An improved dynamic programming tracking-before-detection algorithm based on LSTM network value function

The detection and tracking of small weak maneuvering radar targets in complex electromagnetic environments is still a difficult problem to effectively solve. To address this problem, paper proposes dynamic programming tracking-before-detection method based on long short-term memory (LSTM) network value function(VL-DP-TBD). With the help estimated posterior probability provided by designed LSTM network, calculation function traditional DP-TBD algorithm can be more accurate, effect achieved for improved. Utilizing model estimation target motion state, moving features learned from noisy input data. By incorporating these values into algorithm, accuracy robustness enhanced, so that improved architecture capable recursively accumulating movement trend target. Simulation results show able reduce aggregation improve ability non-cooperative nonlinear dim target.AbbreviationsLSTM: Long memory; DP-TBD: Dynamic programming-based before detection; DBT: Detection tracking; TBD: Tracking HT-TBD: Tracking-before-detection Hough transform; PF-TBD: particle filtering; RFS-TBD: random finite sets; SNR: Signal-to-noise ratio; DP: programming; EVT: Extreme theory; Generalized extreme GLRT: likelihood ratio KT: Keystone transformation; PGA: Phase gradient autofocusing; CFAR: Constant false-alarm rate; J-CA-CFAR: Joint intensity-spatial CFAR; MF: Merit function; CP-DP-TBD: Candidate plot-based DP-TBD; CIT: Coherent integration time; RNN: Recurrent neural network; CS: Current statistical; Pd: probability; Pt: probability.