Admission Control of Video Sessions over Ad Hoc Networks Using Neural Classifiers

The paper proposes an adaptive, distributed admission control scheme for the execution of VBR video sessions over ad hoc networks with heterogeneous video and HTTP traffic. The key idea of the proposed scheme is the use of probabilistic radial basis classifiers, which are specialized types of neural networks. Firstly, the admission control function is defined analytically. Afterwards, the neural network that implements the scheme is set up. Lastly, a thorough, comparative performance evaluation through simulation is performed. Results show that the proposed scheme outperforms state of the art admission control algorithms.