Non-negative sparse deconvolution for localization of airbag surface deformations during inflation

In comparison with traditional TDR methods and VNA methods, the FMCW radar principle provides a cost efficient solution for monitoring rapidly time varied impedance variations on a transmission line. By attaching a part of sensing transmission line to the airbag surface, its real time state will influence the impedance profile of the sensing cable. The measured responses by FMCW radar provide information to explore the airbag surface mechanical behavior. In the previous work [1], it has been shown that the bend on the transmission line generates smoothly distributed impedance variations. It becomes hard to correctly localize the individual bend especially in presence of multiple closely spaced bends under various noise effects. In this paper, a nonnegative l1 norm deconvolution approach is proposed to deal with these problems. In an unfolding experiment the FMCW system presents fast detection ability and the proposed nonnegative sparse deconvolution technique shows accurate high resolution localization results.