Papers
20
Total Citations
374
H-Index
11
About
Zishen Wan is at the forefront of redefining robotic computing through hardware-software co-design, with a primary focus on FPGA-based acceleration for autonomous systems. His work bridges the critical gap between computational demands and real-time performance in robotics, from drones and self-driving cars to manipulators. His highly cited survey on FPGA-based robotic computing (112 citations) established a foundational roadmap for the field, while his "Formula-1" roofline model provides a groundbreaking visual framework for understanding the interplay between sensing, computation, and dynamics in aerial machines. Wan has made significant contributions to energy-efficient, runtime-reconfigurable accelerators for core robotic functions like localization and SLAM, as well as stereo matching. He also addresses the critical challenge of system resilience, developing frameworks like MAVFI to analyze and improve fault tolerance in learning-based navigation systems. His recent work includes ORIANNA, an accelerator generation framework for optimization-based robotics, and RobotPerf, an open-source benchmarking suite for evaluating robotics computing performance across diverse hardware platforms. With over 300 total citations and a portfolio spanning from hardware acceleration to fault analysis, Wan is driving the next generation of efficient, safe, and high-performance autonomous machines.
Research Focus
Key Achievements
Top Papers
- 1A Survey of FPGA-Based Robotic Computing112 citations · 2021
- 2
- 3
- 4
- 5Robotic Computing on FPGAs22 citations · 2021
- 6Analyzing and Improving Fault Tolerance of Learning-Based Navigation Systems22 citations · 2021
- 7
- 8
- 9
- 10