Siva Hari
Senior Research Scientist, NVIDIA

Invited Talk: Resilience and Safety for Autonomous Vehicles
Slides
Video
Abstract:
Autonomous vehicles (AVs) employ a sophisticated set of sensors, computer systems, and algorithms to perceive surroundings, localize, plan, and control the vehicle. These complex systems must meet high-levels of safety as they are deployed to transport or interact with humans. To demonstrate the safety of an AV, it must be designed and tested to tolerate hardware errors, operate within the constraints of the hardware platform (e.g., limited power), and operate in challenging driving conditions. While standards have been established to facilitate system development to address some of these problems (e.g., tolerating hardware errors), addressing others have remained open-ended. In this talk, the presenter will summarize some of his recent work that aims to improve safety of AVs by addressing each of the above listed areas, and discuss interesting future research directions.
Biography:
Siva Hari is a Senior Research Scientist in the Computer Architecture Research Group at NVIDIA. His current research focus is on making GPUs resilient through architecture and software level solutions. He obtained his Ph.D. and M.S. degrees from the Computer Science Department at University of Illinois at Urbana-Champaign and Bachelor’s degree from the Computer Science and Engineering Department at IIT Madras.
He was selected to receive the 2014 David J. Kuck Outstanding Ph.D. Thesis Award by the Computer Science Department at University of Illinois at Urbana-Champaign. He also received the W.J. Poppelbaum Memorial Award by the CS Department at University of Illinois in 2012 for academic merit and creativity in computer architecture. One of the papers he co-authored was selected for the IEEE Micro’s Top Picks 2013 and another was recognized as the Best Paper Award Runner-up at DSN 2018.