White space prediction technique for cognitive radio applications

This paper investigates spectrum occupancy trends of voice services in the band of 450 MHz. An occupancy model of channel occupancy in the frequency and time domain was developed. A robust algorithm is proposed to allocate white spaces within a channel group with by primary users. The algorithm statistically identifies which channels can be used with a higher probability of having a white space, so minimizing the possible collisions. The simulation results indicate clearly the good potential for cognitive radio secondary use in this band.