WP5. Network Optimal Transmit and Receive

In this WP a global QoS optimisation problem is formulated taking into account the whole possible heterogeneous cellular network. The inter-cell interference has a well-defined structure in the space domain. Provided the receiver has multiple antennas, we will exploit the spatial correlation properties of the interference to design receivers that try to reach the best compromise between inter-cell cancellation and diversity gain. In practice, a multiple antenna receiver has several degrees of freedom, in the joint space and frequency domain. If we increase the number of degrees dedicated to interference cancellation, we reduce the maximum achievable diversity gain. Therefore, there must be an optimal trade-off, in the receiver design, which will be the primary subject of our investigations.

Besides, implementation of a unified air interface concept is one of the potential techniques that can significantly increase the uplink and downlink capacity of cellular networks. Nevertheless, in order to achieve a network capacity gain from link level improvements the radio resource management algorithms must be able to automatically capture the link gains. It is expected that the network gain will depend on the type of advanced receiver, the multi-path channel profile, the block error rate target (BLER), etc., aspects that will be investigated in this WP.

In packet oriented wireless systems with QoS provisioning the received interference will have a large variance both in frequency and time - a much more difficult environment than for the single cell case. We will study the potential of using the SURFACE air interface in different packet oriented systems with varying degrees of granularity, i.e., where the radio resource units to be scheduled have different characteristics. We will design a packet scheduler that considers the state of the propagation channel as well as the interference. In this WP we will not consider limitations regarding, signalling, etc., and therefore the performance results will be upper bounds of the results obtained by more realistic algorithms developed in WP6.