An Efficient UAV-Based Data Collection Scheme for Energy-Harvesting Internet of Things

Addressing the data collection challenges in energy-harvesting Internet of Things (IoT), this paper proposes an efficient data collection scheme utilizing unmanned aerial vehicles (UAVs). In this scheme, the UAV serves as a mobile data fusion center, where sensors transmit their observations through coherent multiple access channels. Considering the energy status of each sensor, the problem is formulated as a convex optimization model. A neural network is introduced to predict the number of time slots required for collecting data from varying numbers of sensors. Additionally, a grouping strategy is employed, utilizing the mutual coherence value in the measurement matrix to determine whether grouped data collection is necessary, thereby reducing the scale of the convex problem in large-scale sensor networks. Experimental results demonstrate that the proposed scheme effectively alleviates the difficulties in solving convex optimization problems, significantly improving reconstruction accuracy and reducing algorithmic time complexity under the same number of time slots.