Study And Simulation On Channel Gain Estimation Algorithms In Wireless Sensor Networks

In wireless sensor networks, for some special measurement requirements, the sensors are often deployed in inaccessible terrains that forbid replacement or replenishment of the sensor node power sources. Therefore, energy consumption is a key issue in wireless sensor networks.A large portion of energy is consumed on communications and can be minimized. In order to improve energy utilization efficiency of the network, distributed power control should be performed if the channel gains are known. However, the sensors are deployed randomly which leads to unknown channel gains. Thus, how to estimate channel gains accurately becomes the research subject of this paper.For traditional channel gain estimation algorithm, such as the least mean square (LMS) algorithm, each node updates estimation independently by repeating iteration to gradually reduce the estimation error. A cooperative distributed estimation algorithm:estimation algorithm based on diffusion LMS (D-LMS), has been proposed by making use of the equality and collaboration of sensor nodes. The D-LMS estimation algorithm adopts the collaborative means to enhance the accuracy and convergence of the estimation algorithm. Then, based on the D-LMS estimation algorithm, an estimation algorithm based on adaptive diffusion LMS (AD-LMS) is proposed. The AD-LMS estimation algorithm minimizes the local estimation error in each iteration as far as possible, in order to minimize the final estimation error.In order to optimize performance, considering the global estimation algorithm has the best estimation performance, we take it for reference and minimize the global consumption for purpose as far as possible. We improve the algorithm to make it meet the requirements of distributed estimation. As a result, two distributed estimation algorithms are proposed. They are estimation algorithms based on exchange diffusion combine LMS (EDC-LMS) and exchange combine diffusion LMS (ECD-LMS). Considering the influence of adaptive adjustment for improving the performance of the algorithm, we add adaptive adjustment in the new algorithms.In the end of this paper, we perform simulations of algorithms in the estimation error and convergence by Matlab. Simulation results demonstrate that the algorithms proposed in this paper have expectant performance.