280 likes | 382 Views
Agent-Based Coordination of Sensor Networks. Alex Rogers School of Electronics and Computer Science University of Southampton acr@ecs.soton.ac.uk. Overview. Decentralised Coordination Landscape of Algorithms Optimality vs Communication Costs Local Message Passing Algorithms
E N D
Agent-Based Coordination of Sensor Networks Alex Rogers School of Electronics and Computer Science University of Southampton acr@ecs.soton.ac.uk
Overview • Decentralised Coordination • Landscape of Algorithms • Optimality vs Communication Costs • Local Message Passing Algorithms • Max-sum algorithm • Graph Colouring • Example Application • Wide Area Surveillance Scenario • Future Work & Sensor Testbed
Decentralised Coordination • Multiple conflicting goals and objectives • Discrete set of possible actions • Some locality of interaction Agents
Decentralised Coordination • Multiple conflicting goals and objectives • Discrete set of possible actions • Some locality of interaction Sensors
Decentralised Coordination • Multiple conflicting goals and objectives • Discrete set of possible actions • Some locality of interaction Agents
Decentralised Coordination No direct communication Solution scales poorly Central point of failure Who is the centre? Decentralised control and coordination through local computation and message passing. • Speed of convergence, guarantees of optimality, communication overhead, computability Central point of control Agents
Landscape of Algorithms Optimality Complete Algorithms DPOP OptAPO ADOPT Message Passing Algorithms Sum-Product Algorithm Probability Collectives Iterative Algorithms Best Response (BR) Distributed Stochastic Algorithm (DSA) Fictitious Play (FP) Communication Cost
Sum-Product Algorithm Find approximate solutions to global optimisation through local computation and message passing: A simple transformation: allows us to use the same algorithms to maximise social welfare: Factor Graph Variable nodes Function nodes
Graph Colouring Graph Colouring Problem Equivalent Factor Graph Agent function / utility variable / state
Graph Colouring Equivalent Factor Graph Utility Function
Max-Sum Calculations Variable to Function: Information aggregation Function to Variable: Marginal Maximisation Decision: Choose state that maximises sum of all messages
Wide Area SurveillanceScenario Dense deployment of sensors to detect pedestrian and vehicle activity within an urban environment. Unattended Ground Sensor
Energy Constrained Sensors Maximise event detection whilst using energy constrained sensors: • Use sense/sleep duty cycles to maximise network lifetime of maintain energy neutral operation. • Coordinate sensors with overlapping sensing fields. duty cycle time duty cycle time
Future Work • Continuous action spaces • Not limited to discrete actions • Bounded Solutions • Prune edges from the cyclic factor graph to reveal a tree • Run Max-Sum on this tree • Calculate a bound on how far this solution is from the real optimal solution Factor Graph
Publications • Farinelli, A., Rogers, A., Petcu, A. and Jennings, N. R. (2008) Decentralised Coordination of Low-Power Embedded Devices Using the Max-Sum Algorithm. In: Proceedings of the Seventh International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS-08), Estoril, Portugal. • Waldock, A., Nicholson, D. and Rogers, A. (2008) Cooperative Control using the Max-Sum Algorithm. In: Proceedings of the Second International Workshop on Agent Technology for Sensor Networks, Estoril, Portugal. • Farinelli, A., Rogers, A. and Jennings, N. (2008) Maximising Sensor Network Efficiency Through Agent-Based Coordination of Sense/Sleep Schedules. In: Proceedings of the Workshop on Energy in Wireless Sensor Networks in conjunction with DCOSS 2008, Santorini, Greece.
SunSPOT Network • Chipcon 2431 SoC • 8051 processor, 8KB RAM • SunSPOT network • Java enabled, 180 MHz 32bit ARM • Accelerometers, light, temperature sensors • Programming over-the-air