Automatic driving scene self-recognition method and storage medium

Abstract

The invention relates to an automatic driving scene self-recognition method and a storage medium. The method comprises the steps of data acquisition and storage, data cleaning, events, scene recognition and point burying, data cutting, classification and label labeling, format conversion and data use. According to the invention, automatic identification and extraction, format conversion and labeled classified storage management are carried out on all collected whole vehicle network signals and sensor original signal data; a scene library established according to layering and classification of real vehicle data covers a natural traffic flow scene, an automatic driving failure working condition, driver takeover data, whole vehicle driving data, a working condition outside an operation area range, a traffic accident, traffic participant interaction data and man-machine interaction data. Big data support can be provided for automatic automobile product definition, function planning, demand analysis, software and hardware development, system integration, simulation and real automobile verification, function and performance rapid iteration updating, automatic driving standard regulation making and the like, and the utilization efficiency and the repeated utilization rate of data are improved.

Description

technical field

[0001] The invention belongs to the technical field of data collection and analysis, and in particular relates to a self-identification method and a storage medium of an automatic driving scene. Background technique

[0002] The diversity, complexity, uncertainty, inability to fully reproduce, etc. of traffic scene elements such as road types, lane lines, signs, obstacles, etc., make it difficult for the test and verification of automatic driving functions to cover all traffic scenes or require huge costs. At the same time, it is also a limiting factor for the real mass production of unmanned autonomous driving. Continuous iterative update of data-driven software functions and performance is undoubtedly the most cost-effective way, the fastest iteration speed, and the strongest self-growth ability.

[0003] In order to ensure the safety of the automatic driving function, any optimization and upgrade of the software requires hundreds of thousands of kilometer...

Examples

Embodiment Construction

[0037] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0038] See Figure 1-Figure 5 , a specific embodiment of an automatic driving scene self-identification method, specifically includes the following steps:

[0039] S1: Data collection and storage, upload the vehicle network signals and sensor raw data of the collection vehicle, test vehicle, mass production vehicle and third-party platform to the cloud storage server;

[0040] S2: data cleaning, checking the integrity and continuity of the data in the cloud storage server in step S1, and removing damaged and invalid data;

[0041] Wherein, step S2 also includes:

[0042] S201: Extract the data uploaded in step S1 from the cloud storage server;

[0043] S202: Detect the integrity of the data in step S201, and if the data is complete, perform S203; if the data is incomplete, notify the data administrator to eliminate the data with l...

technical field

[0001] The invention belongs to the technical field of data collection and analysis, and in particular relates to a self-identification method and a storage medium of an automatic driving scene. Background technique

[0002] The diversity, complexity, uncertainty, inability to fully reproduce, etc. of traffic scene elements such as road types, lane lines, signs, obstacles, etc., make it difficult for the test and verification of automatic driving functions to cover all traffic scenes or require huge costs. At the same time, it is also a limiting factor for the real mass production of unmanned autonomous driving. Continuous iterative update of data-driven software functions and performance is undoubtedly the most cost-effective way, the fastest iteration speed, and the strongest self-growth ability.

[0003] In order to ensure the safety of the automatic driving function, any optimization and upgrade of the software requires hundreds of thousands of kilometer...

Examples

Embodiment Construction

[0037] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0038] See Figure 1-Figure 5 , a specific embodiment of an automatic driving scene self-identification method, specifically includes the following steps:

[0039] S1: Data collection and storage, upload the vehicle network signals and sensor raw data of the collection vehicle, test vehicle, mass production vehicle and third-party platform to the cloud storage server;

[0040] S2: data cleaning, checking the integrity and continuity of the data in the cloud storage server in step S1, and removing damaged and invalid data;

[0041] Wherein, step S2 also includes:

[0042] S201: Extract the data uploaded in step S1 from the cloud storage server;

[0043] S202: Detect the integrity of the data in step S201, and if the data is complete, perform S203; if the data is incomplete, notify the data administrator to eliminate the data with l...

Images