Using antenna arrays for direction of arrival (DoA) estimation and source localization is a well-researched topic. In this paper, we analyze virtual antenna arrays for DoA estimation where the antenna array geometry is acquired using data from a low-cost inertial measurement unit (IMU). Performance evaluation of an unaided inertial navigation system with respect to individual IMU sensor noise parameters is provided using a state space based extended Kalman filter. Secondly, using Monte Carlo simulations, DoA estimation performance of random 3-D antenna arrays is evaluated by computing Cramér-Rao lower bound values for a single plane wave source located in the far field of the array. Results in the paper suggest that larger antenna arrays can provide significant gain in DoA estimation accuracy, but, noise in the rate gyroscope measurements proves to be a limiting factor when making virtual antenna arrays for DoA estimation and source localization using single antenna devices.