Consensus-Based Chernoff Test in Sensor Networks
Consensus-Based Chernoff Test in Sensor Networks
复制标题
DOI：
10.1109/cdc.2018.8618910
复制
发表时间：
2018-12
期刊：
影响因子：
--
通讯作者：
A. Rangi;M. Franceschetti;S. Maranò
中科院分区：
文献类型：
--
作者：
A. Rangi;M. Franceschetti;S. Maranò
我们提出了一个顺序和自适应的假设检验，在一个完全分布式的设置，依赖于传感器网络，没有一个单一的数据融合中心。该测试的灵感来自于Bennoff的最佳解决方案，最初是在集中式设置中得出的。我们比较我们的测试与传感器网络中的最优序贯测试的性能，并提供了充分条件，使所提出的测试达到渐近最优，最小化所需的预期成本，以达到一个决定，加上预期成本作出错误的决定，当单位时间的观察成本趋于零。在这些条件下，所提出的测试也被证明是渐近最优的，相对于达到一个决定所需的时间的较高时刻。
We propose a sequential and adaptive hypothesis test that operates in a completely distributed setting, relying on a sensor network where no single data-fusion center is present. The test is inspired by Chernoff's optimal solution, originally derived in a centralized setting. We compare the performance of our test with the optimal sequential test in sensor networks and provide sufficient conditions for which the proposed test achieves asymptotic optimality, minimizing the expected cost required to reach a decision plus the expected cost of making a wrong decision, when the observation cost per unit time tends to zero. Under these conditions, the proposed test is also shown to be asymptotically optimal with respect to the higher moments of the time required to reach a decision.