Team Members:
Study fault-tolerance RF communication in multi-hop routing environments, develop robust dynamic re-routing software architecture by combination of LEGO / RCX kits and Motes
Sensor networks are very common in areas like environment monitoring, system monitoring, etc. Lot of research work has been done in proposing routing algorithms in multi-hop sensor networks. The main disadvantage of multi-hop sensor networks is that the overall power consumption increases with the sparse network due to large distance transmission. Also there can be several retransmission due to increase in transmission errors. Data mule is an approach to reduce these issues in case of sparse sensor networks.
Data mule is a mobile agent which travels around the network along a fixed path and communicates with each sensor nodes, gathers data from these nodes and delivers to the gateway system (which can be a computer). In this case, the sensor nodes need not do long range transmission and save power considerably. The data mule approach itself has its own drawbacks such as increased latency in gathering data at the central gateway and estimating the optimal path to move around in the network.
We aim to combine the advantages of the two approches. Instead of having an entirely multihop network or entire sparse / unconnected group of motes which convey data through data mule, we propose a clustered sparse network, comprising of islands of sensor networks. A cluster / island of sensors in themselves is a multi-hop sensor network. We aim to propose a fault-tolerant multi-hop routing algorithm that can be used within each cluster. At this level, the data mule moves around in the network and communicates only with the group-leader of each cluster. The data mule itself becomes a solution for fault tolerance when the fault-tolerant algorithm reaches its limits because of physical constraints.
References:
Created on: Mar 16, 2009, Updated on April 22, 2009