A Research On Location Of Deep Underground Pipeline With MIMO Ground Penetrating Radar

Ground penetrating radar(GPR)is a non-destructive geophysical detection method,which is widely used in the detection of underground pipelines.However,due to the low pass filtering effect of the earth medium,the traditional pulse ground penetrating radar cannot take into account the detection depth and resolution,leading to the difficulty in the detection of deep underground pipelines.Multiple input multiple output(MIMO)radar has the characteristics of virtual large aperture and high resolution.In this paper,the application of orthogonal MIMO ground penetrating radar in the field of pipeline location related technologies are studied.Waveform orthogonality is the prerequisite for MIMO ground-penetrating radar to achieve high resolution positioning.High correlation sidelobe of existing orthogonal phase coded signals will affect the accuracy of pipeline detection.Aiming at the above problems,a quadrature phase encoding waveform design method based on the improved Harris hawks optimization(HHO)is proposed.MIMO ground penetrating radar has a limited length of the transmitting waveform because underground deep-buried pipelines belong to the near field relative to the transceiver antenna.It is necessary to design a suitable high-orthogonal phase encoding waveform according to the actual detection conditions.The crossover and mutation strategies of genetic algorithm(GA)are introduced in the framework of HHO algorithm to update the discrete individual,which makes the algorithm suitable for searching the discrete polyphase coding matrix.The simulation results show that the waveform designed based on the improved HHO algorithm has better orthogonal performance than the waveforms designed based on similar algorithm.,which is conducive to improving the positioning accuracy of subsequent underground pipelines.In underground environment with low SNR,it is difficult for Multiple Signal Classification(MUSIC)algorithm to obtain an accurate noise subspace through eigenvalue decomposition,and the inaccurate noise subspace will lead to performance degradation of MUSIC algorithm.Aiming at the above problems,a subspace iterative MUSIC algorithm is proposed.Firstly,the ground-sticking detection model is modeled in polar coordinates,and the signal model coupled with angle and distance is obtained;the ground-leaking detection model is modeled in Cartesian coordinates,and the signal model coupled with horizontal distance and depth is obtained.Secondly,in order to obtain the accurate signal subspace,a set of non-zero delay covariance matrices are used to establish the objective function related to the signal subspace.Then,the accurate signal subspace is estimated by iterative solution.Finally,the projection matrix of noise subspace is obtained through the relation between signal and noise subspace,hence an improved MUSIC algorithm is implemented.The simulation results show that the improved MUSIC algorithm has better performance than similar algorithms,which is conducive to improving the resolution of the imaging location of underground pipelines.The search of non-target areas in MUSIC algorithm leads to low detection efficiency,which is not suitable for occasions with high real-time requirements such as vehicle-mounted detection.Aiming at the above problems,the sparrow search algorithm(SSA)with powerful search ability is used to improve the two-dimensional search process.However,the classic SSA algorithm cannot solve the multi-modal problem.An SSA-root-MUSIC algorithm is proposed in the ground-sticking scenario.After eliminating the distance parameter,the root-MUSIC is used to estimate the target angle,and then the SSA algorithm is used for precise positioning;In the ground-leaking scenario where the received signal parameters are inseparable,an improved multimodal SSA-MUSIC algorithm is proposed.The hybrid clustering algorithm is used to classify molecular population in the improved algorithm,where multiple extreme points are located through an intelligent search.The simulation results show that both algorithms improve the positioning efficiency while avoiding the quantization error of grid traversal,have higher computing speed and accuracy,and are suitable for the continuous detection scene of vehicle-mounted ground penetrating radar.