Robot reflex policy improves contact tasks like peg insertion and surface following
Learning Reflexive Behavior for Contact-Rich Manipulation
Robotics
Summary
Robots often struggle when touching objects because they don’t sense forces or shapes well. The authors created a reflex system that learns from simple virtual examples of springs, flat surfaces, and rails, using past movements to control robot joints without force sensors. This helps the robot handle delicate tasks like fitting pegs very precisely and following rough surfaces while keeping forces low. Their approach works under other robot controllers and improves success rates in tough tasks.
What this means in practice
- •For industrial robot programmers: Enhance robot assembly tasks such as precise peg insertion by adding a reflex layer that reduces contact force and improves success rates without extra sensors.
- •For robot teleoperation teams: Improve teleoperated robots’ contact handling by using the reflex policy to reliably manage environmental constraints during delicate manipulations.
Authors
Quan Nguyen, Yunho Kim, Joonho Lee
Abstract
During contact-rich manipulation, interactions between a robot and its environment carry information about local geometry: a surface prevents penetration, a bore guides a peg. A controller that exploits these interactions can comply with environmental constraints while preserving task intent. Robot-earning systems commonly use position, hybrid force-position, or Cartesian impedance control. Their prescribed tracking objectives, stiffness, or force-control directions may not match local constraints and may degrade performance. We learn a proprioceptive reflex policy in simulation on three simple interaction primitives: a spring, a plane, and a rail. The policy maps task-space commands to joint-position targets using state history, without direct force or geometrical measurements. Once trained, it serves as a frozen execution layer beneath higher-level controllers. We evaluate it in dual-arm box lifting, peg insertion, and surface following. In box lifting, the reflex kept the force below the threshold while the baseline failed. In rough-surface following it stayed below the 10 N reference, on par with tuned hybrid force-position control. In 0.02 mm peg insertion It reduced mean estimated contact force to less than half that of the baseline while increasing hardware success rates from at most 22 % to 36-58 %. By separating contact response from command generation, the reflex policy provides motion planners, learned policies, and teleoperators with robust contact-rich execution.