In This Section
Introduction
Introduction

Watch the video from May Mobility's presentation

Join CSAIL Alliances for a talk featuring May Mobility's Edwin Olson.

Abstract
The hype cycle for autonomous vehicles has reached the disillusionment phase. Why have robo-taxis not lived up to their promises? While many point to the many subtle/rare edge-cases that a self-driving vehicle might encounter (like policemen directing traffic or the moral brain teasers known as trolley problems), the reality is more mundane: self-driving cars perform poorly even in routine interactions with other road users. The problem is that other road users dynamically react to the actions of the self-driving car; those interactions must not only be understood but also actively manipulated.

In this talk, I'll describe a planning approach called Multi-Policy Decision Making (MPDM) that helps autonomous vehicles understand and exploit the coupled behavior of the agents on a roadway. I'll present results from controlled experiments using small robots as well as comment on lessons learned through the commercial deployment of MPDM in autonomous vehicles on public roads in downtown Detroit and Columbus. We believe this technology can help push self-driving vehicles past the disillusionment phase, starting with shuttles and eventually tackling robo-taxis.

About the Speaker
Edwin Olson is an Associate Professor of Computer Science and Electrical Engineering at the University of Michigan, and co-founder/CEO of May Mobility, Inc., which develops self-driving shuttles. He earned his PhD from MIT in 2008 for work in robot mapping.  He has worked on autonomous vehicles for over a decade, including work on the 2007 DARPA Urban Challenge, vehicles for Ford and Toyota Research Institute, and now May Mobility. His academic research includes work on perception, planning, and mapping. He was awarded a DARPA Young Faculty Award, named one of Popular Science's "Brilliant 10", and was winner of the 2010 MAGIC robotics competition. He is perhaps best known for his work on AprilTags, SLAM using MaxMixtures and SGD, and Multi-Policy Decision Making.

CSAIL Lecture Edwin Olson
Speakers