LP-NavOA: Integrated Local Navigation and Obstacle Avoidance for Humanoid Robots under Limited Perception
2026-06-22 • Robotics
Robotics
AI summaryⓘ
The authors developed a navigation method called LP-NavOA to help humanoid robots walk safely and efficiently around obstacles without needing detailed maps. They first trained a robot movement controller and then taught a planner to adjust its walking direction using simple local sensors and goal directions. This approach improved the robot's ability to avoid obstacles and reach goals on time in simulated environments. Tests also showed the method works on real hardware without constant manual control.
humanoid robotslocal navigationobstacle avoidanceproprioceptionproximal policy optimizationA-star algorithmwaypoint planningreinforcement learningMuJoCo simulatorrobot control
Authors
Yukun Luo, Jianjun Ma, Yuyao Min, Jinzhe Li, Kaihong Huang, Peng Li
Abstract
Humanoid local navigation in cluttered environments must jointly resolve obstacle avoidance, sparse-goal recovery, and stable whole-body locomotion under short-range and partially observable sensing. Explicit planner-control decompositions introduce latency and can mismatch agile humanoid command-tracking limits, while purely reactive controllers may lose the goal after obstacle occlusion. We present LP-NavOA, a limited-perception navigation and obstacle-avoidance framework for humanoid robots. A raycast-conditioned perception-action proximal policy optimization (PPO) locomotion backbone is first trained with a robot-centered circular heading-speed command and a shared command-side safety filter. With this backbone frozen, A-star and waypoint teachers generate rollouts for distilling a recurrent local planner that overwrites only the heading command at deployment, leaving the whole-body policy intact. At runtime, LP-NavOA uses proprioception, short-range local range sensing, and a body-frame goal direction, requiring no global map, waypoint stream, or external planner. In MuJoCo open-wall and indoor layouts, the distilled planner produces obstacle bypassing and post-avoidance goal recovery, raising teacher-calibrated on-time arrival from 38--40\% to 85--97\% and reducing brush/contact-heavy progress relative to a backbone-only controller. Ablations show that dynamic route shaping, teacher-active data collection, and the circular command interface are important for navigation efficiency and for training the 3.0\,m/s backbone. A Unitree G1 deployment analysis demonstrates hardware executability without continuous joystick steering.