Hybrid TDOA/FDOA and track optimization of UAV swarm based on A-optimality

Abstract

The source location based on the hybrid time difference of arrival (TDOA)/frequency difference of arrival (FDOA) is a basic problem in wireless sensor networks, and the layout of sensors in the hybrid TDOA/FDOA positioning will greatly affect the accuracy of positioning. Using unmanned aerial vehicle (UAV) as base stations, by optimizing the trajectory of the UAV swarm, an optimal positioning configuration is formed to improve the accuracy of the target position and velocity estimation. In this paper, a hybrid TDOA/FDOA positioning model is first established, and the positioning accuracy of the hybrid TDOA/FDOA under different positioning configurations and different measurement errors is simulated by the geometric dilution of precision (GDOP) factor. Second, the Cramer-Rao lower bound (CRLB) matrix of hybrid TDOA/FDOA location under different moving states of the target is derived theoretically, the objective function of the track optimization is obtained, and the track of the UAV swarm is optimized in real time. The simulation results show that the track optimization effectively improves the accuracy of the target position and velocity estimation.

A fast direct locator for radiation source based on composite convolution neural network

“…While the evaluation function for testing network is the mean average error (MAE) function. The definitions of MSE, MAE and signal‐to‐noise ratio (SNR) are same as in reference [2]. For the MSE loss function, L2 regularization penalty term is added to the square of the model weights to prevent overfitting of model training.…”

“…In order to reduce the dependence of positioning accuracy on some single parameter, multiple parameters were used jointly in some works. For instance, TDOA and AOA were used jointly to locate the source by exploring both time-domain and space-domain information in [2]. For position calculation in a two-step locator, the weighted least squares method (WLS), two-stage weighted least squares (TS-WLS), Chan algorithm and Taylor algorithm are usually adopted to solve the equations based on the selected positioning parameters.…”