Joint Transmit and Receive Antenna Position Optimization for Movable Antenna Array Enhanced Wireless Sensing

This letter investigates the movable antenna (MA) array enhanced wireless sensing via flexible array geometry reconfiguration at both the transmitter and receiver, where the weighted sum of Cramér-Rao bounds (WSCRB) with respect to elevation and azimuth angles is adopted to evaluate the sensing performance within a given target region. To this end, an optimization problem is formulated to minimize the WSCRB by jointly optimizing transmit and receive MAs’ positions and the covariance matrix of the transmit signal. To address the problem with coupled variables and intractable objective function, a block coordinate descent (BCD) algorithm with momentum gradient descent and log barrier penalties is proposed. Since the complex objective function prevents closed-form gradient computation, we use PyTorch’s autograd function to automatically compute the gradient. Numerical results demonstrate that the proposed MA scheme achieves lower WSCRB and higher sensing accuracy compared to both sparse and dense arrays employing fixed-position antennas.