Quality of service oriented spatial processing in the Manhattan grid

In this paper, we investigate quality of service based MIMO spatial processing in the Manhattan scenario. To model the requirements arising from the individual users, we apply specific traffic and mobility models for metropolitan areas that represent user movement and service requirements in an adequate way. These demands result in an individual signal to interference and noise requirement that has to be fulfilled for each user on the physical layer. Thus, in a first step we analyze the interference among users on a 4-way junction on the downlink and show that it should not be ignored. Based on this observation, we propose the target SINR sensitive channel representative interference cancellation (TSS-CRIC) scheme. It groups the users in a street and calculates a channel representative. The latter, as a virtual user, reflects the spatial features of all users in the group. Moreover, it is tailored in such a way that the precoding computed on its basis achieves the individual Quality of Service requirements and is able to adapt itself as the users move through the scenario. These advantages justify the higher computational complexity compared to a scheme that achieves similar results with respect to pure interference suppression.