Time-Optimal Ergodic Retrieval

Name: Time-Optimal Ergodic Retrieval
Uploaded: 2026-10-19T19:11:42.687992+00:00
Duration: 320 s
Description: This work proposes a new method, Time-Optimal Ergodic Retrieval (TOER), to solve search tasks where a robot must find an object and return to a base quickly. TOER models detection as a Poisson process and minimizes total mission time by including the necessary return phase in the objective. Results show TOER is up to 19% faster than existing methods, especially when returning is costly.

This work proposes a new method, Time-Optimal Ergodic Retrieval (TOER), to solve search tasks where a robot must find an object and return to a base quickly. TOER models detection as a Poisson process and minimizes total mission time by including the necessary return phase in the objective. Results show TOER is up to 19% faster than existing methods, especially when returning is costly.

Open the full episode