Online Offloading Optimization Method for Image Encryption in Inland Vessel Remote Driving Control Tasks

Objectives Aiming at the problems of limited on-board computing power, fluctuating network bandwidth, and the difficulty of balancing real-time performance and security of traditional encryption algorithms in image encryption for inland waterway shipping remote control systems, a task offloading and encryption optimization method based on edge computing is proposed. Methods A progressive multi-round image encryption scheme based on a chaotic system was designed, which supports configurable encryption strength for multiple rounds to achieve flexible trade-off between security and computational overhead. A multi-objective optimization model incorporating delay, caching, communication cost and security score was constructed, and Lyapunov optimization theory was introduced to convert long-term constraints into queue stability problems. An improved Binary Arithmetic Optimization Algorithm (BAOA) was adopted to directly solve the offloading decision in discrete space, and dynamically allocate each encryption round to be executed on the ship-side or edge server. Experiments were conducted to compare the comprehensive performance of various strategies such as full ship-side execution, full edge execution, Genetic Algorithm (GA) and Simulated Annealing (SA) under dynamic bandwidth conditions. Results The experimental results show that under the bandwidth-fluctuating environment, the BAOA algorithm achieves the optimal balance among the four indicators of delay, communication cost, cache occupancy and security. Its total cost is lower than that of the comparison algorithms, and it maintains good robustness and stability under different network conditions. Conclusions The research results can provide a reference for the edge computing optimization technology of ship image encryption.