@inproceedings{1266,
  author = {Julia Ebert and Florian Berlinger and Bahar Haghighat and Radhika Nagpal},
  title = {A Hybrid PSO Algorithm for Multi-Robot Target Search and Decision Awareness},
  abstract = {<p>Groups of robots can be tasked with identifying a location in an environment where a feature cue is past a threshold, then disseminating this information throughout the group {\textendash} such as identifying a high-enough elevation location to place a communications tower. This is a continuous-cue target search, where multi-robot search algorithms like particle swarm optimization (PSO) can improve search time through paral- lelization. However, many robots lack global communication in large spaces, and PSO-based algorithms often fail to consider how robots disseminate target knowledge after a single robot locates it. We present a two-stage hybrid algorithm to solve this task: (1) locating a target with a variation of PSO, and (2) moving to maximize target knowledge across the group. We conducted parameter sweep simulations of up to 32 robots in a grid-based grayscale environment. Pre-decision, we find that PSO with a variable velocity update interval improves target localization. In the post-decision phase, we show that dispersion is the fastest strategy to communicate with all other robots. Our algorithm is also competitive with a coverage sweep benchmark, while requiring significantly less inter-individual coordination.</p>},
  year = {2022},
  journal = {IEEE RSJ International Conference on Intelligent Robots and Systems (IROS)},
  address = {Kyoto, Japan},
}