An Efficient Differential Evalutionary Algorithm Based Localization in Wireless Sensor Networks

A Wireless Sensor Network (WSN) is a wireless network that contains several low cost tiny devices such as sensors to senses environmental circumstances. In many situations, each node of the WSN has to know its location in the real world. Several cost effective localization techniques can be used to locate each sensor. Among various techniques, the range based techniques are known for their accurate prediction of sensor location. The traditional gradient descent method like Gauss Newton updates the required parameters using derivative information. This kind of strategy works only if the derivative of the cost function exists. In addition to that, though the gradient techniques converge fast, they may be trapped by local minima. In such a case, Optimization Techniques (OTs) can be better alternatives. Among various OTs, the Differential Evolutionary Algorithm (DEA) is a popular one because it can find exact solution irrespective of initial parameters, fast convergence and consists of limited control parameters. Thus, this paper proposes an efficient DEA localization technique to outperform Gauss Newton based localization.