Abstract: To solve the vehicle network spectrum sensing and the problem of dynamic spectrum management, and improve vehicle network radio channel effective communication and data transmission performance, a predictive control of Cognitive Radio（CR） vehicle network is proposed. Firstly, it sets point calculations of constraint variables or target value expectations for predictive variables. Then, according to the expectation, in the framework of  Model Predictive Control（MPC）, the probability of predictive variables is calculated in the control system, and calculation and prediction of variable probability in real systems are adopted in the mathematical model. Finally, the prediction decision is based on practical operation in real system, and the operation results are feedback to improve the quality of forecast. The simulation results show that, compared with method of preemptive MAC and method of reactive channel allocation, control message power consumption of the proposed method is lower, the number of sensing channels for secondary user nodes is fewer, and the channel utilization is enhanced. Therefore, it can realize an efficient CR vehicle network.

Key words: spectrum sensing, vehicle network, cognitive radio, model predictive control, secondary users

摘要： 为解决车载网络中频谱感知和动态频谱管理问题，改善车载网络中无线电频谱信道中有效感知和数据传输性能，提出了一种预测控制的认知无线电（CR）车载网络。对所有预测变量，设定点计算的约束变量或目标值的期望值；根据目标期望值，在基于模型预测控制（MPC）框架中，通过控制计算预测变量的概率，使用数学模型计算预测变量在现实系统中的概率；基于现实系统预测做出决策操作，并将操作结果进行反馈，通过反馈提高预测质量。仿真实验结果表明：与抢占式MAC方法和反应式信道分配方法相比，提出方法的控制消息能耗更低，次级用户节点需要感知的信道数量更少，信道利用率得到增强。因此，可实现一个高效的CR车载网络。

关键词: 频谱感知, 车载网络, 认知无线电, 模型预测控制, 次级用户