Unscented transform based low complexity performance assessment for adaptive linearly constrained minimum variance filters

Linearly Constrained Minimum Variance (LCMV) filters are applied in communication and RADAR systems. In order to evaluate the performance of these filters, Monte Carlo (MC) simulations are commonly employed despite their high computational complexity. This paper proposes a low complexity performance assessment based on the Unscented Transform (UT). With only 32 iterations, the performance evaluation curves of the UT based approach superpose the curves of a thousand MC iterations. Since the computational complexity of one UT iteration is approximately the same as that of a MC iteration, the proposed solution drastically reduces the required time for performance evaluations.