Distributed estimation with energy budgets
Date of Issue2016
School of Electrical and Electronic Engineering
In this dissertation we aim to study the performance of a random tree network with changing noise profiles and when the nodes in the network are in motion with a certain velocity. In the project, we use multi-hop diffusion adaptation strategy for distributed estimation. The estimation method takes into account local and network-wide energy constraints. The diffusion strategy considered here is the multi-hop adapt-then-combine algorithm. This algorithm aims to find an optimal path for information sharing. This is achieved by determining the optimal information neighbourhood based on the combination weights. This results in lower energy budgets. The simulations were run to observe the robustness of the network. Every node in the network is associated with a standard noise deviation within an upper limit. We also introduced mobility into our network. Every node was made to follow a trajectory with a certain velocity. The nodes followed a projectile path. With every instance of time, the topology of the network changes. This allows breaking and making of links between the nodes. As the time increases for a certain velocity the communication links between the nodes start breaking and this results in increasing mean square deviation.