Development system for original model of automatic driving system

Abstract

The invention discloses a development system for an original model of an automatic driving system. The system comprises a rapid prototype controller, a camera, a millimeter wave radar, a first sensingfusion module, m sensing modules and n sensing fusion modules; the n sensing fusion modules fuse lane line and barrier information detected by n sensing groups respectively; a control algorithm module, a signal processing module, a first sensing fusion analysis module and n sensing fusion analysis modules are integrated in the rapid prototype controller; the n sensing groups are formed by combining the camera, the millimeter wave radar and 1-m sensing modules; and the n sensing fusion analysis modules receive and analyze fusion information sent by the n sensing fusion modules respectively, and output target-level lane line and barrier signals, and the signal processing module selects and processes the signals, and outputs the signals to the control algorithm module. Thus, the automatic driving function can be developed flexibly and rapidly.

Description

technical field [0001] The invention belongs to the field of automatic driving, and in particular relates to an original model development system of an automatic driving system. Background technique [0002] In the prototyping process of L3 autonomous driving, ADAS (Advanced Driving Assistant System) control algorithms are usually developed in a modular and platform-based mode, and changes in external signals do not affect the control module as much as possible. [0003] The original model development system of the existing automatic driving system includes a rapid prototype controller, a camera, a millimeter-wave radar, and the first sensor fusion module. The rapid prototype controller integrates a control algorithm module, a signal processing module, and the first sensor fusion module. The fusion analysis module and the vehicle CAN analysis module, the signal processing module, the first sensor fusion analysis module and the vehicle CAN analysis module form the control int...

AI Technical Summary

Problems solved by technology

[0004] Due to the great uncertainty in the types of perception modules such as sensors and high-precision maps on the vehicle, each time the perception platform is switched, for example, from the perception platform composed of the camera, millimeter-wave radar and the first sensor fusion module When switching to a perception platform consisting of a camera, millimeter-wave radar, lidar, and a second sensing fusion module (used to fuse lane lines and obstacle information detected by the camera, millimeter-wave radar, and lidar respectively), it is necessary to When changing the first sensor fusion analysis module in the control interface module to the second sensor fusion analysis module corresponding to the second sensor fusion module, or switching to other perception platforms, the first sensor fusion analysis module still needs to be It is cumbersome to change the module to other corresponding sensor fusion analysis modules, especially when comparing the control effects of different sensing platform inputs, it is necessary to frequently add sensor fusion modules, frequently change sensor fusion analysis modules, and frequently compile control models, which consumes a lot of time. long time

Images