The deployment and configuration of wireless access networks represent an increasingly significant cost factor which will force the telecommunications industry to adopt new methodologies for network deployment and configuration. In System (UMTS) or IEEE 802.11. In combination with the need for interoperability between heterogeneous systems (e.g., different access technologies), this will inevitably increase the relative costs for deployment and configuration to an end where current quasimanual methods, using a mixture of off-line planning tools, expensive drive testing, and economic rules of thumb, become uneconomical and unsustainable. Self-configuration of access nodes helps to control these costs. However, in addition, there is a strong need for the novel concept of a self-deploying network [13]. Figure 1 shows the future vision of a selfdeploying radio access network architecture. While the deployment of base stations could be performed
IntroductionNext-generation cellular networks will have to support higher data rates to meet the demand for upcoming data services. One way to achieve this is to improve the signal processing capability of transmitters and receivers, up to a point where the channel efficiency is close to the Shannon bound. Other options are the use of multiple input-multiple output (MIMO) radio communications or beam forming to increase the achievable data rates in rich scattering or line-of-sight environments, respectively. A simpler, more pragmatic approach is a reduction of the cell size with a commensurate increase in the total number of cells. This trend is already emerging today with the increasing number of picocells for hot-spot coverage using Universal Mobile Telecommunications manually, based on the optimal positions identified by the network, a different approach is investigated here: mobile wireless access nodes that are able to plan and modify their positions without human interaction are considered. This allows the network to adapt autonomously to changes in user positions and demand. In the future, this concept may even be further extended to fully "robotic" base stations with a high degree of autonomy, intelligence, and powers of reasoning [8]. To enable base station mobility, a wireless backhaul is considered that supports both direct and peer-to-peer connections.While base station mobility might seem futuristic for commercial wireless communication systems today, this concept has near-term applications in the field of military and emergency communications, where fast network deployment is required in highrisk areas or in environments that are difficult to access. For example, using unmanned area vehicles (UAVs) as base station carriers, a self-deploying network could rapidly provide coverage over large areas.Base station positioning has been studied extensively in the past, using simulated annealing [11,19], evolutionary algorithms [21], integer linear programming [12], and greedy algorithms [20]. Other work has explored the trade-offs among coverage, cell count, and capac...