Physical layer metrics estimation for CSMA/CA networks using a Markov modeling and source enumeration

• The aim is to estimates some vertical handover metrics.
• These metrics help to take decision on the suitable access point to choose.
• All the metrics are estimated from the physical layer signal.
• To achieve the estimation we propose a new number of source tracking algorithm based on a Markov chain modeling.
• The proposed algorithm is a cognitive radio algorithm dedicated to CSMA/CA based networks.

The existence of multiple wireless networks with different radio access technologies and protocols makes the radio environment heterogeneous. In order to provide the best Quality of Service available from the active networks, and satisfy the concept of always best connected, one can take advantage of this heterogeneity by developing multi-mode terminals able to smartly switch from one interface to another. This switching process, known as vertical handover (VHO), requires some relevant metrics to be measured by the terminal in order to decide whether to trigger a VHO or not. Using multiple antennas, we propose to track the number of active sources and employ the results in CSMA/CA networks for VHO. The proposed algorithm is developed using a Markov chain model for sources enumeration at any given time. We also use a three state Markov model for CSMA/CA networks and show how this algorithm can be applied to recursively obtain two informative metrics about the channel state, namely the channel occupancy rate and the collision rate. Numerical simulations confirm that the proposed algorithm performs well for practical SNR values. The proposed algorithm relies on a physical layer sensing and requires no connection to the access point, no synchronization, no signal demodulation and no frame decoding. This particularity ensures a seamless handover with a time/energy economy.