Low computational cost matching pursuit algorithm for LCX-based Intruder Detection System

The leaky coaxial cable (LCX) based intruder detection system detects intruders by making efficient use of time-variation of the channel impulse response among the cables. Recently, a compressed sensing (CS)-assisted intruder detection system has been proposed for improving the accuracy of the intruder detection. Although the CS-assisted system can improve the false detection probability performance efficiently, it is difficult to implement it because of its huge computational cost requirement. This paper proposes a modified matching pursuit (MMP) algorithm for reducing the computational cost of the CS-assisted intruder detection scheme. MMP can reduce the computational cost by limiting the search range in the vicinity of impulse response peaks estimated at the last measurement. Computer simulation results show the proposed system can reduce the computational cost for the conventional CS-based algorithm without degradation in the false detection performance.