Contains an algorithm that calculates the sampling, forwarding and routing actions of each node within a network such that maximum information of data is collected. Two algorithms are presented, one which assumes fixed path for data from a node to the base station (i.e underlying tree structure) and one which has flexible routing.
In both cases dynamic programming is used to propagate message from the nodes in the network to the base station. In the tree algorithm messages are propagated from the root (base station) to the leaves and then back to the root again:
1. Starting from the root, node sends to all its children the energy it uses to forward data and its own energy budget (i.e. how much energy it can use). When a node receives this information from its parent it sends its own information to all of its children.
2. Each node sends its D message to its parent, expressing it utility calculated based on its own ability to send data and in conjunction with its children.
3. base station calculates the optimal coordination for the sensors based on whether they forward some of their data or their childrens data. Corrdination message propagates through the network.
Using this as a starting point for calculating flow within the network. I am going to start with a simple tree network of 4 nodes and assume there are no transmission line limits. Need to express a nodes ability to either use power from elsewhere or use its own power to satisfy a load.
No comments:
Post a Comment