The Satellite Downlink Replanning Problem: A BP Neural Network and Hybrid Algorithm Approach for IoT Internet Connection

Abstract

In the era of the Internet of Things, the role of a satellite has become increasingly important. The use of satellite resources can meet the needs of high-speed Internet connection of the Internet of Things. Effective communication between satellites and the ground is crucial. This paper proposes an appropriate combination approach that is based on an improved genetic algorithm (GA) for a satellite downlink replanning problem. The initial population of the GA is optimized using a backpropagation (BP) neural network (NN). First, a variety of scheduling schemes is used to train the BP NN, and characteristics are extracted through self-adaptive learning. The NN model after training provides a good initial solution for the input information of the GA. This hybrid algorithm (HA), which consists of a scheduling GA and a local search algorithm, can quickly complete the replanning of the downlink task sequence. The ability of the HA is enhanced by the BP NN. A series of experiments is used to prove the validity of the HA. As revealed in the results of different scale instances, the proposed algorithm performs better than other scheduling algorithms.

Keywords

• Satellite downlink replanning
• hybrid algorithm
• BP neural network
• genetic algorithm
• local search
• Internet of Things (IoT)