This paper considers a downlink multiple-input multiple-output (MIMO) multi-carrier code division multiple access (MC-CDMA) system with pilot aided channel estimation (PACE) and iterative channel estimation (ICE) in the receiver. Exploiting orthogonal space-time block coding (STBC), we investigate ICE schemes as a simple extension of PACE using estimated data chips as additional pilots. Due to the superposition of different users' spread data signals, zero-valued chips can occur after spreading, which can cause noise enhancement when using data estimates as reference signals in ICE. Hence, we propose MIMO channel estimation methods to overcome the above problem. Simulations with a realistic outdoor MIMO channel model, over a wide range of data rates and speeds, show that ICE can outperform PACE at the cost of increased complexity.
