Cognitive radio technology deals with the problem of spectrum scarcity. Such technology allows the secondary users (SUs) to access the primary users’s (PUs) channels without causing harmful interference to the PU’s transmission. In this paper, a novel multi-class SU cognitive radio model is introduced where each SU class has the ability to sense and operate on a subset of the network channels. Such multi-class SUs’ network could exist due to hardware limitations or network regulations. The model is analysed based on the mixed preemption/nonpreemption resume priority M/M/1 queueing model. First, the analysis is carried out assuming no priority between the SU classes. Next, a priority scheme is introduced to balance the performance between the different classes. The numerical results show the delay performance trade-off between the different classes and the overall delay enhancement to the network after prioritizing the classes. Moreover, comparative studies to related work in literature and simulation results for the effect of sensing errors on SU and PU networks are introduced. Based on these results, it is recommended to apply the priority-based model for low and medium PU’s loads when the system time is considered as a performance metric. As far as we know, the case of different channel access capabilities between the SUs has not been analysed in literature before.

认知无线电技术处理频谱稀缺的问题。这种技术允许次要用户（SU）访问主要用户（PU）通道，而不会对PU的传输造成有害干扰。在本文中，介绍了一种新颖的多类SU认知无线电模型，其中每个SU类都具有感知和操作网络信道子集的能力。由于硬件限制或网络法规的限制，此类多类SU的网络可能存在。基于混合抢占/非抢占恢复优先级M / M / 1排队模型对模型进行了分析。首先，假设SU类之间没有优先级进行分析。接下来，引入优先级方案以平衡不同类之间的性能。数值结果表明，在优先考虑类别之后，不同类别之间的延迟性能之间存在权衡关系，并且网络的整体延迟增强。此外，还介绍了对相关工作的比较研究以及仿真结果对传感误差对SU和PU网络的影响。基于这些结果，建议在将系统时间视为性能指标时，将基于优先级的模型应用于中低PU负载。据我们所知，SU之间不同信道访问能力的情况以前没有在文献中进行过分析。基于这些结果，建议在将系统时间视为性能指标时，将基于优先级的模型应用于中低PU负载。据我们所知，SU之间不同信道访问能力的情况以前没有在文献中进行过分析。基于这些结果，建议在将系统时间视为性能指标时，将基于优先级的模型应用于中低PU负载。据我们所知，SU之间不同信道访问能力的情况以前没有在文献中进行过分析。