Research On Weak Target Detection Method For Vehicle Mounted Millimeter Wave Radar In Complex Environments

The technology for detecting and locating weak targets on mobile platforms in complex environments has broad application prospects in counter-terrorism,urban warfare,disaster relief,and intelligent transportation,among other fields.In practical urban environments,traditional target detection methods based on cameras face challenges such as strong lighting dependencies,low target recognition rates,limited viewing angles,and line-of-sight detection limitations,primarily due to obstacles,scene clutter,and diverse target types.Millimeter-wave radar,as a small-sized,all-weather,and high-resolution sensor,serves as a critical tool for target detection on vehicular platforms.Leveraging millimeter-wave radar for accurate localization and tracking of weak targets in complex environments has become a prominent research focus worldwide.Consequently,this thesis conducts research on weak target detection using vehicular millimeter-wave radar in complex environments,with the main objectives and innovations outlined as follows:1.Addressing the problem of non-line-of-sight(NLOS)weak target detection in typical road scenarios,this study analyzes the electromagnetic propagation characteristics of vehicular millimeter-wave radar.It formulates a NLOS forward vehicle echo model based on multipath propagation and validates this model using XFDTD electromagnetic simulation software.Furthermore,it analyzes the impact of radar-target relative positions on multipath signals,providing theoretical support for the design of localization and tracking algorithms and the practical implementation of systems.2.Tackling the challenges associated with weak target detection in complex environments,including weak signal energy,high clutter interference,and time-varying Doppler effects,this thesis proposes a series of algorithms.Firstly,it introduces a clutter suppression algorithm based on the Gaussian Laplacian operator to mitigate clutter interference in NLOS echo signals.Subsequently,a multi-target echo separation strategy based on singular value decomposition is introduced to enhance weak target signals.Finally,a density-based spatial clustering target localization algorithm is designed.3.To address the issue of low tracking accuracy of weak targets by vehicular millimeter-wave radar in complex road environments,this thesis presents two target tracking methods.Specifically,it proposes a target tracking method based on Bayesian multi-object filtering and another based on Bayesian multi-object association.These methods achieve efficient tracking of weak targets by incorporating time-varying attenuation factors and multi-object cooperative association.All of the aforementioned weak target detection methods for vehicular millimeter-wave radar in complex environments are tested and validated using either simulated or real-world data.The results demonstrate that the proposed NLOS weak target detection methods and weak target tracking algorithms effectively improve the detection accuracy of weak targets on mobile platforms and hold practical application value.