Coverage Evaluation for the Single Agent and Multi-Agent Persistent Turning Walker Model under a range of Parameter Values
MetadataShow full item record
In our work, we take inspiration from animals, particularly groups of fish to devise coverage algorithms for autonomous robots. Persistent Turning Walker (PTW) is a model which characterized the movement of fish in two-dimensions. We propose the extension of this model to three-dimensions and prove that the model can be used as effective coverage strategy. Methods such as Generalized Cell Mapping (GCM) are used to convert a non-linear model into its respective Markov Chain and then the coverage in a bounded search space can be showed by using reachability theory. For the multi-agent PTW model, techniques such as the steady state behavior can be used to generalize the system and then, prove the reachability of the system. The coverage of the PTW model can also used as an exploration strategy in source seeking algorithm, such as the Hybrid Speeding Up Slowing Down (Hybrid SUSD). Moreover, the aforementioned reachability concept can also extended to the more non-trivial multi-view reachability problem, where an agent has to cover a given grid cell in the search space from all possible directions to ensure reachability. The PTW model can also be used to achieve multi-view coverage by a single agent in two-dimensions. The use of PTW model as a randomized search strategy can enable coverage in GPS-denied environments while being more efficient that other random search strategies, such as the Random Walk and the Correlated Random Walk. The aforementioned work can be extended to evaluate the multi-view coverage concept to three dimensions and show its effectiveness in the multi-agent models as well.