Humanoids improve cart pushing using adaptive object motion planning
Belief-Overlap Partial Inversion: Object-Centric Flow Replanning for Humanoid Loco-Manipulation
Robotics
Summary
Moving an unknown object is hard because you don’t know exactly how it will respond. The authors studied robots pushing carts whose weight and wheels were unknown at first. They created a system that continuously updates its understanding while planning the movement, but only changes the plan when necessary to avoid interrupting ongoing actions. Their method led to better accuracy in simulations and worked well on a real robot pushing a cart.
What this means in practice
- •For robotics engineers: Enable robots to push unknown carts more accurately by updating plans based on sensed object behavior without disrupting ongoing movements.
- •For automated warehouse teams: Improve mobile robots' ability to maneuver uncertain loads by integrating real-time sensing with adaptive motion planning.
Authors
Chen Yu, Sudarshan S Harithas, Juan Borbon, Dingqi Zhang, Jiuguang Wang
Abstract
People often learn how an unfamiliar object moves through a small exploratory push, using its response to guide their next motion. Inspired by this behavior, we study humanoids pushing carts whose mass and caster stiffness are initially unknown. Our object-centric framework closes online estimation, flow-based planning, and whole-body control into a loop, creating a tension: replacing the plan whenever the belief shifts disrupts the maneuver under way, while keeping it strands the robot on a path its evidence has ruled out. Belief-Overlap Partial Inversion (BOPI) scales revision to agreement, inverting the flow under the old belief only as far as the beliefs disagree before reconstructing under the new one. Across 18 simulated goals it lowers pooled final position and yaw error by roughly 33% and 47% against fixed-depth replanning, and on hardware it translates and turns a non-holonomic cart.