Autonomous deployment of heterogeneous mobile agents with arbitrarily anisotropic sensing patterns

Abstract

This article examines the problem of autonomous optimal deployment of the nodes in a sensor network with heterogeneous anisotropic patterns. The sensing footprints of the latter are allowed to be any arbitrary convex set instead of circular, while the members of the network are considered heterogeneous, as far as the scaling factor of the aforementioned pattern is concerned. The proposed coordination algorithm relies on suitable partitioning of the sensed space, based on certain Helly-type theorems for planar convex curves, guaranteeing distributed information flow. Results are further confirmed via simulation studies in comparison to circular-approximationbased ones.

Cooperative positioning/orientation control of mobile heterogeneous anisotropic sensor networks for area coverage

“…In contrast to the majority of the works in coverage control, the sensing patterns considered in this work are not required to be node-centred circular ones (standard disc model). In regard to the extensions provided in comparison with the authors' work [12], in this article:…”

“…An extension to this issue has been recently proposed by the authors [12], bypassing the standard uniform model. More specifically, a partitioning scheme (followed by the proper coordination algorithm) has been proposed, that takes into account the maximal inscribed convex set, rather than the circular one.…”

“…Unlike the cells resulting from distance-based partitioning techniques [15] or certain pattern-based ones [12], the sets defined in (5) are, in the generic case, noncompact.…”

“…Extension to arbitrary patterns has been introduced by the authors in [12], where partitioning was performed in a pattern-based manner rather than a distance-based one. However, in that case the sensors were considered to have common orientation, while no rotation of the sensing domain was taken into account.…”

“…This article is an extension of the authors' work [12], by allowing the rotation of the patterns, thus dropping the conservative demand for common orientation. Furthermore, the sensing footprint for the nodes is not required to be a convex set (since the homothetic requirement [14] is dropped), while heterogeneity in the network lies in the scaling factor of the (common) pattern among the nodes.…”