Cooperative Exploration under Communication Constraints

The cooperative exploration problem necessarily involves communication among agents,while the spatial separation inherent in this task places fundamental limits on the amountof data that can be transmitted. However, the impact of limited communication on theexploration process has not been fully characterized. Existing exploration algorithms do notrealistically model the tradeoff between expansion, which allows more rapid explorationof the area of interest, and maintenance of close relative proximity among agents, whichfacilitates communication.This thesis develops new algorithms applicable to the problem of cooperative explo-ration under communication constraints. The exploration problem is decomposed into twoparts. In the first part, cooperative exploration is considered in the context of a hierarchi-cal communication framework known as a mobile backbone network. In such a network,mobile backbone nodes, which have good mobility and communication capabilities, pro-vide communication support for regular nodes, which are constrained in movement andcommunication capabilities but which can sense the environment. New exact and approx-imation algorithms are developed for throughput optimization in networks composed ofstationary regular nodes, and new extensions are formulated to take advantage of regularnode mobility. These algorithms are then applied to a cooperative coverage problem.In the second part of this work, techniques are developed for utilizing a given level ofthroughput in the context of cooperative estimation. The mathematical properties of theinformation form of the Kalman filter are leveraged in the development of two algorithmsfor selecting highly informative portions of the information matrix for transmission. Onealgorithm, a fully polynomial time approximation scheme, provides provably good resultsin computationally tractable time for problem instances of a particular structure. The other,a heuristic method applicable to instances of arbitrary matrix structure, performs very wellin simulation for randomly-generated problems of realistic dimension. Thesis Supervisor: Jonathan P. HowTitle: Professor of Aeronautics and Astronautics Thesis Supervisor: Eytan ModianoTitle: Associate Professor of Aeronautics and Astronautics