Adaptive Reliability-Based Splitting Algorithms with Collision Inference for Sequential Detection

We consider the problem of distributed detection in a large, single-hop, wireless sensor network in which the fusion center (FC) is not able to collect observations from all sensor nodes. We propose an ordered sequential detection algorithm that combines a reliability-based splitting algorithm, an ordered transmission strategy, and a sequential probability ratio test (SPRT). In the proposed scheme, the FC uses a random access protocol to collect local observations in descending order of their reliabilities. Unlike many schemes, where packet collisions are treated as errors and ignored, the FC in the proposed scheme can partially retrieve the observations from collisions. As a result, the FC sequentially decides whether to make a global decision or to continue collecting more local observations by using both the successfully received observations and these partially retrieved observations. Numerical results show that the proposed approach significantly outperforms a conventional SPRT scheme.