Multi-Target Joint Range-Velocity Estimation with Graph Signal Processing in OFDM Sonar

By using orthogonal frequency division multiplex(OFDM) sonars, a new target parameter estimation method based on graph signal processing (GSP) is presented. By building full-connected graph based on OFDM signals, an objective function from the coefficients of the signal represented in an eigenvector basis is formed, basing on the graph Fourier transform. Therefore, this method realizes the joint range-velocity estimation, both for single target and multi-targets. Moreover, a filtering algorithm is also proposed to resist noise and false points. Simulation results have showed that, after filtering, 2D-GSP method has the best performance in single target estimation and gets the similar performance to the two-dimensional multiple signal classification (2D-MUSIC) algorithm in multi-targets estimation, with the computational complexity lower than that of the 2D-MUSIC algorithm.