On the performance of random antenna arrays for direction of arrival estimation

A single antenna based virtual antenna array at the receiver can be used to find direction of different incoming radio signals impinging at the receiver. In this paper, we investigate the performance of random 3D virtual antenna arrays for DoA estimation. We have computed a Cramér-Rao Lower Bound (CRLB) for DoA estimation if the true antenna positions are not known, but these are estimated with an uncertainty. Position displacement is estimated with an extended Kalman filter (EKF) by using simulated data samples of acceleration and rotation rate which are corrupted by stochastic errors, such as, white Gaussian noise and bias drift. Furthermore, the effect of position estimation error on the DoA estimation performance is evaluated using the CRLB. The results show that the number of useful elements in the antenna array is limited, because the standard deviation of the position estimation error grows over time.