For the multi-cell case, diverse CR devices from different networks may be
involved in cooperative sensing. The cooperative devices’ sensing moments and locations are different as well as their sensing abilities. When the cooperative sensing procedure is triggered, the fusion center sends out a sensing
instruction to the cooperative CR devices. The cooperative CR devices will attach the sensing information with a specific tag indicating the
sensing moments and sensing locations, and send it back to the fusion center to make a final decision. For the multi-cell case, heterogeneous sensing
exploits both spatial and temporal diversities. The combination of sensing information in the fusion center plays an important role in cooperative
sensing. For example, the probability of channel availability varies as time goes on and location changes. Based on this observation, we introduce weight parameters to show the importance of sensing information. If the timing and
location of the obtained sensing information are close to the targeted timing or location, the associated sensing information has higher importance
and is assigned larger weight. Therefore, cooperative sensing is able to exploit the temporal and spatial diversities of channel availability. To fit the heterogeneous topology of 5G networks, cooperative sensing could operate in a centralized or decentralized manner. Decentralized cooperative sensing operates as follows.A CR device invites some of its neighboring CR devices
to organize a local cooperation team. Each team member serves as a volunteer to help the others detect the targeted channels. A cooperative device only helps the others in its idle time; hence, the overhead during sensing cooperation is low. The decentralized cooperative sensing scheme can be utilized in D2D networks or small cell networks where low-battery devices need to find available spectrum without communicating with a BS/AP.