A data acquisition and processing system and a data acquisition and processing method

Abstract

The present invention relates to a data acquisition and processing system and a data acquisition and processing method, which are composed of a vehicle-side controller and a cloud server. The vehicle-side controller is connected to the environmental perception system and the vehicle bus system on the vehicle, and is used to collect environmental perception data and vehicle driving status data, and upload the data to the cloud server through the network. The data acquisition and processing method of the data acquisition and processing system of the present invention includes vehicle-side data acquisition and processing and cloud data storage management and analysis. Data value identification is directly performed at the data source end, and the judgment standard of the value has a certain universality and stability, and can continuously discover value and then collect it. The collected data is valuable data; the data acquisition flow cost and data storage cost are reduced, and the time for data processing and analysis is saved, and a database with direct data index is established; due to the stability and universality of the value judgment standard, it can run for a long time at the data source end without frequent updates.

Description

Technical Field

[0001] The present invention belongs to the technical field of autonomous driving, and in particular relates to a data acquisition and processing system and a data acquisition and processing method. Background Art

[0002] At present, big data technology has been widely used in all walks of life, but its data volume is large and the data value density is low. Often, a large amount of data is collected and accumulated, but when you want to analyze the data in a certain aspect, it is like looking for a needle in a haystack. You can't find it or it takes a lot of effort to find valuable data. On the one hand, it wastes traffic and data storage space to collect and store these massive data; on the other hand, it creates difficulties for using and extracting valuable data.

[0003] In many industries, especially the automotive industry, the general data collection and processing method is to collect and store all the data, and then find the required data for analysis through data cleaning. The prior art discloses a method, system and computer storage medium for evaluating a regulatory control algorithm. The evaluation method of the regulatory control algorithm includes: the vehicle side identifies abnormal scenes based on vehicle data, and the vehicle data includes external data and internal data; after the vehicle side identifies the abnormal scene, it generates abnormal scene data and transmits it to the offline side; the offline side constructs a virtual scene based on the abnormal scene data, and evaluates the regulatory control algorithm corresponding to the abnormal scene based on the virtual scene. It mainly performs data cleaning and machine learning training or virtual scene construction for the collection of machine learning training data. This technology does not give a clear explanation of the design of data collection events and data labels, nor does it have a database comparison function for checking duplicates. Summary of the invention

[0004] The purpose of the present invention is to provide a data acquisition and processing system that directly performs data value identification at the data source end, and also to provide a data acquisition and processing method for the data acquisition and processing system, so as to solve the problems of waste of data storage and data flow resources caused by indiscriminate data acquisition, and the inability to retrieve and low utilization rate of big data after acquisition and storage.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] A data acquisition and processing system consists of a vehicle-side controller and a cloud server; the vehicle-side controller is connected to the environmental perception system and the vehicle bus system on the vehicle, and is used to collect environmental perception data and vehicle driving status data; the cloud server has a data transceiver module, a storage module, a retrieval module and a playback analysis module, which are used for data transceiver, storage, retrieval and playback analysis; the vehicle-side controller uploads data to the cloud server via the network; the vehicle-side controller is connected to the driver status monitoring system, and is used to collect driver control data.

[0007] A data collection and processing method of a data collection and processing system, including vehicle-side data collection and processing and cloud-side data storage management and analysis;

[0008] The vehicle-side data collection and processing includes the following steps:

[0009] A1. The vehicle-side controller collects and records the environmental perception sensor data, vehicle driving status data and driver control data in real time over the recent period;

[0010] A2. Analyze the environmental perception sensor data. If an event with low recognition credibility occurs, or the difference between the autonomous driving trajectory or speed planning data and the driver's actual driving trajectory and speed exceeds a threshold, or a driver takeover event occurs during autonomous driving, the next step of the analysis process will be triggered.

[0011] A3. Intercept the data of a period of time before and after the above event and the current event information from the environmental perception sensor data, and extract key information from the environmental perception sensor data and vehicle driving status data to form a scene label;

[0012] A4. Compare the scene data tag generated in step A3 with the scene tag database stored in the cloud server. If they have the same scene, add 1 to the scene data count corresponding to the database. If the cumulative count value of the scene tag is less than the calibration value, compress and package the data of the calibration time before and after the event with the scene tag and upload them to the cloud server. Otherwise, discard the scene tag.

[0013] The cloud data storage management analysis includes the following steps:

[0014] B1. The cloud server receives the uploaded data, extracts the scene tags in the data, stores the scene tags and original data in the scene tag database and the original database respectively, and establishes a link;

[0015] B2. Search the scene tag content, enter the search keyword, find the corresponding original data for analysis, and display the original data in the form of curves, icons, video streams, etc.;

[0016] B3. Collect data uploaded by multiple vehicles, and update the tag database according to the data tag uploaded by each vehicle each time, including the corresponding count value, and update the tag database in the vehicle-side controller through the network with the updated tag database.

[0017] Furthermore, in step A2, the low credibility event includes the credibility of the lane line recognition signal, the credibility of the obstacle target recognition signal, etc. When the credibility value is lower than a certain calibration value, it is considered that the low credibility event of the signal occurs.

[0018] Furthermore, the calibration value is initially set to 30%.

[0019] Further, in step A2, the difference between the automatic driving trajectory or speed planning data and the actual trajectory and speed of the driver exceeds a threshold.

[0020] Furthermore, the lateral deviation of the trajectory is greater than 1 meter or the speed deviation is greater than 30 km / h.

[0021] Furthermore, in step A3, the key information includes target information, road information and vehicle control information.

[0022] Furthermore, the target information is that there are pedestrians crossing the road, the road information is a 2-lane straight section, and the vehicle control information is braking 50% and turning 90 degrees.

[0023] Further, in step A4, the accumulated count value of the scene is less than a certain calibration value, and the calibration time before and after the event is a certain calibration value before the event and a certain calibration value after the event.

[0024] Furthermore, the calibration value is initially set to 50, and the calibration time before and after the event is 60 seconds before the event and 30 seconds after the event. Compared with the prior art, the beneficial effects of the present invention are:

[0025] The data acquisition and processing system and the data acquisition and processing method of the present invention directly identify the value of data at the data source end, and the judgment standard of the value has a certain universality and stability, and can continuously discover value and then collect it, and the collected data is valuable data; it reduces the data acquisition flow cost and data storage cost, and saves the time of data processing and analysis, and establishes a database directly with data indexes; due to the stability and universality of the value judgment standard, it can run for a long time at the data source end without frequent updates. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0027] Figure 1 A schematic diagram of the structure of the data acquisition and processing system of the present invention;

[0028] Figure 2 Vehicle-side controller data collection flow chart;

[0029] Figure 3Cloud server data processing flow chart. DETAILED DESCRIPTION

[0030] The present invention will be further described below in conjunction with embodiments:

[0031] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0032] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0033] Based on the ultimate goal of autonomous driving, the present invention is to drive the vehicle like a human, and is designed to be a stable data source processing method and system that can continuously and automatically discover problems, record and upload problems. It is mainly used in the automotive electronic control industry or the autonomous driving industry to solve problems related to data collection and data management.

[0034] like Figure 1 As shown, the data acquisition and processing system of the present invention is composed of a vehicle-side controller and a cloud server used in conjunction with the vehicle-side controller.

[0035] The vehicle-side controller is connected to the vehicle's environmental perception system and vehicle bus system, and may also include a driver status monitoring system. It can collect environmental perception sensor data, vehicle driving status data and driver control data, and can also upload this data to a cloud server through the network.

[0036] The cloud server has a data transceiving module, a storage module, a retrieval module and a playback analysis module, which are used for data transceiving, storage, retrieval and playback analysis; the vehicle-side controller uploads the data to the cloud server through the network.

[0037] The data collection and processing method of the present invention includes vehicle-side data collection and processing and cloud data storage management and analysis. The vehicle-side data collection and processing includes data collection, data processing, event triggering and tagging methods of the vehicle-side controller. The cloud data storage management and analysis includes data storage, retrieval and analysis methods of the cloud server.

[0038] like Figure 2 Specifically, the vehicle-side data collection and processing includes the following steps:

[0039] Step 1: The vehicle-side controller collects and records the environmental perception sensor data, vehicle driving status data and driver control data in real time over the recent period, and processes and analyzes the above data. First, the environmental perception sensor data is analyzed. If an event with low recognition quality (credibility) occurs, the next step of the analysis process will be triggered;

[0040] Step 2: Extract the target information being used from the environmental perception sensor data and generate the scene label of the data after processing;

[0041] Step 3: intercept the data of a period of time before and after the above event occurs, as well as the current event information, from the environmental perception sensor data, and extract key information from the environmental perception sensor data and vehicle driving status data to form a scene data label;

[0042] Step 4: Compare the scene data tag generated in step 3 with the scene tag database stored in the cloud server. If they have the same scene, add 1 to the scene data count corresponding to the database. If the cumulative count value of the scene tag is less than a certain calibration value, compress and package the data of a certain calibration value before the event and a certain calibration value after the event with the scene tag and upload them to the cloud server. Otherwise, discard the scene tag.

[0043] like Figure 3 As shown, the cloud data storage management analysis includes the following steps:

[0044] Step 1: The cloud server receives the uploaded data;

[0045] Step 2: Extract scene labels from the data, store the scene labels and original data in the scene label database and the original database respectively, and establish links;

[0046] Step 3: The cloud server has a data retrieval function. By searching the scene tag content, the corresponding original data can be found for analysis. For example, by entering the search keywords: pedestrian crossing the road + 2 lanes + straight driving, the corresponding data can be found;

[0047] Step 4: The cloud server has a data playback function, which can display the original data through curves, icons, video streams, etc.

[0048] Step 5: The cloud server can collect data uploaded by multiple vehicles, and update the tag database according to the data tags uploaded by each vehicle each time, including the corresponding count values, and update the tag database in the vehicle-side controller through the network with the updated tag database.

[0049] Example 1

[0050] Vehicle-side data collection and processing:

[0051] Step 1: The vehicle-side controller collects and records the environmental perception sensor data, vehicle driving status data and driver control data in real time over the recent period, and processes and analyzes the above data. First, the environmental perception sensor data is analyzed. If the credibility of the lane line recognition signal is less than 30%, the next step of the analysis process will be triggered;

[0052] Step 2: Extract information about pedestrians crossing the road from the environmental perception sensor data, and generate scene labels for the data after processing;

[0053] Step 3: intercept the data of a period of time before and after the above event occurs, as well as the current event information, from the environmental perception sensor data, and extract key information from the environmental perception sensor data and vehicle driving status data to form a scene data label;

[0054] Step 4: Compare the scene data tag generated in step 3 with the scene tag database stored in the cloud server. If they have the same scene, add 1 to the scene data count corresponding to the database. If the cumulative count value of the scene tag is less than 50, compress and package the data 60 seconds before and 30 seconds after the event with the scene tag and upload it to the cloud server. Otherwise, discard the scene tag.

[0055] Cloud Data Storage Management Analysis:

[0056] Step 1: The cloud server receives the uploaded data;

[0057] Step 2: Extract scene labels from the data, store the scene labels and original data in the scene label database and the original database respectively, and establish links;

[0058] Step 3: The cloud server has a data retrieval function. By searching the scene tag content, the corresponding original data can be found for analysis. For example, by entering the search keyword: pedestrians crossing the road, the corresponding data can be found;

[0059] Step 4: The cloud server has a data playback function, which can display the original data through curves, icons, video streams, etc.

[0060] Step 5: The cloud server can collect data uploaded by multiple vehicles, and update the tag database according to the data tags uploaded by each vehicle each time, including the corresponding count value, and update the tag database in the vehicle-side controller through the network with the updated tag database.

[0061] Vehicle-side data collection and processing:

[0062] Step 1: The vehicle-side controller collects and records the environmental perception sensor data, vehicle driving status data and driver control data in real time over the recent period, and processes and analyzes the above data. First, the environmental perception sensor data is analyzed. If the lateral deviation of the autonomous driving trajectory is greater than 1 meter and the speed deviation is greater than 30km / h, the next step of the analysis process will be triggered;

[0063] Step 2: Extract the 2-lane straight section information from the environmental perception sensor data and generate the scene label of the data after processing;

[0064] Step 3: intercept the data of a period of time before and after the above event occurs, as well as the current event information, from the environmental perception sensor data, and extract key information from the environmental perception sensor data and vehicle driving status data to form a scene data label;

[0065] Step 4: Compare the scene data tag generated in step 3 with the scene tag database stored in the cloud server. If they have the same scene, add 1 to the scene data count corresponding to the database. If the cumulative count value of the scene tag is less than 50, compress and package the data 60 seconds before and 30 seconds after the event with the scene tag and upload it to the cloud server. Otherwise, discard the scene tag.

[0066] Cloud Data Storage Management Analysis:

[0067] Step 1: The cloud server receives the uploaded data;

[0068] Step 2: Extract scene labels from the data, store the scene labels and original data in the scene label database and the original database respectively, and establish links;

[0069] Step 3: The cloud server has a data retrieval function. By searching the scene tag content, the corresponding original data can be found for analysis. For example, by entering the search keyword: 2 lanes, the corresponding data can be found;

[0070] Step 4: The cloud server has a data playback function, which can display the original data through curves, icons, video streams, etc.

[0071] Step 5: The cloud server can collect data uploaded by multiple vehicles, and update the tag database according to the data tags uploaded by each vehicle each time, including the corresponding count values, and update the tag database in the vehicle-side controller through the network with the updated tag database.

[0072] Example 3

[0073] Vehicle-side data collection and processing:

[0074] Step 1: The vehicle-side controller collects and records the environmental perception sensor data, vehicle driving status data and driver control data in real time over the recent period, and processes and analyzes the above data. First, the environmental perception sensor data is analyzed. If the driver takes over during autonomous driving, the next step of the analysis process will be triggered;

[0075] Step 2: Extract information about pedestrians crossing the road, 2-lane straight sections, 50% braking, and 90-degree turns from the environmental perception sensor data, and generate scene labels for the data after processing;

[0076] Step 3: intercept the data of a period of time before and after the above event occurs, as well as the current event information, from the environmental perception sensor data, and extract key information from the environmental perception sensor data and vehicle driving status data to form a scene data label;

[0077] Step 4: Compare the scene data tag generated in step 3 with the scene tag database stored in the cloud server. If they have the same scene, add 1 to the scene data count corresponding to the database. If the cumulative count value of the scene tag is less than 50, compress and package the data 60 seconds before and 30 seconds after the event with the scene tag and upload it to the cloud server. Otherwise, discard the scene tag.

[0078] Cloud Data Storage Management Analysis:

[0079] Step 1: The cloud server receives the uploaded data;

[0080] Step 2: Extract scene labels from the data, store the scene labels and original data in the scene label database and the original database respectively, and establish links;

[0081] Step 3: The cloud server has a data retrieval function. By searching the scene tag content, the corresponding original data can be found for analysis. Enter the search keywords: pedestrian crossing the road + 2 lanes + straight driving, and the corresponding data can be found;

[0082] Step 4: The cloud server has a data playback function, which can display the original data through curves, icons, video streams, etc.

[0083] Step 5: The cloud server can collect data uploaded by multiple vehicles, and update the tag database according to the data tags uploaded by each vehicle each time, including the corresponding count values, and update the tag database in the vehicle-side controller through the network with the updated tag database.

[0084] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A data acquisition and processing method for a data acquisition and processing system, the data acquisition and processing system, consisting of a vehicle-side controller and a cloud server; the vehicle-side controller is connected to the environmental perception system and the vehicle bus system on the vehicle, and is used to collect sensor data and bus data; the cloud server has a data transceiver module, a storage module, a retrieval module and a playback analysis module, which are used for data transceiver, storage, retrieval and playback analysis; the vehicle-side controller uploads data to the cloud server via the network; the vehicle-side controller is connected to the driver status monitoring system, and is used to collect driver status data. The data acquisition and processing method, It is characterized in that Including vehicle-side data collection and processing and cloud data storage, management and analysis; The vehicle-side data collection and processing includes the following steps: A1. The vehicle-side controller collects and records the environmental perception sensor data, vehicle driving status data and driver control data in real time over the recent period; A2. Analyze the environmental perception sensor data. If an event with low recognition credibility occurs, or the difference between the autonomous driving trajectory or speed planning data and the driver's actual driving trajectory and speed exceeds a threshold, or a driver takeover event occurs during autonomous driving, the next step of the analysis process will be triggered. A3. Intercept the data of a period of time before and after the above event and the current event information from the environmental perception sensor data, and extract key information from the environmental perception sensor data and vehicle driving status data to form a scene label; A4. Compare the scene data tag generated in step A3 with the scene tag database stored in the cloud server. If they have the same scene, add 1 to the scene data count corresponding to the database. If the cumulative count value of the scene tag is less than the calibration value, compress and package the data of the calibration time before and after the event with the scene tag and upload them to the cloud server. Otherwise, discard the scene tag. The cloud data storage management analysis includes the following steps: B1. The cloud server receives the uploaded data, extracts the scene tags in the data, stores the scene tags and the original data in the scene tag database and the original database respectively, and establishes a link; B2. Search the scene tag content, enter the search keyword, find the corresponding original data for analysis, and display the original data in the form of curves, icons, video streams, etc.; B3. Collect data uploaded by multiple vehicles, and update the tag database according to the data tag uploaded by each vehicle each time, including the corresponding count value, and update the tag database in the vehicle-side controller through the network with the updated tag database.

2. The data acquisition and processing method of the data acquisition and processing system according to claim 1, Features: In step A2, the low credibility events include the credibility of the lane line recognition signal, the credibility of the obstacle target recognition signal, etc. When the credibility value is lower than a certain calibration value, it is considered that the low credibility event of the signal occurs.

3. The data acquisition and processing method of the data acquisition and processing system according to claim 2, Features: The calibration value is initially set to 30%.

4. The data acquisition and processing method of the data acquisition and processing system according to claim 1, Features: Step A2: The difference between the automatic driving trajectory or speed planning data and the actual trajectory and speed of the driver exceeds a threshold.

5. The data acquisition and processing method of the data acquisition and processing system according to claim 4, Features: The lateral deviation of the trajectory is greater than 1 meter or the speed deviation is greater than 30 km / h.

6. The data acquisition and processing method of the data acquisition and processing system according to claim 1, Features: Step A3, the key information includes target information, road information and vehicle control information.

7. The data acquisition and processing method of the data acquisition and processing system according to claim 6, Features: The target information is that there are pedestrians crossing the road, the road information is a 2-lane straight section, and the vehicle control information is braking 50% and turning 90 degrees.

8. The data acquisition and processing method of the data acquisition and processing system according to claim 1, Features: Step A4: the accumulated count value of the scene is less than a certain calibration value, and the calibration time before and after the event is a certain calibration value before the event and a certain calibration value after the event.

9. The data acquisition and processing method of the data acquisition and processing system according to claim 8, Features: The calibration value is initially set to 50, and the calibration time before and after the event is 60 seconds before the event and 30 seconds after the event.

Images