Project: #96 F1/10 Autonomous Racing Course and Competition Progress Report - Reporting Period Ending: Sept. 30, 2022 Principal Investigator: Rahul Mangharam Status: Active Start Date: July 1, 2017 End Date: June 30, 2023 Research Type: None Grant Type: Education Grant Program: FAST Act - Mobility National (2016 - 2022) Grant Cycle: Mobility21 - University of Pennsylvania Progress Report (Last Updated: Oct. 4, 2022, 3:32 p.m.) % Project Completed to Date: 50 % Grant Award Expended: 50 % Match Expended & Document: 50 USDOT Requirements Accomplishments This effort focuses on the development of a toolchain called TunerCar which jointly optimizes safe driving strategy, planning methods, control algorithms, and vehicle parameters for an autonomous racecar. In this paper, we detail the target hardware, software, simulators, and systems infrastructure for this toolchain. Our methodology employs a parallel implementation of CMA-ES which enables simulations to proceed 6 times faster than real-world rollouts. We show our approach can reduce the lap times in autonomous racing, given a fixed computational budget. For all tested tracks, our method provides the lowest lap time, and relative improvements in lap time between 7-21%. We demonstrate improvements over a naive random search method with equivalent computational budget of over 15 seconds/lap, and improvements over expert solutions of over 2 seconds/lap. We further compare the performance of our method against hand-tuned solutions submitted by over 30 international teams, comprised of graduate students working in the field of autonomous vehicles. Finally, we discuss the effectiveness of utilizing an online planning mechanism to reduce the reality gap between our simulation and actual tests. By developing tools that automatically tune reusable core autonomy components, we are able to both reduce the cost of developing high-performance autonomous vehicles and also improving the safety across a range fo driving scenarios. The results of this work have been published in over 6 peer-reviewed journals and conferences. Impacts We have established the Autoware Center of Excellence for Autonomous Driving the the Pennovation Center, Philadelphia to serve as a physical laboratory for collaborative research and development of open-source autonomous vehicle software. The Autoware Foundation is a non-profit consortium of over 60 industry, academic and government members that develop autonomous driving software as a community. One approved patent - Control of multi-drone fleets with temporal logic objectives. US20200348696A1. 2020-11-05 One PhD student, Matthew O-Kelly, started a company https://trustworthy.ai which was acquired by Waymo/Google in 2021. His thesis work was supported in part by this DoT grant. Other We are organizing the 2nd Workshop on Challenges and Opportunities in Autonomous Racing and the 10th Autonomous Racing GrandPrix at the 2022 IEEE International Conference on Robotics and Automation (ICRA). https://icra2022.f1tenth.org/ and https://icra2022-race.f1tenth.org/ We have released a new course website for anyone interested in teaching autonomous racing as a way to explain tradeoff in safety and performance with autonomous vehicles - https://courses.f1tenth.org/ The community website https://f1tenth.org has a high number of contributors and visitors. This effort has over 60 university partners internationally that either use the F1Tenth platform in research/teaching or participate in the international racing competitions. Outcomes New Partners The Autoware Foundation - https://autoware.org Issues N/A