Spectrum Sensing Improvement in Cognitive Radio Networks for Real-Time Patients Monitoring

Abstract

Regular monitoring of vital signs guarantees a preventive treatment of common diseases ensuring better health for people. Most of the proposed solutions in e-health context are based on a set of heterogeneous wireless sensors, fitting the patient and his environment. Often, these sensors are connected to a local smart node acting as a gateway to the outside (contacts, servers). When the patient is mobile, one of the issues we may face is the guarantee of a permanent connectivity between local smart node and the outside. To overcome this problem, we need to define a robust communications architecture able to benefit from different technologies and standards. This provides equipments with the ability to dispose of free-bands to perform their transmission any-time and anywhere. Cognitive radio, although appropriate technology, requires taking into account the interdependence between the patient’s mobility and frequency band changes. Our proposal, is an anticipation model, a decision-making function that predicts the state of frequency bands occupancy. The model combines the machine learning techniques to the Grey Model system to provide low cost algorithm for spectral prediction which facilitates or guarantees permanent connectivity.