Cognitive radio networks (CRNs) provide an effective solution to address the increasing demand for spectrum resources. The cooperation among secondary users (SUs) improves the sensing performance and spectrum efficiency. In this paper, we study the traffic-demand-based cooperation strategy of SUs in multichannel CRNs, in which each SU makes its own cooperative sensing decision according to its traffic demand. When an SU has high traffic demand, it can choose to sense multiple channels in the sensing period and obtain more chances to use spectrum resources. If an SU has no data to transmit, it can choose not to perform spectrum sensing. We formulate this problem as a nontransferable utility (NTU) overlapping coalitional game. In this game, each SU implements a cooperation strategy according to its expected payoff, which takes into account the expected throughput and energy efficiency. We propose two different algorithms, namely, an overlapping coalition formation (OCF) algorithm and a sequential coalition formation (SCF) algorithm, to obtain a coalition structure. The OCF algorithm guarantees that the coalition structure is stable, whereas the SCF algorithm has lower computational complexity and less information exchange among SUs. Moreover, when being assigned a certain channel, SU implements an adaptive transmission power control scheme to further improve its energy efficiency. Simulation results show that our proposed algorithms achieve a higher aggregate throughput than the disjoint coalition formation algorithm from the literature, where each SU can join only one coalition.