Reproduce “Communication-Aware Robot Execution for Cloud Inference under Spatially Heterogeneous Connectivity” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Fengkai Liu, Yuichi Ohsita, Masayuki Murata, Hideyuki Shimonishi
A communication-aware execution framework coordinates robot motion with cloud inference by selecting where to submit the next request and estimating whether its response can arrive before the current navigation instruction runs out. In Gazebo simulations with spatially varying Wi-Fi, it achieved 94.5%–100% task success across three waypoint spacings; removing request-point selection reduced success to 65.5%–90.9%.
Key insights
- Cloud requests are both timing and motion decisions: scheduling alone does not ensure that a request is sent from a viable location.
- The current task primitive provides a finite execution window that can be used to assess whether a cloud response is likely to arrive in time.
- Maintaining strong connectivity across the entire route is not necessary if the robot can plan suitable submission and retrieval opportunities.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.