Summary
Cognitive radio (CR) was developed to solve the issue of spectrum scarcity because CR users are unlicensed and can use the underutilized licensed spectrum without interfering with licensed users. During handoff, however, this CR creates a target channel sequence (TCS) to transfer the channel to the multiusers. It must also deal with issues such as equally assigning TCS to different secondary users (SUs), channel access collisions, and channel obsolesces during TCS development. TCS is issued to SUs in the same and different primary user (PU) areas. In this paper, the hybrid multilayer perceptron (MLP)–convolutional neural network (CNN) (MLP‐CNN) technique is specifically proposed to provide services to the SU even under proactive TCS constraints, as well as to address the aforementioned challenges. The MLP can increase cell quality by resolving the transmission time gap problem that occurs during the proactive stage. The MLP determines the characteristics of each SU in a specific area with high reliability for channel allocation, while the CNN allocates the channel based on the characteristics of the entire SU present in a specific area. This hybrid MLP‐CNN approach, on the other hand, is unable to achieve the optimal degree of optimized TCS generation accuracy due to the need for parameter fine‐tuning, which is accomplished using the golden eagle optimization technique. Different parameters, such as single‐user scenario, multiuser scenario, multichannel scenario, and network access overhead, are used to measure the performance of the proposed model. The handoff delay, average interval per target channel, and the overhead that occurs during the transmission of data by our proposed method are 0.56%, 0.17%–0.03%, and 50%–10% respectively. Thus, our proposed method outperformance all the existing approaches.