Speaker:

Dr. Zheng Dong, Assistant Professor, Computer Science, Wayne State University

Title:

Turbocharging real-time DNN workloads for autonomous driving: Optimization on heterogeneous embedded multicore platforms

Abstract:

The landscape of autonomous vehicles is poised to reach a market valuation of $36.15 billion by 2025, delineated into $31.02 billion for partially autonomous vehicles and $5.13 billion for fully autonomous models. Despite significant strides made in integrating these vehicles across various sectors, their widespread adoption in everyday life remains a distant objective due to critical safety concerns. A pivotal aspect driving these vehicles is the capacity for sensors to amass approximately 4TB of data daily, a vital resource for monitoring driving conditions. However, the true challenge lies in real-time processing of this colossal data to swiftly make critical driving decisions. This obstacle is exacerbated by the computational demands of current Deep Neural Network (DNN) perception algorithms, further constrained by limited computing resources on embedded systems.

Biograpy:

Dr. Zheng Dong is an Assistant Professor in the Computer Science Department of Wayne State University, where he is the co-director of the CAR Lab, the director of the Real-time Systems Group and the WSU site co-director of the NSF IUCRC eCAT center. His educational background includes a Ph.D. in Computer Science from the University of Texas at Dallas, a master’s degree in computer software engineering from the University of Science and Technology of China, and a bachelor’s degree in computer science from Wuhan University. He is full of enthusiasm about realizing timing-predictable, safety-critical cyber-physical systems through algorithm-SoC co-design, with autonomous driving being the focused application domain. His research has been funded by the National Science Foundation.