On Mobile Ad Hoc Networks for Coverage of Partially Observable Worlds

TL;DR

The paper tackles online deployment of mobile agents to form a connected MANET in initially unknown environments by framing the problem as the Partially Observable Cooperative Guard Art Gallery Problem (POCGAGP). It introduces two algorithms, CADENCE (centralized) and DADENCE (decentralized), and validates them through a large-scale benchmark of 1,500 orthogonal dungeon environments, augmented with a quadtree-based complexity measure. The results show that both approaches achieve complete coverage with maintained connectivity, with CADENCE excelling in minimising steps and DADENCE maximising decentralization and agent-efficiency. The work demonstrates that geometric abstractions, particularly valid corners and visibility graphs, enable scalable, provably robust MANET deployment in partially observable spaces, with promising avenues for extending to non-orthogonal and 3D domains.

Abstract

This paper addresses the movement and placement of mobile agents to establish a communication network in initially unknown environments. We cast the problem in a computational-geometric framework by relating the coverage problem and line-of-sight constraints to the Cooperative Guard Art Gallery Problem, and introduce its partially observable variant, the Partially Observable Cooperative Guard Art Gallery Problem (POCGAGP). We then present two algorithms that solve POCGAGP: CADENCE, a centralized planner that incrementally selects 270 degree corners at which to deploy agents, and DADENCE, a decentralized scheme that coordinates agents using local information and lightweight messaging. Both approaches operate under partial observability and target simultaneous coverage and connectivity. We evaluate the methods in simulation across 1,500 test cases of varied size and structure, demonstrating consistent success in forming connected networks while covering and exploring unknown space. These results highlight the value of geometric abstractions for communication-driven exploration and show that decentralized policies are competitive with centralized performance while retaining scalability.

Figures (23)

• Figure 1: Three robots forming a MANET in a partially observed indoor environment. The bright yellow areas represent observed cells, while the shadowed regions are uncovered. The two holes (polygons in the middle) and outer walls are physical barriers that cannot be covered. Green arrows indicate the agents' communication links.
• Figure 2: Transformation of the polygon $W$ into $W'$ to remove all invalid reflexive corners (in green) while retaining only the valid corners (in red). The added walls for $W'$ are in dark blue.
• Figure 3: Forbidden diagonal line-of-sight: Assumption \ref{['asp3']} rules out configurations where a top-left cell sees a bottom-right cell through a single diagonal corner between two obstacle cells (cf. Definition \ref{['WFOV']}).
• Figure 4: Flowchart illustrating the DADENCE algorithm. The numbered blocks (1--11) represent its main components.
• Figure 5: Illustration of the queue-jump step corresponding to block 10 in Figure \ref{['flowchart']}. The deployment point is marked in red, and the agents are shown in black. The black arrows represent the individual agent shifts towards $x^*$, while the red arrow depicts the effective result: the agent at the deployment point effectively leaping over the subsequent agent(s).

Theorems & Definitions (29)

• Definition 1: Partially-observable CGAGP
• Definition 2: Visibility graph
• Definition 3: Interior angle
• Definition 4: Orthogonal polygon
• Definition 5: Orthogonal world
• Remark 1
• Remark 2
• Definition 6: Field of view
• Definition 7
• Definition 8: Valid corner