An Analytical Technique to Determine the Decision Thresholds of Multi-bit Distributed Detection in Sensor Networks

Abstract

In this paper we propose a technique to determine the decision thresholds in multi-bit distributed detection. Detection thresholds are required to quantize the acquired information from the environment to send them to a fusion center. In multi-bit detection, decision making is complicated and in most cases, methods based on simulation or Person by Person Optimizations are applied to find the thresholds. In such methods, the thresholds are usually determined by estimating environmental conditions and conducting the necessary simulations and carrying out complex computations. In our proposed method we have adopted a simple analytical technique in place of these complicated methods with a detection performance close to the optimal. Compared with the detection performance of optimized solutions, the difference is slight while the simplicity and the speed of computations in our method allows the possibility of using it in a sensor network after it has deployed in the environment and the actual conditions have been determined.