Handover for Multi-Beam LEO Satellite Networks: A Multi-Objective Reinforcement Learning Method
Yang Sun, Yuqing Zhai, Wenjun Wu, Pengbo Si, F. Richard Yu
Abstract
In multi-beam low-earth orbit (LEO) satellite networks, frequent handovers between intra-satellite and inter-satellite beams are inevitable. In this letter, we design a beam handover strategy based on the multi-objective reinforcement learning (MORL) method to achieve seamless and effective handover between multiple beams of LEO satellites. We first model the handover optimization problem of the multi-beam LEO satellite networks as a multi-objective optimization (MOO) problem to jointly maximize throughput, minimize the handover frequency, and keep the network load balanced. On this basis, we convert the MOO problem into a multi-objective Markov decision process (MOMDP), and utilize an MORL method, called multi-objective deep Q-learning network (MODQN), to learn and achieve the optimal solution. Simulation results show the effectiveness and superiority of the proposed handover scheme.