data is fused using a multimodal feature fusion algorithm based on the zebra optimization algorithm to obtain the fused vehicle state data information;

[0095] R4. Based on the fused vehicle state data and the fused road state data, an extended Kalman filter algorithm incorporating a golden sine factor is used to fuse the vehicle and road state data, and a fused state matrix is ​​constructed to obtain fused vehicle and road state matrix data information;

[0096] R5. Based on the fusion state matrix data information of the vehicle and the road, construct the vehicle fusion positioning function W, characterize the vehicle fusion positioning data, and obtain the vehicle fusion positioning data information.

[0097] In this embodiment, if Figure 3 As shown, in step R3, the multimodal feature fusion algorithm based on the zebra optimization algorithm is used to fuse the vehicle's motion state data, including:

[0098] R31. Based on the absolute positioning data information of the vehicle, the posture data information of the vehicle and the data information of the vehicle speed and wheel angle, establish the vehicle motion state relationship function L,

[0099]

[0100] Among them, a1 is the absolute positioning data information of the vehicle, a2 is the posture data information of the vehicle, a3 is the data information of the vehicle speed and wheel angle, γ1 and γ2 are relationship factors, which characterize the relationship between the vehicle's motion state data and obtain the relationship data information of the vehicle's motion state data;

[0101] R32 based on the relationship data information of the vehicle's motion state data, establish the relationship between the vehicle motion state data zebra optimization function Q,

[0102]

[0103] Wherein, b is the relationship data information of the vehicle's motion state data, η1, η2 and η3 are the relationship optimization factors of the vehicle's motion state, and the relationship of the vehicle's motion state data is optimized to obtain the optimized relationship data information of the vehicle's motion state data;

[0104] R33. Based on the optimized relationship data information of the vehicle motion state data, a multimodal fusion function P of the vehicle state is constructed.

[0105]

[0106] Among them, c is the relational data information of the optimized vehicle motion state data, λ1, λ2 and λ3 are the multimodal fusion feature factors of the vehicle motion state, and the vehicle motion state data is fused to obtain the fused vehicle state data information.

[0107] In this embodiment, the constraint function g of the multimodal fusion characteristic factors λ1, λ2 and λ3 of the vehicle motion state is,

[0108]

[0109]

[0110] Among them, the value range of the constraint function g is (1,2).

[0111] In this embodiment, if Figure 4 As shown, in step R4, the use of the extended Kalman filter algorithm integrated with the golden sine factor to fuse the vehicle and road state data includes:

[0112] R41. Based on the fused vehicle status data information and the fused road status data information, establish the state variable function S of the extended Kalman filter for vehicle fusion positioning,

[0113]

[0114] Among them, r1 is the state data information of the fused vehicle, r2 is the state data information of the fused road, A is the state matrix, and B is the control matrix. The state parameters of the extended Kalman filter of the vehicle fusion positioning are characterized to obtain the state parameter data information of the extended Kalman filter of the vehicle fusion positioning;

[0115] R42. Based on the fused vehicle status data information and the fused road status data information, establish the observation variable function R of the extended Kalman filter for vehicle fusion positioning,

[0116]

[0117] Among them, r1 is the state data information of the fused vehicle, r2 is the state data information of the fused road, and U is the observation matrix. The observation parameters of the extended Kalman filter of the vehicle fusion positioning are characterized to obtain the observation parameter data information of the extended Kalman filter of the vehicle fusion positioning;

[0118] R43 based on the observation parameter data information of the extended Kalman filter of the vehicle fusion positioning and the state parameter data information of the extended Kalman filter of the vehicle fusion positioning, establish a fusion function O,

[0119]

[0120] Among them, g1 is the observation parameter data information of the extended Kalman filter for vehicle fusion positioning, g2 is the state parameter data information of the extended Kalman filter for vehicle fusion positioning, μ1, μ2 and μ3 are the golden sine factors of the extended Kalman filter. The state data of the vehicle and the road are fused, and the fusion state matrix is ​​constructed to obtain the fusion state matrix data information of the vehicle and the road.

[0121] In this embodiment, the golden sine factors μ1, μ2 and μ3 of the extended Kalman filter are,

[0122]

[0123] Among them, g1 is the observation parameter data information of the extended Kalman filter for vehicle fusion positioning, and g2 is the state parameter data information of the extended Kalman filter for vehicle fusion positioning.

[0124] In this embodiment, the present invention provides a computer-readable storage medium storing a computer program programmed or configured to execute any one of the multi-sensor fusion positioning methods based on extended Kalman filtering.

[0125] In this embodiment, the present invention provides a computer-readable storage medium storing a computer program programmed or configured to execute any one of the multi-sensor fusion positioning methods based on extended Kalman filtering.

[0126] Any reference to memory, storage, database or other media used in the embodiments provided herein may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0127] In summary, the present invention not only provides more robust navigation estimation than using a single sensor by combining complementary or redundant information from different sensors, but also provides stable and reliable positioning information in all operational scenarios.

[0128] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A multi-sensor fusion positioning method based on extended Kalman filtering, characterized in that: The method comprises: R1. When a vehicle is driving on a road, it acquires real-time point cloud data of the road using the onboard lidar, image data of the road using the onboard camera, absolute positioning data of the vehicle using the onboard integrated navigation, posture data of the vehicle using the onboard IMU, and speed and wheel angle data of the vehicle using the onboard speedometer. R2. Based on the point cloud data information of the road and the image data information of the road, the point cloud and image of the road are fused using a fusion algorithm of an Archimedean optimized neural network model to obtain fused road status data information; R3. Based on the absolute positioning data information of the vehicle, the posture data information of the vehicle, and the data information of the vehicle speed and wheel angle, the vehicle motion state data is fused using a multimodal feature fusion algorithm based on the zebra optimization algorithm to obtain the fused vehicle state data information; R4. Based on the fused vehicle state data and the fused road state data, an extended Kalman filter algorithm incorporating a golden sine factor is used to fuse the vehicle and road state data, and a fused state matrix is ​​constructed to obtain fused vehicle and road state matrix data information; R5. Based on the fusion state matrix data information of the vehicle and the road, construct the vehicle fusion positioning function W, characterize the vehicle fusion positioning data, and obtain the vehicle fusion positioning data information.

2. The multi-sensor fusion positioning method based on extended Kalman filtering according to claim 1, characterized in that: The fusion positioning function W of the vehicle is, Among them, x is the fusion state matrix data information of the vehicle and the road, and α1, α2 and α3 are the fusion positioning factors of the vehicle.

3. The multi-sensor fusion positioning method based on extended Kalman filtering according to claim 2, characterized in that: The constraint function f of the vehicle's fusion positioning factors α1, α2 and α3 is, Among them, the value range of the constraint function f is (2,3).

4. The multi-sensor fusion positioning method based on extended Kalman filtering according to claim 1, characterized in that: In step R2, the fusion algorithm using the Archimedean optimized neural network model to fuse the point cloud and the image of the road includes: R21. Based on the point cloud data information of the road, construct a point cloud parameter matrix of the road, based on the image data information of the road, construct an image parameter matrix of the road, and obtain the parameter matrix data information of the point cloud and image of the road; R22. Based on the parameter matrix data information of the point cloud and image of the road, construct the Archimedean target optimization function G, Where y1 is the parameter matrix data information of the road point cloud, y2 is the parameter matrix data information of the road image, β1, β2 and β3 are the target optimization factors of Archimedes, and the point cloud and image parameters of the road are optimized to obtain the optimized point cloud and image parameter data information of the road; R23. Input the optimized road point cloud and image parameter data information into the neural network model for training and learning, and determine the feature fusion function H. Among them, z1 is the point cloud parameter data information of the optimized road, z2 is the image parameter data information of the optimized road, δ1, δ2 and δ3 are the optimization factors of the point cloud and image parameters of the road. The point cloud and image of the road are fused to obtain the state data information of the fused road.

5. The multi-sensor fusion positioning method based on extended Kalman filtering according to claim 4 is characterized in that: The constraints of the optimization factors δ1, δ2 and δ3 of the point cloud and image parameters of the road are: The target optimization factors β1, β2 and β3 of Archimedes are: Among them, y1 is the parameter matrix data information of the point cloud of the road, and y2 is the parameter matrix data information of the image of the road.

6. The multi-sensor fusion positioning method based on extended Kalman filtering according to claim 1, characterized in that: In step R3, the multimodal feature fusion algorithm based on the zebra optimization algorithm is used to fuse the vehicle's motion state data, including: R31. Based on the absolute positioning data information of the vehicle, the posture data information of the vehicle and the data information of the vehicle speed and wheel angle, establish the vehicle motion state relationship function L, Among them, a1 is the absolute positioning data information of the vehicle, a2 is the posture data information of the vehicle, a3 is the data information of the vehicle speed and wheel angle, γ1 and γ2 are relationship factors, which characterize the relationship between the vehicle's motion state data and obtain the relationship data information of the vehicle's motion state data; R32 based on the relationship data information of the vehicle's motion state data, establish the relationship between the vehicle motion state data zebra optimization function Q, Wherein, b is the relationship data information of the vehicle's motion state data, η1, η2 and η3 are the relationship optimization factors of the vehicle's motion state, and the relationship of the vehicle's motion state data is optimized to obtain the optimized relationship data information of the vehicle's motion state data; R33. Based on the optimized relationship data information of the vehicle motion state data, a multimodal fusion function P of the vehicle state is constructed. Among them, c is the relational data information of the optimized vehicle motion state data, λ1, λ2 and λ3 are the multimodal fusion feature factors of the vehicle motion state, and the vehicle motion state data is fused to obtain the fused vehicle state data information.

7. The multi-sensor fusion positioning method based on extended Kalman filtering according to claim 6, characterized in that: The constraint function g of the multimodal fusion characteristic factors λ1, λ2 and λ3 of the vehicle motion state is: Among them, the value range of the constraint function g is (1,2).

8. The multi-sensor fusion positioning method based on extended Kalman filtering according to claim 1, characterized in that: In step R4, the use of the extended Kalman filter algorithm integrated with the golden sine factor to fuse the vehicle and road state data includes: R41. Based on the fused vehicle status data information and the fused road status data information, establish the state variable function S of the extended Kalman filter for vehicle fusion positioning, Among them, r1 is the state data information of the fused vehicle, r2 is the state data information of the fused road, A is the state matrix, and B is the control matrix. The state parameters of the extended Kalman filter of the vehicle fusion positioning are characterized to obtain the state parameter data information of the extended Kalman filter of the vehicle fusion positioning; R42. Based on the fused vehicle status data information and the fused road status data information, establish the observation variable function R of the extended Kalman filter for vehicle fusion positioning, Among them, r1 is the state data information of the fused vehicle, r2 is the state data information of the fused road, and U is the observation matrix. The observation parameters of the extended Kalman filter of the vehicle fusion positioning are characterized to obtain the observation parameter data information of the extended Kalman filter of the vehicle fusion positioning; R43 based on the observation parameter data information of the extended Kalman filter of the vehicle fusion positioning and the state parameter data information of the extended Kalman filter of the vehicle fusion positioning, establish a fusion function O, Among them, g1 is the observation parameter data information of the extended Kalman filter for vehicle fusion positioning, g2 is the state parameter data information of the extended Kalman filter for vehicle fusion positioning, μ1, μ2 and μ3 are the golden sine factors of the extended Kalman filter. The state data of the vehicle and the road are fused, and the fusion state matrix is ​​constructed to obtain the fusion state matrix data information of the vehicle and the road.

9. The multi-sensor fusion positioning method based on extended Kalman filtering according to claim 8, characterized in that: The golden sine factors μ1, μ2 and μ3 of the extended Kalman filter are, Among them, g1 is the observation parameter data information of the extended Kalman filter for vehicle fusion positioning, and g2 is the state parameter data information of the extended Kalman filter for vehicle fusion positioning.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program that is programmed or configured to execute the multi-sensor fusion positioning method based on extended Kalman filtering as described in any one of claims 1 to 9.

Images