Hybrid RF Propagation Model using ITM and Gaussian Processes for Communication-Aware Planning

Spencer Watza, Ramya Kanlapuli, Eric Frew

May 23, 2017 (modified: Jul 17, 2017) RSS 2017 RCW Workshop - Proceedings Submission readers: everyone
  • Student paper: Yes
  • Abstract: Communication-aware robotics is a necessary step to successfully implement multi-agent missions with UAS. While distributed systems have been investigated, previous research focused on using simple RF propagation models or ignored it with assumptions. To develop higher understandings of the affects of wireless communication between agents, higher fidelity models are required. This work proposes a hybrid approach for simulation and in-situ mapping of the Path Loss fields. The architecture combines a prediction step using traditional propagation models with a learning step to correct errors. A version of this architecture is used for simulation and shows the plausibility of using Gaussian Processes to find the structure of an RF with sparse measurements.
  • TL;DR: A hybrid RF propagation model is designed and simulated using SPLAT! and Gaussian Processes to add simulation capability and in-situ use for Communication Aware applications
  • Keywords: Communication-Aware, RF Propagation, Path Loss Communication Model
  • Conflicts: colorado.edu, wmich.edu