Event
Proposes a locomotion strategy diagnosis method based on effective rank analysis by gait phase
1 reports · 1 sources
Felipe Tommaselli et al. propose using the effective rank of the network Jacobian to diagnose the representation health of legged locomotion reinforcement learning policies, finding that conditioning on gait phase exposes representational structure invisible to the global effective rank. The study observes that architectures with normalization plus residual connections allocate more than one effective-rank dimension to swing-dominated states than plain MLPs do. Based on this, the proposed diagnostic method reduces joint jitter by up to about 2 times in simulation and on a physical Spot robot; the paper has been accepted to CoRL 2026.
Report timeline
- Improving legged locomotion policy health via gait-phase-conditioned effective rank of the network Jacobian
This study analyzes reinforcement learning policies for legged locomotion using the effective rank of the network Jacobian, finding that conditioning on gait phase exposes representation structures invisible from the global rank. Architectures with observation normalization and residual connections allocate more than one additional effective-rank dimension to swing-dominant states than vanilla MLPs do. Based on this, the proposed diagnostic method reduces joint jitter by up to about 2x in simulation and on a physical Spot, and the paper has been accepted to CoRL 2026.
arXiv cs.RO ·