Event
A Robot Metareasoning Method with Adaptive Computational Budgets
1 reports · 1 sources
Lorenzo Magnino et al., in the arXiv paper "Learning Robot Metareasoning," propose modeling the determination of computational budgets in real-time robot control as a semi-Markov decision process (SMDP), where a single decision simultaneously determines the action commitment duration and the solver allocation time, and train a reinforcement learning policy conditioned on the current state, the in-progress plan, and low-cost planner statistics. On three types of tasks—ground robot navigation, agile quadrotor flight, and contact-rich manipulation—the authors claim that this policy can adjust the budget according to scene demands and outperforms fixed-budget baselines.
Report timeline
- Learning Robot Metareasoning: Adaptive Computation Budgets for Real-Time Control
The arXiv paper Learning Robot Metareasoning proposes modeling the determination of the optimal computation budget as a semi-Markov decision process (SMDP), in which a single decision jointly determines the robot's action commitment duration and the time allocated to the solver. The authors trained a reinforcement learning policy conditioned on the current state, the executing plan, and low-cost planner statistics. Across three task types—ground robot navigation, agile quadrotor flight, and contact-rich manipulation—the policy adjusts its budget according to scene demands and outperforms fixed-budget baselines.
arXiv cs.RO ·