Max Robotics — Global Service Robot Directory + US Certification

About

Kwanho Kim is a leading researcher in energy-efficient, high-performance computer vision architectures for intelligent robotic systems. His work focuses on developing specialized hardware processors that enable real-time object recognition under severe power constraints. Kim’s most significant contributions include the design of a groundbreaking 81.6 GOPS object recognition processor, which leverages a memory-centric Network-on-Chip (NoC) architecture to efficiently support both task-level and data-level parallelism for mobile robot applications. He further advanced the field by creating a 201.4 GOPS multi-object recognition processor that incorporates a bio-inspired neural perception engine, mimicking the human visual attention mechanism to dramatically reduce computational costs. This work, which has garnered over 15 citations, demonstrated the ability to recognize multiple objects in a single frame—a critical step beyond prior single-object systems. Kim’s research has consistently pushed the boundaries of power efficiency, achieving 496mW operation at peak performance. In more recent work, he has explored dynamic obstacle avoidance using Particle Swarm Optimization combined with Kalman Filters, introducing the novel concept of an inertia rate to balance local and global search. His contributions are foundational to the development of truly autonomous, vision-guided mobile robots.