Traffic simulation is an efficient and cost-effective way to test AutonomousVehicles in a complex and dynamic environment . The current simulation environments fall behind on two fronts — the background vehicles fail to simulate naturalistic driving behavior and the existing environments do not test the entire pipeline in a modular fashion . We propose to extend our developed software by adding CARLA, which in turn will enrich the perception of the ego vehicle by providing realistic sensors outputs of the AVs surrounding environment . Using the proposed framework, AVs perception, planning, and control could be testedin a complex, realistic driving environment. The performance of the proposedframework in constructing output generation and AV evaluations are demonstrated using several case studies. The proposed framework has been demonstrated using a series of case studies . The proposal is available on the OpenAI platform. For more information, please visit http://www.openAI.com/autonomousautomotive-simulation-sensors-to-vehicle-simulating-vehicles-assemble-as-preventive-assistant-assassible-assual

Author(s) : Pei Li, Arpan Kusari, David J. LeBlanc

Links : PDF - Abstract

Code :
Coursera

Keywords : simulation - framework - vehicles - environment - proposed -

Leave a Reply

Your email address will not be published. Required fields are marked *