Off-grid radar node placement for target localization in radar networks

The selection of the optimal nodes positions in multistatic radars is an important design task. The reason for this is a strong correlation between system performance and radar nodes topology, which in turn is limited mainly by economical constraints and operational envirinment. Conventional topology optimization approaches are based on the gridding of areas of potential radars positions. This affects quality of the solution, allowing the radar nodes to be placed only on the available positions from the fixed set. In order to mitigate the effect of gridding, we use the first-order Taylor series approximation of the cost function around the potential coordinates of radar nodes locations. This results in the representation of the radar area as a set of bins allowing one radar position per bin to be selected. Provided numerical analysis shows that this approach lead to radar network topology selection that results in lower errors of target position estimation, compared to approach based on gridding of the area of potential radar nodes positions.