Positioning data processing method and device, cloud equipment and vehicle

Abstract

The invention relates to the technical field of vehicles, in particular to a positioning data processing method and device, cloud equipment and a vehicle, and the method comprises the steps: obtaining positioning data packets of all target vehicles; classifying the positioning data packets of all the target vehicles according to the data acquisition time and the vehicle identifiers; the target number of the containers is determined according to the total number of the positioning data packets and the classification result, the containers with the target number are started, the configuration files and the positioning data packets are transmitted into the corresponding containers, the containers conduct fusion positioning processing on the positioning data packets based on the configuration files and the multi-thread technology, the cloud computing power is fully utilized, and the positioning efficiency is improved. Positioning related data are loaded and processed efficiently in a large scale, the processing speed is increased, and the positioning data accuracy is improved through a fusion method based on pose map optimization. Therefore, the problems that at present, vehicles are usually limited by local hardware computing power, large-scale rapid positioning data processing cannot be supported, and the precision of sensors of the vehicles is limited are solved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle technology, and in particular to a positioning data processing method, device, cloud equipment and vehicle. Background Art

[0002] In the development of autonomous driving technology, data collection and processing are extremely critical links. With the advancement of technology, more and more automobile companies and startups focusing on the field of autonomous driving have begun to use vehicles as mobile data collection platforms. These vehicles include not only collection vehicles specially designed for data collection, but also mass-produced models sold on the market. In this way, perception information (such as images, radar signals, etc.) and GPS positioning data under various driving conditions can be widely collected, providing rich materials for training more accurate autonomous driving algorithms.

[0003] However, for mass-produced models, data collection and processing face some challenges. The first is the limitation of local hardware computing power: mass-produced vehicles are usually equipped with computing resources required to meet basic driving assistance functions. Therefore, when performing real-time data processing, especially when it involves large-scale data sets or needs to execute complex algorithms, the existing on-board computing power may not be sufficient to support efficient data analysis tasks; the second is the problem of sensor accuracy: although modern cars are generally equipped with a variety of types of sensors, due to cost considerations or design compromises, the performance indicators of these sensors often cannot reach the optimal level. Summary of the invention

[0004] One of the purposes of the present invention is to provide a positioning data processing method to solve the problems that current vehicles are often limited by local hardware computing power, cannot support large-scale rapid positioning data processing, and the accuracy of their own sensors is also limited; the second purpose is to provide a device; the third purpose is to provide a cloud device; and the fourth purpose is to provide a vehicle.

[0005] In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention is as follows: a positioning data processing method, comprising the following steps: obtaining positioning data packets of all target vehicles; classifying the positioning data packets of all target vehicles according to data collection time and vehicle identification; determining the target number of containers according to the total number of positioning data packets and the classification results, opening the target number of containers, and transferring the configuration file and the positioning data packet to the corresponding container, wherein the container performs fusion positioning processing on the positioning data packet based on the configuration file and multi-threading technology.

[0006] According to the above-mentioned technical means, the present invention collects and classifies the positioning data packets of all target vehicles, utilizes the pose graph optimization fusion algorithm of container technology and multi-threading technology, transfers the configuration file and the positioning data packet into the corresponding container, and performs fusion positioning processing, thereby realizing fast and efficient processing of positioning data and improving the accuracy of positioning data.

[0007] Furthermore, the container performs fusion positioning processing on the positioning data packets based on the configuration file and multi-threading technology, including: identifying the target processing quantity in the configuration file; performing time-series segmentation and grouping on the positioning data packets to be processed by the container according to the target processing quantity; treating the segmented and grouped data as a fusion positioning task, placing the fusion positioning task into the thread pool of the container, and starting at least one thread to process the fusion positioning task of the thread pool.

[0008] According to the above technical means, the present invention can efficiently perform fusion positioning processing on positioning data packets, improve the speed and efficiency of data processing, and ensure the effective use of resources.

[0009] Furthermore, at least one thread is started to process the fusion positioning task of the thread pool, including: reading the position data, lane line data and map data in the data corresponding to the fusion positioning task; performing perception matching based on the lane line data and the map data to obtain a semantic positioning result; fusing the position data, the semantic positioning result and the track calculation data into a pose graph, and determining the fusion positioning result of the corresponding positioning data packet based on the fused pose graph.

[0010] According to the above technical means, the present invention constructs a fused pose graph, and the fused positioning result of each positioning data packet can be determined more accurately through the fused pose graph.

[0011] Furthermore, before the position data, the semantic positioning result and the dead reckoning data are integrated into the pose graph, the method further includes: if the dead reckoning data does not exist in the data corresponding to the fused positioning task, the wheel speed and angular velocity in the position data are identified, and the dead reckoning is performed based on the wheel speed and angular velocity.

[0012] According to the above technical means, the present invention can perform track calculation through the wheel speed and angular velocity in the position data, ensuring that even in the absence of direct track information, the positioning data set can be improved through supplementary calculations, thereby improving the accuracy and reliability of the positioning system.

[0013] Furthermore, before placing the fusion positioning task into the thread pool of the container and starting at least one thread to process the fusion positioning task of the thread pool, it also includes: determining the size of the thread pool according to the available CPU cores and memory; and determining the number of threads to be started according to the size of the thread pool and the target processing number.

[0014] According to the above technical means, it can be ensured that computing resources are optimally utilized, which not only avoids resource waste but also prevents performance degradation caused by excessive competition.

[0015] Furthermore, after the container performs fusion positioning processing on the positioning data packet based on the configuration file and the multi-threading technology, the container also includes: verifying the fusion positioning result of the positioning data packet; and optimizing the fusion positioning result according to the verification result.

[0016] According to the above technical means, the present invention can optimize and adjust the fused positioning result according to the fused positioning result of the verification positioning data packet, thereby ensuring the accuracy and reliability of the positioning information.

[0017] Furthermore, the configuration file includes internal parameters of the positioning program, a location where the positioning result is saved, and a location where the positioning program log is saved, wherein the positioning program is used to perform fusion positioning processing on the positioning data packet.

[0018] A positioning data processing device comprises: an acquisition module for acquiring positioning data packets of all target vehicles; a classification module for classifying the positioning data packets of all target vehicles according to data collection time and vehicle identification; a determination module for determining a target number of containers according to the total number of positioning data packets and classification results, opening the target number of containers, and transferring configuration files and positioning data packets to corresponding containers, wherein the containers perform fusion positioning processing on the positioning data packets based on the configuration files and multi-threading technology.

[0019] A cloud device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the program to implement the above-mentioned positioning data processing method.

[0020] A vehicle uploads a positioning data packet to the above-mentioned cloud device.

[0021] Beneficial effects of the present invention:

[0022] The present invention obtains the positioning data packets of all target vehicles, classifies the positioning data packets of all target vehicles in coordination with vehicle identification, determines the target number of containers according to the total number of positioning data packets and the classification results, opens the target number of containers, and transfers the configuration file and the positioning data packet to the corresponding container. The container performs fusion positioning processing on the positioning data packet based on the configuration file and multi-threading technology, makes full use of the cloud computing power, efficiently loads and processes positioning related data, improves the processing speed, and improves the accuracy of positioning data based on the fusion method optimized by the posture graph. Thus, it solves the problems that current vehicles are often limited by local hardware computing power, cannot support large-scale rapid positioning data processing, and the accuracy of built-in sensors is also limited. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of the flow of the positioning data processing method provided by the present invention;

[0024] Figure 2 A schematic diagram of the overall cloud-based fusion positioning process provided by an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of a fusion positioning process provided by an embodiment of the present invention;

[0026] Figure 4 A block diagram of a positioning data processing device provided by the present invention;

[0027] Figure 5 This is a schematic diagram of the cloud platform structure provided by the present invention. DETAILED DESCRIPTION

[0028] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0029] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0030] The embodiment of the present invention proposes a positioning data processing method, such as Figure 1 As shown, the method comprises the following steps:

[0031] In step S101, the positioning data packets of all target vehicles are obtained.

[0032] Among them, the positioning data packet contains a data unit of vehicle location information, including location coordinates, timestamp, etc.; the present invention obtains the positioning data packet addresses of all target vehicles through automatic cloud scheduling to obtain the positioning data packets of all target vehicles.

[0033] In step S102, the positioning data packets of all target vehicles are classified according to the data collection time and the vehicle identification.

[0034] The data collection time is the timestamp recorded in each positioning data packet, that is, the specific moment when the data packet is collected; the vehicle identification is the information used to identify each vehicle, which may be a vehicle identification code, etc.

[0035] It can be understood that the embodiment of the present invention can identify the timestamp recorded in each positioning data packet to determine the data collection time, and at the same time identify the vehicle identification, and classify the positioning data packets of all target vehicles according to the data collection time and the vehicle identification. It should be noted that there are many methods for classifying the positioning data packets of all target vehicles, and classifying the positioning data packets of all target vehicles using the data collection time and the vehicle identification is only one of them.

[0036] In step S103, the target number of containers is determined according to the total number of positioning data packets and the classification results, the target number of containers are opened, and the configuration file and the positioning data packets are transferred to the corresponding containers, wherein the container performs fusion positioning processing on the positioning data packets based on the configuration file and multi-threading technology.

[0037] Among them, the container is a lightweight virtualization technology used to run applications in an isolated environment. Each container can execute tasks independently and can be started and stopped quickly. The configuration file is passed in from the outside, which will be described in detail below and will not be repeated here. Multi-threading technology allows a program to execute multiple threads at the same time to improve processing efficiency.

[0038] It can be understood that the embodiment of the present invention can determine the number of containers that need to be opened based on the total amount of all positioning data packets collected and the results of classification by time and vehicle identification, and start the corresponding number of containers, and pass the configuration file and the corresponding positioning data packets into these containers. Inside each container, based on the parameters in the configuration file, the positioning data packets are fused and positioned using multi-threading technology.

[0039] In the embodiment of the present invention, the configuration file includes internal parameters of the positioning program, a positioning result storage address and a positioning program log storage address, wherein the positioning program is used to perform fusion positioning processing on the positioning data packet.

[0040] It can be understood that the embodiment of the present invention uses a positioning program to perform fusion positioning processing on positioning data packets, and the configuration file includes internal parameters of the positioning program, a positioning result storage address, and a positioning program log storage address.

[0041] In an embodiment of the present invention, a container performs fusion positioning processing on a positioning data packet based on a configuration file and multi-threading technology, including: identifying a target processing quantity in a configuration file; performing time-series segmentation and grouping on the positioning data packets to be processed by the container according to the target processing quantity; treating the segmented and grouped data as a fusion positioning task, placing the fusion positioning task into a thread pool of the container, and starting at least one thread to process the fusion positioning task of the thread pool.

[0042] Among them, the thread pool is a design pattern that creates a group of threads in advance to execute a series of tasks, which can avoid the overhead caused by frequent creation and destruction of threads.

[0043] It can be understood that in the embodiment of the present invention, when the container performs fusion positioning processing on the positioning data packets based on the configuration file and multi-threading technology, the container first reads the parameters in the configuration file, obtains information such as the target processing quantity, and according to the target processing quantity specified in the configuration file, the positioning data packets to be processed in the container are divided and grouped in chronological order to form multiple subtasks, and the divided and grouped data are taken as a fusion positioning task and put into the thread pool of the container, and at least one thread is started to execute the fusion positioning task in the thread pool.

[0044] In an embodiment of the present invention, at least one thread is started to process a fusion positioning task of a thread pool, including: reading position data, lane line data and map data in data corresponding to the fusion positioning task; performing perception matching based on the lane line data and the map data to obtain a semantic positioning result; fusing the position data, the semantic positioning result and the track calculation data into a pose graph, and determining the fusion positioning result of the corresponding positioning data packet based on the fused pose graph.

[0045] Among them, the position data can be satellite positioning information data such as GPS data and Beidou data, which are not specifically limited here; the lane line data includes road marking information data captured by sensors such as cameras, which is used to assist vehicle positioning; perception matching is to calculate the exact position of the vehicle on the map by comparing the difference between the real-time perceived information and the pre-stored map data; the semantic positioning result is to provide more precise location information than traditional GPS through understanding the surrounding environment; track calculation is to predict the current position of the vehicle based on the vehicle's speed, direction and other motion parameters, which is used to supplement or correct GPS positioning; the pose graph is a data structure that represents the posture of an object at different times and the relationship between them, which is used to optimize positioning accuracy.

[0046] It can be understood that when the embodiment of the present invention starts at least one thread to process the fusion positioning task in the thread pool, it first reads the position data, lane line data and map data from the data packet, uses the lane line data and map data for perception matching to obtain the precise position information of the vehicle relative to the environment, and obtains the semantic positioning result. Subsequently, the data from the GPS, the semantic positioning result obtained by perception matching, and the track calculation data are combined, and fusion processing is performed through the pose graph optimization algorithm. Finally, the final fusion positioning result of each positioning data packet is determined based on the fused pose graph, thereby achieving more accurate correction and optimization of the original positioning data.

[0047] In an embodiment of the present invention, before the position data, semantic positioning results and track-reckoning data are integrated into the pose graph, it also includes: if there is no track-reckoning data in the data corresponding to the fused positioning task, the wheel speed and angular velocity in the position data are identified, and track-reckoning is performed based on the wheel speed and angular velocity.

[0048] It can be understood that before the embodiment of the present invention integrates the position data, semantic positioning results and track-reckoning data into the pose graph, if the track-reckoning data is missing, it is necessary to supplement the generation of track-reckoning data by analyzing the wheel speed and angular velocity information in the position data to ensure that the position data, semantic positioning results and track-reckoning data can be combined and input into the pose graph for optimization processing. It should be noted that in practice, the track-reckoning data can also be obtained by other means. The calculation method provided by the embodiment of the present invention to supplement the generation of track-reckoning data by analyzing the wheel speed and angular velocity information in the position data is only one of them.

[0049] In an embodiment of the present invention, before placing the fusion positioning task into the thread pool of the container and starting at least one thread to process the fusion positioning task of the thread pool, it also includes: determining the size of the thread pool according to the available CPU cores and memory; determining the number of threads to be started according to the size of the thread pool and the target processing number.

[0050] Among them, the CPU core is the part inside the central processing unit that can execute instructions independently.

[0051] It can be understood that before placing the fusion positioning task into the thread pool of the container and starting at least one thread to process the fusion positioning task of the thread pool, the embodiment of the present invention needs to first evaluate the CPU core and memory status of the system to set an appropriate thread pool size; then, according to the capacity of the thread pool and the total amount of tasks to be processed, the number of working threads to be started is reasonably arranged to ensure that all positioning data packets are processed efficiently and smoothly.

[0052] In the embodiment of the present invention, after the container performs fusion positioning processing on the positioning data packet based on the configuration file and the multi-threading technology, the method further includes: verifying the fusion positioning result of the positioning data packet; and optimizing the fusion positioning result according to the verification result.

[0053] It can be understood that after the positioning data packet is fused, the embodiment of the present invention will verify the generated fused positioning result to ensure its reliability; for those results that do not meet the requirements, measures will be taken to further optimize them until a satisfactory position estimate is obtained, thereby ensuring the accuracy of a single data point and improving the stability and reliability of the overall positioning system.

[0054] According to the positioning data processing method proposed in the embodiment of the present invention, by acquiring the positioning data packets of all target vehicles, the positioning data packets of all target vehicles are classified in coordination with the vehicle identification, and the target number of containers is determined according to the total number of positioning data packets and the classification results, the target number of containers is opened, and the configuration file and the positioning data packet are transferred to the corresponding container, and the container performs fusion positioning processing on the positioning data packet based on the configuration file and multi-threading technology, which makes full use of the cloud computing power, efficiently loads and processes the positioning related data, improves the processing speed, and improves the accuracy of the positioning data based on the fusion method optimized by the pose graph.

[0055] The positioning data processing method is further described below through a specific embodiment. Figure 2 As shown, the positioning data processing method specifically includes the following steps:

[0056] Step S201, open the container.

[0057] The cloud obtains the addresses of all pending positioning data packets, classifies them by time and vehicle frame number, and assigns the addresses of data packets with the same vehicle frame number and the same day to a container. In practice, there can be different ways to assign them, and this is just one method.

[0058] The number of tasks N that should be started is obtained by the total number of data packets and the number of data packets allocated to each container. The system starts N containers. N containers may not be started at the same time. The tasks started at the same time are determined based on the hardware resources, and each task corresponds to a container. Other tasks are in the queue. When a task is completed and the hardware resources are released, the container is started to perform the positioning task.

[0059] Step S202: Start the fusion program.

[0060] After starting the container, call the fusion algorithm script. You can also call the positioning program directly here. For easy management, use the script to call the positioning program. Read the configuration file passed in from the outside. The configuration file can be in YAML or JSON format. The configuration file contains the internal parameters of the positioning program, the location where the positioning results are saved, the location where the positioning program log is saved, and the location where the positioning status code is saved. At the same time, pass in the previously roughly allocated data packet address in the form of a string or configuration file

[0061] Step S203: grouping data packets.

[0062] Fusion positioning further groups the addresses of incoming data packets, and divides and groups them according to the default number of packets M processed by the positioning program read from the configuration file and the time order of the packets, with each group containing M packets.

[0063] Step S204, start multithreading.

[0064] The thread pool size is determined based on the actual available CPU cores and memory of the container, and K threads can be started at the same time. Each group of data is used as the input of a subtask, and all data packets of each group are put into the thread pool, the positioning program is started, and K threads are started in parallel. The remaining groups of data are queued into the task queue. When the positioning processing of a data packet is completed, it is added to the thread pool and the existing thread is called.

[0065] Step S205: Start the fusion thread and fusion program.

[0066] The fusion positioning process is as follows Figure 3 As shown, the details are as follows:

[0067] Step S301: data acquisition.

[0068] Fusion positioning reads position data, camera-perceived lane data, and high-precision map data from the data packet. The data needs to be verified when reading. If any of the above three types of data is missing, the packet is skipped and the data packet name and error type are recorded.

[0069] It should be noted that dead reckoning is performed based on the wheel speed and angular velocity in the position data, and dead reckoning data obtained by other methods may also be used directly.

[0070] Step S302: Perception matching and pose graph optimization.

[0071] 3.1 During initialization, the initial position of the vehicle is determined through position data.

[0072] 3.2 Traverse the perception data and pre-process the perception data of the current frame, such as projecting the lane line onto the ground to obtain the lane line point. Other perception results including stop lines, curbs, arrows, etc. are also processed accordingly to obtain point clouds. Use KdTree or other search methods on the high-precision map to find the high-precision map points of the corresponding elements based on the current vehicle position. Then perform ICP matching or other point cloud matching on the points of the camera perception elements and the corresponding point cloud of the high-precision map to obtain the matching results, thereby updating the current posture as the semantic positioning result.

[0073] 3.3 Use dead reckoning or GPS as the initial value of the next perceived pose, and repeat the matching process in 3.2 above. Continuously obtain semantic positioning results. The matching process needs to check whether all matching results are reasonable. If they are unreasonable, such as large matching errors, mark the point as an unreliable vertex. Reinitialize using GPS pose and repeat 3.1.

[0074] The GPS, dead reckoning, and semantic positioning results are added to the pose graph optimization equation for fusion. Based on the GPS state value, vehicle speed and steering angle during dead reckoning, and semantic positioning matching error, they are used as the weight value source in the fusion and jointly optimized. The optimization result is obtained.

[0075] Step S303: Result verification.

[0076] Verify the fusion result by comparing it with the GPS-fixed point and semantic positioning fixed point at the corresponding time. If the distance or direction exceeds the threshold, the optimization is considered unreliable and the point is marked.

[0077] Step S304: Save the positioning data.

[0078] Step S206: output the positioning result and save it.

[0079] Save all fusion positioning results including whether they are reliable and output them to a file.

[0080] If all data is processed, save the status code of each data packet processing for later summary of processing status

[0081] Next, the positioning data processing device proposed according to the embodiment of the present application is described with reference to the accompanying drawings.

[0082] Figure 4 is a block diagram of a positioning data processing device according to an embodiment of the present application. Figure 4 As shown, the positioning data processing device 10 includes: an acquisition module 401 , a classification module 402 and a determination module 403 .

[0083] Among them, the acquisition module 401 is used to obtain the positioning data packets of all target vehicles; the classification module 402 is used to classify the positioning data packets of all target vehicles according to the data collection time and the vehicle identification; the determination module 403 is used to determine the target number of containers according to the total number of positioning data packets and the classification results, open the target number of containers, and transfer the configuration file and the positioning data packet to the corresponding container, wherein the container performs fusion positioning processing on the positioning data packet based on the configuration file and multi-threading technology.

[0084] In an embodiment of the present invention, the determination module 403 is further used to: identify the target processing quantity in the configuration file; perform time-sequential segmentation and grouping of the positioning data packets to be processed by the container according to the target processing quantity; use the segmented and grouped data as a fused positioning task, put the fused positioning task into the thread pool of the container, and start at least one thread processing thread pool's fused positioning task.

[0085] In an embodiment of the present invention, at least one thread is started to process a fusion positioning task of a thread pool, including: reading position data, lane line data and map data in data corresponding to the fusion positioning task; performing perception matching based on the lane line data and the map data to obtain a semantic positioning result; fusing the position data, the semantic positioning result and the track calculation data into a pose graph, and determining the fusion positioning result of the corresponding positioning data packet based on the fused pose graph.

[0086] In an embodiment of the present invention, a calculation module is also included, wherein the calculation module is further used to: before fusing the position data, the semantic positioning result and the track calculation data into the pose graph, if there is no track calculation data in the data corresponding to the fused positioning task, identify the wheel speed and angular velocity in the position data, and perform track calculation based on the wheel speed and angular velocity.

[0087] In an embodiment of the present invention, the determination module 403 is further used to: before placing the fusion positioning task into the thread pool of the container and starting at least one thread to process the fusion positioning task of the thread pool, determine the size of the thread pool according to the available CPU cores and memory; determine the number of threads to be started according to the size of the thread pool and the target processing number.

[0088] In the embodiment of the present invention, the determination module 403 is further used for: after the container performs fusion positioning processing on the positioning data packet based on the configuration file and the multi-threading technology, it also includes: verifying the fusion positioning result of the positioning data packet; and optimizing the fusion positioning result according to the verification result.

[0089] In the embodiment of the present invention, the configuration file includes internal parameters of the positioning program, a positioning result storage address and a positioning program log storage address, wherein the positioning program is used to perform fusion positioning processing on the positioning data packet.

[0090] According to the positioning data processing device proposed in the embodiment of the present invention, the positioning data packets of all target vehicles are acquired through the coordinated action of the acquisition module, the classification module and the determination module, the positioning data packets of all target vehicles are classified in coordination with the vehicle identification, and the target number of containers is determined according to the total number of positioning data packets and the classification results. The target number of containers is opened, and the configuration file and the positioning data packet are transferred to the corresponding container. The container performs fusion positioning processing on the positioning data packet based on the configuration file and multi-threading technology, fully utilizes the cloud computing power, efficiently loads and processes the positioning related data, improves the processing speed, and improves the accuracy of the positioning data based on the fusion method optimized by the pose graph.

[0091] Figure 5 A schematic diagram of the structure of a cloud device provided in an embodiment of the present application. The cloud device may include:

[0092] A memory 501 , a processor 502 , and a computer program stored in the memory 501 and executable on the processor 502 .

[0093] When the processor 502 executes the program, the positioning data processing method provided in the above embodiment is implemented.

[0094] Furthermore, the cloud device also includes:

[0095] The communication interface 505 is used for communication between the memory 501 and the processor 502 .

[0096] The memory 501 is used to store computer programs that can be executed on the processor 502 .

[0097] The memory 501 may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile memory, such as at least one disk memory.

[0098] If the memory 501, the processor 502 and the communication interface 505 are implemented independently, the communication interface 505, the memory 501 and the processor 502 can be connected to each other through a bus and communicate with each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0099] Optionally, in a specific implementation, if the memory 501, the processor 502 and the communication interface 505 are integrated on a chip, the memory 501, the processor 502 and the communication interface 505 can communicate with each other through an internal interface.

[0100] The processor 502 may be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0101] An embodiment of the present invention further provides a vehicle, which uploads a positioning data packet to the above-mentioned cloud device.

[0102] The above embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Any equivalent substitution or change made by a person skilled in the art based on the present invention is within the protection scope of the present invention.

Claims

1. A positioning data processing method, characterized in that: The following steps are involved: Obtain positioning data packets of all target vehicles; Classify the positioning data packets of all target vehicles according to the data collection time and vehicle identification; The target number of containers is determined according to the total number of positioning data packets and the classification results, the target number of containers are opened, and the configuration file and the positioning data packets are transferred to the corresponding containers, wherein the containers perform fusion positioning processing on the positioning data packets based on the configuration file and multi-threading technology.

2. The positioning data processing method according to claim 1, characterized in that: The container performs fusion positioning processing on the positioning data packet based on the configuration file and multi-threading technology, including: identifying a target process quantity in the configuration file; Performing time-sequential segmentation and grouping of the positioning data packets to be processed by the container according to the target processing quantity; The segmented and grouped data is used as a fusion positioning task, the fusion positioning task is put into the thread pool of the container, and at least one thread is started to process the fusion positioning task of the thread pool.

3. The positioning data processing method according to claim 2, characterized in that: The starting of at least one thread to process the fusion positioning task of the thread pool includes: Reading the position data, lane line data and map data in the data corresponding to the fusion positioning task; Performing perceptual matching on the lane line data and the map data to obtain a semantic positioning result; The position data, the semantic positioning result and the dead reckoning data are fused into a pose graph, and a fused positioning result of a corresponding positioning data packet is determined according to the fused pose graph.

4. The positioning data processing method according to claim 3, characterized in that: Before fusing the position data, the semantic positioning result and the dead reckoning data into the pose graph, the method further includes: If the dead reckoning data does not exist in the data corresponding to the fused positioning task, the wheel speed and angular velocity in the position data are identified, and dead reckoning is performed according to the wheel speed and the angular velocity.

5. The positioning data processing method according to claim 2, characterized in that: Before placing the fusion positioning task into the thread pool of the container and starting at least one thread to process the fusion positioning task of the thread pool, the method further includes: Determine the size of the thread pool based on available CPU cores and memory; The number of threads to be opened is determined according to the size of the thread pool and the target processing quantity.

6. The positioning data processing method according to claim 1, characterized in that: After the container performs fusion positioning processing on the positioning data packet based on the configuration file and the multi-threading technology, the container further includes: Verifying the fused positioning result of the positioning data packet; The fusion positioning result is optimized according to the verification result.

7. The positioning data processing method according to any one of claims 1 to 6, characterized in that: The configuration file includes internal parameters of the positioning program, a positioning result storage address and a positioning program log storage address, wherein the positioning program is used to perform fusion positioning processing on the positioning data packet.

8. A positioning data processing device, characterized in that: include: An acquisition module is used to acquire positioning data packets of all target vehicles; A classification module, used to classify the positioning data packets of all target vehicles according to the data collection time and vehicle identification; A determination module is used to determine a target number of containers according to the total number of positioning data packets and classification results, open the target number of containers, and transfer the configuration file and the positioning data packet to the corresponding container, wherein the container performs fusion positioning processing on the positioning data packet based on the configuration file and multi-threading technology.

9. A cloud device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the positioning data processing method according to any one of claims 1 to 7.

10. A vehicle, characterized in that: The vehicle uploads a positioning data packet to the cloud device described in claim 9.

Images