Steve Lemke

Simulation Platform for 2021 IEEE AD AI Test Challenge

SVL Simulator is a co-sponsor of the 2021 IEEE Autonomous Driving AI Test Challenge in conjunction with IEEE, San Jose State University (SJSU), Baidu and others. SJSU faculty members have been using the SVL Simulator for several years in autonomous classes and for student projects, so when SJSU professors decided to create a global challenge for the 2021 IEEE AI Test Conference they chose LG’s SVL Simulator as the simulation platform for the competition. Organizers from industry and academia in a dozen countries met in the Fall of 2020 to plan and prepare for this challenge. The challenge was launched in January, 2021 and by the time registration ended in mid-March, 234 individuals had registered to participate and self-organized into 119 different teams.

Background on the challenge


In the industry today, simulation is needed to facilitate the development and testing of autonomous driving systems. In order to perform efficient simulation testing, techniques for optimizing and accelerating testing processes are in great demand. This challenge was set up as a platform to address this demand, and advocate the importance of and need for quality validation and automation for autonomous vehicles. This platform provides a global competition opportunity for international student teams and professional teams to develop diverse simulation testing techniques and approaches in test scenario generation and automation.

Goals of the challenge


In this challenge, LG’s SVL Simulator provides the simulation testing and scenario execution environment, and is fully integrated with Baidu Apollo, which will be the autonomous 'system under test'. Teams will create test cases and then use LG’s SVL Simulator and Apollo to evaluate those test cases. See more details on challenge goals and deliverables here.

Two phases of the challenge


The first phase is “Training and team selection” where training materials including tutorial videos, documentation, and a total of four two-hour live online training and Q&A sessions (hosted by SVL Simulator group) were presented so that everyone could get familiar with the usage of SVL Simulator and Baidu Apollo, and how to create test case scenarios. Phase 1 requires teams to create a few simple test scenarios to demonstrate their proficiency using SVL Simulator. Phase 1 runs through the end of April 2021; teams selected to advance to phase 2 were notified by May 15th 2021.

The second phase is the actual competition which runs through mid-July, during which teams will leverage available traffic accident/crash database (and potentially other data sources) to analyze and create scenarios and automation scripts to be submitted for evaluation. When the all-new cloud-based SVL Simulator (2021.1) was released in late March 2021, it was chosen to be the platform for phase 2, and thus the basis for the last two training sessions. Challenge results and winners will be presented in late August at the 2021 IEEE AI Test Conference.

The SVL Simulator team is thrilled to be a partner on this project. You can find more information about the Challenge here.