Design of Collision-free Nearest Neighbor Assertion and Load Balancing in Sensor Network System

Abstract

Collision free transmission of raw sensor signal in application like detection and estimation of environmental problems is difficult and one of the major challenges is to balance the load and identify the interference nodes to achieve high performance. Ensuring multipath transmission among the nodes guarantees end-to-end packet transfer. However, such multipath transmission also gives rise to multiple copies of same packet resulting in interference and collisions among the nodes. While, certain routing protocols have high probability of delivery by maintaining an accurate neighbor list, they fail to achieve load balance which extensively degrades the performance of the network. In this paper, we present a Collision-Free Nearest Neighbor Assertion (CNNA) method to avoid collision and interference by significantly identifying the multiple copies sending nodes CNNA method initially identifies the collision and interference nodes in the sensor networks using the n-d tree data structure. The n-d tree search uses the binary search tree mechanism to easily identify the collision and interference nodes with minimal search time. A genetic optimization approach is formulated with an objective to balance the load on the neighboring nodes of the sensor network by measuring the weighted variance for the collision free n-d tree data structure. With the weighted variance using n-d tree data structure, Crossover and Mutation operation is applied to balance the load on the nodes. Extensive simulation results are carried out through analysis to show that the proposed algorithm satisfies both load balancing and collision removal in sensor network system. CNNA method achieves highest performance rate and performance measure is done on the factors such as collision removal rate, load balancing efficiency and search time taken for identifying the collision in network.