Papers
2
Total Citations
22
H-Index
2
About
Ziyun Li is pioneering energy-efficient, fully-on-chip intelligence for micro-robotic vision systems. Their research centers on domain-specific system-on-chip (SoC) design, embedded memory technology, and hybrid processing architectures that enable autonomous navigation at the edge. Li’s major contribution is the development of flexible SoCs that seamlessly handle both convolutional neural networks (CNNs) and classic vision tasks, eliminating reliance on off-chip components. Their 2022 work, cited 16 times, introduced a 22nm SoC with 2MB eMRAM achieving 3.5 TOPS/W—a breakthrough for retentive, on-chip weight storage in micro-robots. The follow-up RoboVisio SoC (2024) further refined this architecture for fully autonomous navigation, demonstrating how hybrid processing elements can balance efficiency and flexibility. By integrating non-volatile memory directly into the compute fabric, Li’s designs address critical power and area constraints in millimeter-scale robots. Their work has immediate implications for swarm robotics, environmental monitoring, and medical devices where real-time vision must be processed without cloud connectivity. Li’s achievements represent a significant step toward truly intelligent, self-contained micro-robots.
Research Focus
Key Achievements
Top Papers
- 1
- 2