How to perform efficiency your combinatorics for Autonomous Vehicle Testing
Based on a Renault-Nissan methodology deployment, SVA, and other R&D projects, this presentation explains how to deal with the huge combinatorics meet in ADAS simulation and to compute related autonomous vehicle expected behavior in order to generate tests cases.
I will present the MaTeLo Model-Based Testing process where we model weather, road infrastructures, other motorist behaviors, EGO system, road events, etc. and I explain how to design determinist test cases with great variability.
From this MaTeLo model, I'll show the automatic generation of thousand scenery samples using Gibbs sampler algorithms, to build automatically Oktal SCANeR scripts, for 3D simulation run.
This methodology can simulate million kilometers of driving.
Fabrice TROLLET
See his recent activity : / recent-activity
Download MaTeLo for free : www.all4tec.net/matelo