Two expected characteristics of future wireless networks aresupport for multiple bearer services, which in turn enablemultiple end-user services, and the parallel use of multipleradio access technologies. This dissertation discusses radioresource management principles to improve capacity for suchmulti-bearer service, multi-access networks.
More specifically, it is first focused on how to mostefficiently share the radio resource between bearer servicegroups within one access technology. A general principle forsharing resources in interference limited systems is proposed,and its expected performance estimated. The proposedinterference balancing principle maximizes capacity byadjusting the power budgets per bearer service group so thatthe maximum tolerable interference levels are equal for allbearer services. To verify its validity, the interferencebalancing principle is applied to the 3rd generation cellularsystems GSM/EDGE and WCDMA in a set of multi-bearer servicecase studies. It is seen that interference balancing maystraightforwardly be introduced in these systems, and thatsignificant capacity gains over non-balanced scenarios can beachieved.
Secondly, how to best share traffic load between thedifferent sub-systems in a multi-access scenario isinvestigated. The capability to handle bearer services, andthereby also end-user services, typically differs betweensub-systems. The overall multi-access system capacity istherefore affected by the allocation of bearer services on tosub-systems. Based on this, a simple principle for findingfavorable, under certain constraints near-optimum, sub-systembearer service allocations is derived. It is seen that for agiven service mix combined capacities beyond the sum of thesub-system capacities may be achieved by using the favorablebearer service allocations. Significant capacity gains are alsoseen in a case study in which the bearer service allocationprinciple is applied to a combined GSM/EDGE and WCDMAmulti-access system. The bearer service type may be said toreflect an expected radio resource cost for supporting a userin each sub-system. By taking into account the actual radioresource cost, which may differ from the expected cost, whenassigning users to subsystems, further capacity gains areachievable.
Stockholm: Signaler, sensorer och system , 2003. , ix, 157 p.