Elcano - Starting the Project: Lessons Learned

Tyler Folsom from University of Washington's Elcano project discusses the lessons he learned from building self driving vehicles. He draws upon his experiences from two teams that participated in DARPA Grand Challenge and other similar projects.

Transcript

1. The Elcano Project Open Source Autonomy Tyler C. Folsom, Ph.D, P.E. QUEST Integrated, Inc.
2. Goals: • Promote benefits of self-driving vehicles • Open-source self-drive technology • Safe, sustainable transportation system • One thousand mpg equivalent
6. Target Vehicles: • Tricycles • Human powered vehicles (HPV) • Automobiles • Modular buses • Personal Rapid Transit (PRT) • Trucks • Wheelchairs • Radio-controlled toy cars • Lawn mowers, vacuum cleaners,etc.
7. Current Status Elcano #1 is in Seattle (Jigsaw Renaissance) All mechanical, electrical and software systems for drive-by-wire are in place. Second generation electronics for AI, controls, and panel are under design. First generation navigation and path finding software are in place. Demo scheduled for July at CORVA, Stanford Elcano #2 is in Bothell (U of Washington) Most parts are on-hand Configuration is starting Simulator is not configured
8. Pitfalls: Project complexity. Unexpected electrical behavior. Mechanical glitches. Low level control can eat up resources. The fun stuff is the AI, but you may burn out before getting there.