Frontend
Learning and Control in Physically Coupled Dynamical Systems
arXiv cs.RO0 views
arXiv:2611.23502v1 Announce Type: new Abstract: We introduce TAG (Testbed for Autonomous Games), a compact, low-cost robotic platform for studying learning and control in physically coupled dynamical systems. TAG combines an actively tilted and steered ball-in-maze board with a top-mounted rotary inverted pendulum. The pendulum's reaction torque and vibration disturb the board and marble, while the board's tilt, driven by the maze-solving policy, simultaneously disturbs the pendulum's balance. The main challenge in learning control policies for each subsystem is accounting for the impact of one subsystem's dynamics on the other's learning process. To address this challenge, we present complementary reinforcement learning paradigms, leveraging model-based and model-free approaches to learn the control policies for the maze and pendulum, respectively. Our key finding is that introducing the pendulum causes an immediate, sharp drop in the maze policy's success rate, but the policy adapts online and recovers to its pre-coupling performance within a small number of learning episodes. Experimental results show that both policies, trained independently, successfully control the maze board and the pendulum despite their physical coupling.
Source:https://arxiv.org/abs/2611.23502PreviousBiCAT: Bimanual Coordination Action TokenizationNextRageVLA: Augmentation-Trained Perception and Action for State-of-the-Art Robustness on LIBERO
Related posts
Python Day 4: Bitwise Operators and Ternary Operator
0dev.to
The Complete CBT Mental Health Toolkit: Free Tools for Developers
0dev.to
Service Worker Keeps Stopping? That's the Contract, Not a Bug
0dev.to
iPhone Duo vs. Samsung Galaxy Z Fold8: Which is Better?
0MacRumors
Show HN: A 145-line 2D rigid-body physics engine in TypeScript and Canvas demo
0Hacker News Show
Test analytics consent across a planner visit and a reload
0dev.to