A neural network-based approach for predicting connectivity in wireless networks

Prediction of wireless network connectivity using a Taylor Kriging approach

“…The applicability of the CDSS framework includes cellular networks, ad hoc networks, and wireless IEEE 802.11-based networks. By using a neural network (NN) approach, Nasereddin et al (2005) proposed a CDSS based on generated connectivity maps. The paper used the signal strength data from active wireless users to train an NN and then predict the signal strengths or coverage for the locations where no active user is reporting.…”

“…This approach is compared with a radial basis function artificial NN using several problems. Following the work of Nasereddin et al (2005), Bartolacci et al (2004) and Konak (2009), this paper uses a Kriging model based on Taylor expansion to predict the quality of connectivity for wireless networks of 14-towers, 27-towers, and 45-towers. The prediction results from the new Kriging model are compared with those of the references.…”

“…The data of wireless networks with 14, 27 and 45-towers from (Nasereddin et al, 2005;Bartolacci et al, 2004;Konak, 2009) are used as the test problems. The inputs are the power of the tower serving that point (that is, the closest tower) and the Euclidean distance to the tower.…”

“…Wireless network research deals with all types of connections without cables. Network connectivity is defined as the ability to connect to a network connection point with some minimum threshold of signal strength and quality that allows ongoing communication (Nasereddin et al, 2005).…”

Predicting coverage in wireless local area networks with obstacles using kriging and neural networks

“…There has been limited work in the literature to estimate the network coverage in wireless networks using empirical approaches. Nasereddin et al (2005) has developed a radial basis function artificial neural network (ANN) to estimate the signal-to-noise ratio, which is an important indicator for quality of service in cellular wireless networks. To predict the signal-to-noise ratio at a point p, this ANN approach utilizes four inputs: the x-y coordinates (indices) of point p, the index of the transmitter with highest transmitted power at point p, and the transmission power.…”