Multisource data mining-based vehicle intelligent driving system validity evaluation method

Abstract

The invention discloses a multisource data mining-based vehicle intelligent driving system validity evaluation method. The method comprises the steps of 1) obtaining a to-be-evaluated vehicle model, a random traffic scene model and a passenger injury model through multisource data mining; 2) based on the random traffic scene model and the vehicle model, simulating a specific scene of an accident; 3) simulating the scene by utilizing accident reconstruction software, and outputting a post-collision vehicle state; 4) in combination with the passenger injury model, calculating a passenger injury risk and a unit mileage risk after the accident; and 5) replacing the to-be-evaluated vehicle model with a conventional model, and repeating the steps 1) to 4) to obtain unit mileage and passenger injury risks of a vehicle free of a to-be-evaluated system; and then comparing injury results under the conditions that the to-be-evaluated system exists and does not exist, thereby obtaining the validity of the system. An evaluation platform depends on multisource data and known software, can realize evaluation of multiple intelligent driving systems, is low in running cost and high in speed, and has relatively good universal applicability.

Description

technical field [0001] The invention relates to a vehicle intelligent driving technology, in particular to a method for evaluating the effectiveness of an automobile intelligent driving system in reducing the risk of injury to passengers. Background technique [0002] For the evaluation of the effectiveness of the vehicle intelligent driving system in reducing the risk of occupant injury, the existing evaluation methods can be divided into two categories in terms of equipment conditions: one is an experimental method, and the other is a simulation method. The experimental methods can be divided into two categories: real vehicle experiments and standard tests. The former: universally applicable to different intelligent driving systems, but expensive, time-consuming, and dangerous; the latter: unified standards, objective evaluation, and simple operation , the repeatability is good, but the working condition is single, and the statistical analysis under multiple working condit...

AI Technical Summary

Problems solved by technology

[0003] Existing evaluation methods can be divided into two categories in terms of data sources: accident data and FOT data. Accident data: can reflect various dangerous accident types and can establish occupant injury models, but the data is incomplete and the database is not perfect. A large number of unified sorting and classification work; FOT data: the data is complete and large in number, and random simulation scenarios can be established

Images