Kohei Honda
Papers
2
Total Citations
9
H-Index
2
About
Kohei Honda is a rising researcher in autonomous robotic navigation, with a focus on integrating learning-based methods with classical control to overcome long-standing challenges in the field. His work centers on two critical problems: adaptive replanning and local minima avoidance. In his highly cited 2024 paper, "When to Replan? An Adaptive Replanning Strategy for Autonomous Navigation using Deep Reinforcement Learning," Honda introduced a novel framework that uses deep reinforcement learning to dynamically decide when to trigger global replanning—a key bottleneck in hierarchical navigation systems. This work has already garnered 6 citations, reflecting its immediate relevance. Building on this, his 2025 paper, "Towards Local Minima-free Robotic Navigation," addresses the persistent issue of model-based controllers becoming trapped in suboptimal solutions. By augmenting Model Predictive Path Integral (MPPI) control with repulsive potentials, Honda proposes a method that maintains solution quality while escaping local minima, a significant departure from purely reactive approaches. With 3 citations in its first year, this work is poised to influence future navigation systems. Honda’s contributions are notable for their practical, real-world applicability, bridging the gap between theoretical control and robust autonomous operation.
Research Focus
Key Achievements
Top Papers
- 1
- 2