Error analysis for optimal distributed detection in multi-hop sensor networks

We consider a sensor network in which the sensor motes first make local decisions based on their observations of a noisy environment. They then transmit these local decisions over multiple noisy wireless hops to a cluster head (CH). The CH collects them and fuses them into a globally optimal decision. In this paper, multiple techniques are used to characterize the decision error probability when the CH uses a likelihood ratio test. The results obtained include easily evaluated expressions that provide upper and lower bounds on the exact performance of the optimal distributed detector. These results will enable rapid analysis of system-level tradeoffs involving the number of sensors, the level of error, and energy usage.