AI-Based Optimization Method for Coverage of IoT Upload Link Communication

  To address the degradation in uplink communication quality of the Internet of Things (IoT) caused by long-distance signal attenuation, multipath interference, and dynamic environmental changes, this paper proposes an artificial intelligence (AI)-based optimization method for IoT uplink communication coverage. First, service sampling points, base stations, and antenna parameters are defined within a target area, and a signal propagation model is constructed. The Sigmoid function is then used to smoothly evaluate service quality, and the coverage optimization problem is formulated as an objective function that maximizes coverage. The Q-learning algorithm is employed to achieve dynamic optimization by designing the state space (e.g., coverage rate, number of devices), action space (e.g., antenna parameter adjustments), and reward function (e.g., coverage improvement). Through iterative Q-value updates, the base station controller adaptively adjusts antenna configurations to balance coverage and interference, thereby improving long-term communication performance. To further enhance optimization efficiency, prioritized experience replay and an adaptive exploration mechanism are introduced to dynamically adjust the exploration–exploitation trade-off during learning, accelerating convergence and improving adaptability in dynamic IoT environments. Experimental results show that the proposed method significantly improves the communication performance of the uplink: the number of complete sensor data upload packets increases by 20%–25%, the number of transmission interruptions decreases by 70%–80%, and the average communication duration is reduced by 30%–40%, verifying the effectiveness and robustness of the method.