A sensor array, a collection of sensors spreading out over a field in some geometrical configurations, transmitting and receiving signals, is primary used to enhance the detection of signals with weak signal-to-noise ratio (SNR). With good signal processing methods, undesired signals such as noise can be suppressed and the combining output SNR is amplified by several times. One of the well-known approaches is signal combining, whose main idea is finding a set of combining weights to achieve the maximum output SNR.
Until now, there are two main approaches employed to attain the combining weights. One is eigen-based algorithm, which has heavy computational burden. SUMPLE, on the other hand, can significantly reduce computational cost while providing a comparable performance at the same time.
However, the combining loss of SUMPLE will increase if a sensor in an array fails. To circumvent the drawback, SHEN Caiyao from Information Engineering University presented a modified α coefficient SUMPLE (α-SUMPLE) method in 2011. Although α-SUMPLE can reduce combining loss for an array with a failure sensor, it may incur extra combining loss when no failure in an array.
Recently, WANG Leiou and his team from the Institute of Acoustics of the Chinese Academy of Sciences proposed a novel failure detection circuit for SUMPLE using variability index, i.e. VI-SUMPLE circuit, to detect failure sensor with lower computational cost and no extra combining loss if no sensor fails.
Based on the mathematical model and performance analysis for a failure sensor in an array, researchers find out the problem of α-SUMPLE that the unconditional correcting coefficient employed cannot effectively judge whether a sensor in the array failure or not.
Then, researchers used variability index (VI), a second order statistic, as the failure detection criterion. Combining VI with hypothesis test techniques, a failure detection threshold was determined. After that, the corresponding VI-SUMPLE circuit, which only consumes a trivial computational cost, was produced to detect failure sensor.
Simulation results indicated that this VI-SUMPLE circuit can effectively judge whether sensors fail or not in an array. Moreover, the weight amplitude of failure sensor can be compensated to reduce the combining loss, and no extra combining loss incurred when it turns out no sensor fails in an array.
WANG Leiou, WANG Donghui, HAO Chengpeng. A Novel Failure Detection Circuit for SUMPLE Using Variability Index. IEICE Transactions on Electronics (VOL. E101-C, NO. 2, February 2018). DOI: 10.1587/transele.E101.C.139.
Institute of Acoustics, Chinese Academy of Sciences, 100190, Beijing, China