Robotics

Penn Jerboa robot achieves 1.77 m/s with novel hip-powered hopping

New control method turns pitch torque into hopping energy, hitting 8.85 leg lengths/second.

Deep Dive

Researchers at the University of Pennsylvania — Shane Rozen-Levy, Griffon McMahon, and Daniel Koditschek — have introduced a novel control strategy for pitch-unlocked planar monopeds, presented in the paper "Hip Energized Monopedal Hopping." The key insight is that reaction torques from a conventional PD + feedforward stabilizing controller can be recruited to counteract energy losses from damping. By moving the mass center, the controller increases pitch stabilization torque, thereby injecting energy directly into the gait. A new stepping policy then redistributes energy between the radial and angular degrees of freedom to achieve a user-specified balance between steady-state fore-aft speed and apex height. The team validated their approach using hybrid averaging analysis, deriving closed-form expressions for the fixed points and eigenvalues of the resulting gait, which reveal how physical and control parameters influence performance. Simulations on a generic 5-link biped and a high-fidelity model of the Penn Jerboa confirmed these analytical predictions.

Physical experiments on the Penn Jerboa robot demonstrated stable locomotion speeds ranging from 1.02 m/s to 1.77 m/s (5.10 to 8.85 leg lengths per second), with behavior effectively matching the mathematical analysis. The paper includes 35 pages and 11 figures, exploring the interplay between control design and mechanical design for efficient legged locomotion. This work is significant because it shows how a controller can harness stabilizing torques that are typically viewed as a side-effect of keeping the body upright, turning them into a source of propulsive energy. This could lead to more energy-efficient and agile hopping robots that require less external actuation per step, potentially extending battery life and enabling longer autonomous operations in real-world environments.

Key Points
  • Achieved stable hopping at speeds from 1.02 to 1.77 m/s (5.10–8.85 leg lengths/s) on the physical Penn Jerboa robot.
  • Hybrid averaging analysis produced closed-form fixed points and eigenvalues that predict gait stability and performance.
  • Controller recruits pitch stabilization torque by shifting mass center, counteracting damping losses without additional leg actuation.

Why It Matters

Efficient hopping control could enable more agile and energy-efficient legged robots for real-world navigation.

📬 Get the top 10 AI stories daily