ECE Department Seminar 

 

Planning, Control and Test of Autonomous Vehicles 

 

Shaobing Xu 

Ph.D., Assistant Research Scientist 

Department of Mechanical Engineering & Mcity 

University of Michigan, Ann Arbor 

 

Tuesday, March 9, 2021 

10:30 am – 11:30 am 

Zoom Meeting Link: 

https://gmu.zoom.us/j/96509096651 

 

Abstract: Autonomous vehicles are a promising technology that will redefine future transportation. In this talk, I will present the motion control, planning, and test system I developed at Mcity, but focus on three main challenges. The challenge of motion control is posed by the high time delay, resulting in nonlinearity and worse accuracy and stability. I will introduce a delay-and-dynamics-aware preview control algorithm for improved performance, which can generate, in the format of closed-form, steering compensation for the delay and feedforward steering of the future path curvatures. In motion planning, the challenge being followed is how to generate a scalable algorithm stack that adapts to diverse traffic scenarios. I will present a hierarchical planning framework and focus on a deterministic sampling trajectory generation algorithm, which resolves the nonholonomic planning problem with bounded computing time and near-optimality. Deployments and experiments on the Mcity self-driving car fleet will also be involved.  To test the motion control and planning system, we target the challenge of how to create diverse traffic scenarios at the sensor layer. I will present our solution, an AI-driven sensor-level augmented reality (AR), enabling a CAV to experience various weather/traffic conditions semi-realistically and efficiently.  Finally, I will briefly discuss some salient research directions of interest to me, such as safe autonomy against uncertainty and synergism of cyberphysical multi-agent systems. 

 

Bio: Shaobing is an assistant research scientist (from 1/2019) and was a postdoc (10/2016-12/2018) at the Department of Mechanical Engineering and Mcity, University of Michigan, Ann Arbor. He received a Ph.D. degree in Mechanical Engineering from Tsinghua University in 2016. His research lies in the learning, control, and design of autonomous systems with an emphasis on connected automated vehicles. Shaobing received the outstanding Ph.D. dissertation award (2016) and the Best Paper Award of AVEC’18 (2018). He has served as the PI or co-PI of 3 research projects, with total funding of 1.2 million dollars. He has more than 30 technical publications, including 24 in journals and transactions.