Edge Computing Offloading at Middle-sea Scenario for Maritime Video Surveillances

Abstract

Video surveillance has broad application prospects in maritime rescue, ship transportation and other fields. Mobile edge computing can effectively guarantee low-latency and highly reliable data transmission in maritime video surveillance. This paper comprehensively formulates the edge computing offloading schemes in the middle-sea scenario. The middle-sea scenario has sufficient edge computing nodes as in the offshore scenario and delay constraints due to limited network connectivity as in the far-sea scenario. Taking into account these characteristics, a single-user single-hop unicast offloading model is established and extends to a multi-user model. In addition, for the sufficient and limited edge computing nodes, the multi-user model is further divided into the multi-user single-hop unicast situation 1 and 2 models. We split the mixed-integer nonlinear programming problem and approximate the optimal transmission power by the binary search method. We use the offloading decision allocation algorithm based on alternating selection, offloading decision allocation algorithm based on multi-objective alternating selection, and offloading decision allocation algorithm based on node redistribution to optimize the offloading decisions of the above models. Subsequently, we analyze the simulation results from algorithm comparisons, changes in the number of subtasks, data, and oceanic user equipments. We verify the effectiveness of the proposed schemes and algorithms in saving delay.