Improved gradient boosting hybrid spectrum sharing and actor critic channel allocation in 6G CR-IOT

The fast advancement of wireless communication technology and the growth in the reputation of Internet of Things (IoT) applications have led to the introduction of 6th Generation (6G) cognitive radio (CR) networks. To effectively use the inadequate radio spectrum, sharing spectrum is a crucial component of 6G CR IoT systems. This research proposes a unique method for hybrid spectrum sharing and channel allocation that combines deep reinforcement learning (DRL) with a double-sided auction process. This research comprises several components: a fusion center, base station, primary users, and SUs belongs to IoT network. Hybrid detectors called the Robust Estimator-Correlated detector and the Robust Generalized Likelihood Detector are used for spectrum detection. The fusion center makes spectrum decisions utilizing the Improved Gradient Boosting Decision Tree technique. The Twin Actor Twin Delay Deep Deterministic Policy Gradient technique is used to auction off the channel once the channel list has been decided. Simulations through network simulator assess the suggested hybrid spectrum sharing and channel distribution design. The simulation findings show how the DRL-based strategy is more successful than conventional spectrum allocation methods at increasing spectrum usage and improving interference control.