Papers
2
Total Citations
11
H-Index
2
About
Qingsong Liu is a leading researcher in energy-efficient artificial intelligence hardware, with a focus on reconfigurable processors for autonomous systems. His work centers on developing specialized neural network accelerators that enable smart drones and robots to perform real-time object detection and tracking with exceptional energy efficiency. Liu’s major contributions include the design of the RAODAT processor, which integrates dedicated processing engines for bounding box generation and online object learning—capabilities absent in conventional NN processors. His most-cited paper, "An Energy-Efficient Reconfigurable AI-Based Object Detection and Tracking Processor Supporting Online Object Learning" (2022, 6 citations), and its precursor (2021, 5 citations) demonstrate his pioneering approach to combining reconfigurable architecture with adaptive learning, allowing embedded systems to update their object recognition models on the fly without cloud dependency. This innovation addresses critical bottlenecks in autonomous navigation and surveillance, achieving high throughput while minimizing power consumption. Liu’s work has significant implications for edge AI, pushing the boundaries of what compact, low-power processors can accomplish in real-world robotics and IoT applications.
Research Focus
Key Achievements
Top Papers
- 1
- 2