Wireless Modal Analysis With Backscatter Sensors Subjected to Clock Instability

Abstract

Modal analysis using many accelerometers is a popular and essential testing of artificial structures and machines. The elimination of wire harness to power and collect data from many accelerometers attached to the object under testing is strongly demanded to reduce the time for testing and also to avoid the disturbances caused by the weights and constraints of the accelerometers and harnesses. The synchronization among acceleration measurement streams is important to derive a frequency response in terms of transfer matrix. This paper examines the practicality of concurrent acceleration data streaming with backscatter sensors to realize wireless and batteryless modal analysis. Concurrent backscatter streaming is possible by using multiple subcarriers, however the major synchronization error caused by the instability of backscatter sensor clock needs to be compensated. Synchronization of the streams by resampling is necessary but a naive implementation may produce artificial delay caused by the difference of anti-alias filters tailored to the individual sample rates of backscatter sensors. In this paper, a set of signal processing to synchronize concurrent backscatter signals is designed and implemented in a prototype system. The prototype system comprises three backscatter sensors, interrogator and modal analyzer. The signal processing in the modal analyzer features a recursive least square regression to recover sampling rate of backscatter sensor and a zero-phase-shift filtering in overlapand-add approach. The proposal is evaluated by comparing the transfer matrices generated from the prototype system and a commercial modal analysis system where wired accelerometers are used. The experimental results show that the proposed signal processing can successfully synchronize concurrent backscatter streams by eliminating the artificial delay.