Multi-Angle Recognition of Vehicles Based on Carrier-Free UWB Sensor and Deep Residual Shrinkage Learning

Carrier-free ultra-wideband impulse radars are increasingly used in target detection due to their high resolution and strong interference capabilities. This letter designs a deep residual shrinkage network to achieve the multiangle intelligent vehicle recognition. First, geometric models of the armored car, the truck, and the tank are established. Echo signals under the excitation of the carrier-free ultra-wideband (UWB) signal are obtained. Second, echo signals from different azimuth and pitch angles are arranged and combined into a 2-D matrix so that multiangle information of vehicles is strengthened and embodied. All matrices are converted to grayscale images and adopted as input of subsequent recognition network. Third, a deep residual shrinkage network that utilizes the soft threshold in each channel is designed to recognize vehicles under low signal-to-noise ratios. High precision and outstanding robustness of the designed network is proven by means of comparing with state-of-the-art deep neural networks (DNNs).