21 May 2007 in IEEE Transactions on Signal Processing
We design Tomlinson-Harashima precoding for decentralized receivers and frequency-selective channels based on the minimum mean square error criterion, where the feedforward filter is restricted to have finite length. Contrary to most other publications on Tomlinson-Harashima precoding which rely on solutions for decision feedback equalization to find the corresponding precoding filters in a heuristic manner, we deduce the optimization for Tomlinson-Harashima precoding from the optimization for the linear minimum mean square error transmit filter. Thereby, we include the precoding order explicitly in the problem formulation and thus obtain the precoding filter solutions, together with the algorithms to compute the latency time, i.e., the time difference between application of the precoder at the transmitter and detection at the receiver and the precoding order from a single optimization. Since the algorithm for THP filter computation resulting from the optimization has a high computational complexity, we present an alternative algorithm to compute the Tomlinson-Harashima precoding filters based on a Cholesky factorization with symmetric permutation, resulting in an order of complexity that is the same as for the computation of the linear transmit filters. The simulations reveal that the latency time optimization can be omitted without performance degradation for most practical channel models, i.e., the latency time can be chosen to be the order of the feedforward filter
18 June 2008 in EURASIP Journal on Advances in Signal Processing
The problem of determining an optimal parameter setup at the physical layer in a multiuser, multiantenna downlink is considered. An aggregate utility, which is assumed to depend on the users' rates, is used as performance metric. It is not assumed that the utility function is concave, allowing for more realistic utility models of applications with limited scalability. Due to the structure of the underlying capacity region, a two step approach is necessary. First, an optimal rate vector is determined. Second, the optimal parameter setup is derived from the optimal rate vector. Two methods for computing an optimal rate vector are proposed. First, based on the differential manifold structure offered by the boundary of the MIMO BC capacity region, a gradient projection method on the boundary is developed. Being a local algorithm, the method converges to a rate vector which is not guaranteed to be a globally optimal solution. Second, the monotonic structure of the rate space problem is exploited to compute a globally optimal rate vector with an outer approximation algorithm. While the second method yields the global optimum, the first method is shown to provide an attractive tradeoff between utility performance and computational complexity.