Research
A Nonasymptotic Theory of Gain-Dependent Error Dynamics in Behavior Cloning
arXiv:2604.14484v1 Announce Type: new Abstract: Behavior cloning (BC) policies on position-controlled robots inherit the closed-loop response of the underlying PD controller, yet the effect of control
arXiv:2604.14484v1 Announce Type: new Abstract: Behavior cloning (BC) policies on position-controlled robots inherit the closed-loop response of the underlying PD controller, yet the effect of controller gains on BC failure lacks a nonasymptotic theory. We show that independent sub-Gaussian action errors propagate through the gain-dependent closed-loop dynamics to yield sub-Gaussian position errors whose proxy matrix X_infty(K) governs the failure tail. The probability of horizon-T task failure factorizes into a gain-dependent amplification index Gamma_T(K) and the validation loss plus a generalization slack, so training loss alone cannot predict closed-loop performance. Under shape-preserving upper-bound structural assumptions the proxy admits the scalar bound X_infty(K)preceqPsi(K)ar X with Psi(K) decomposed into label difficulty, injection strength, and contraction, ranking the four canonical regimes with compliant-overdamped (CO) tightest, stiff-underdamped (SU) loosest, and the stiff-overdamped versus compliant-underdamped ordering system-dependent. For the canonical scalar second-order PD system the closed-form continuous-time stationary variance X_infty^{c}(alpha,eta)=sigma^2alpha/(2eta) is strictly monotone in stiffness and damping over the entire stable orthant, covering both underdamped and overdamped regimes, and the exact zero-order-hold (ZOH) discretization inherits this monotonicity. The analysis provides the first nonasymptotic explanation of the empirical finding that compliant, overdamped controllers improve BC success rates.
Related
- Robust Adaptive Backstepping Impedance Control of Robots in Unknown Environments
- Iteratively Learning Muscle Memory for Legged Robots to Master Adaptive and High Precision Locomotion
- Model-Based Reinforcement Learning Exploits Passive Body Dynamics for High-Performance Biped Robot Locomotion
- DA-PTQ: Drift-Aware Post-Training Quantization for Efficient Vision-Language-Action Models
- CLAW: Composable Language-Annotated Whole-body Motion Generation
Source: arXiv cs.RO | 2026-04-17