The Multi-sensor Fusion Automatic Driving Test Scene Algorithm based on Cloud Platform

Abstract

The production of autonomous vehicles needs to accumulate a large amount of mileage experience while the researching, development and testing methods of the traditional vehicle terminals cannot meet the requirements. The industry generally adopts the new Cloud Platform to research and develop iterative mode, which collects massive road environment data through the test fleet, and conducts model and algorithm development and simulation verification, so as to meet the product safety standards. Based on collaboration capability of the Cloud, Pipe and Terminal, this paper supports to carry out data collection, storage, processing and analysis through high-performance basic platform, application platform and AI development platform and other methods. Through cloud training and deployment of the autonomous vehicle , it is completed for scenario development to improve model training speed and development efficiency.