Generalized Optimal Pilot Allocation for Channel Estimation in Multicarrier Systems

This paper addresses the design of MSE-optimal preambles for multicarrier channel estimation under a maximum likelihood or minimum mean squared error criterion. The derived optimality condition gives insight on how to allocate the pilots that compose the preamble. While many papers show that equispaced and equipowered allocation is optimal, the generalized condition demonstrates that there exist many different configurations that offer the same optimal performance. Furthermore, the condition applies not only to maximum likelihood but also to minimum mean squared error channel estimation. An application of the generalized condition in the presence of inactive subcarriers (virtual subcarriers problem) is shown such that a non equispaced allocation can achieve the same optimal performance as if an equispaced one could be used.