Minimum Neighbour with Extended Kalman Filter Estimator (MINEK): Performance Evaluation

The recent emergence of high density wireless local area network (WLAN) deployments is a testament to both the insatiable demands for wireless broadband services and the ubiquity of WLAN technology. However, the increased capacity and extended coverage comes with a corresponding increase in contention and interference. These can cause a significant degradation in throughput unless an effective channel assignment scheme is employed. In an earlier work, we proposed a practical distributed channel assignment scheme for high density WLANs. The proposed minimum neighbour with extended Kalman filter estimator (MINEK) scheme maximizes throughput by selecting the channel with the minimum number of active neighbour nodes (nodes associated with interfering access points). The latter is estimated in-situ using an extended Kalman filter. In this paper, we present extensive performance evaluation of the MINEK scheme. Specifically, the scheme’s performance is evaluated in terms of upper-bound performance, normalized density, non-saturated load, unequal load, fairness and scalability. Extensive packet-level simulations using OPNET have shown that the MINEK scheme not only provides significant throughput improvement when compared to other existing schemes but is also highly robust across various deployment scenarios.