Development of a Cochlea-inspired Artificial Filter Bank (CAFB) and evaluation of its performance via real-time dynamic response tests on a two-span bridge

Abstract

In this paper, a Cochlea-inspired Artificial Filter Bank (CAFB) was developed in order to efficiently acquire dynamic response of structure, and it was also evaluated via dynamic response experiments. To sort out signals containing significant modal information from all the dynamic responses of structure, it was made to adopt a Band-pass filter Optimizing Algorithm (BOA) and a Peak-Picking Algorithm (PPA). Optimally designed on the basis of Kobe earthquake signals, it was then embedded into the Wireless Variety Measurement System (WiVMS). In order to evaluate the performance of the developed CAFB, a vibration test was conducted using the Kobe earthquake signals, and structural responses of a two-span bridge were obtained and analyzed simultaneously by both a wired measurement system and a CAFB-based WiVMS. The test results showed that the compressed dynamic responses acquired by the CAFB-based WiVMS matched with those of the wired system, and they included significant modal information of the two-span bridge. Therefore this study showed that the developed CAFB could be used as a new, economic, and efficient measurement device for Wireless Sensor Networks (WSNs) based real time structural health monitoring because it could reconstruct the whole dynamic response using the compressed dynamic response with significant modal information