Optimization Design Framework for In-Vehicle Time-Sensitive Networking Architecture

The in-vehicle network (IVN) architecture significantly impacts the performance of the high-level autonomous vehicles. This article proposed an innovative optimization design framework for the in-vehicle time-sensitive networking (TSN) architecture. This pioneering framework is specifically designed to optimize switch port assignment, load distribution, and end-to-end delay. To balance port allocation and load distribution, a multiobjective optimization problem is formulated. An adaptive nondominated sorting genetic algorithm (NSGA-II), which incorporates the adaptive crossover and mutation probabilities, is employed to identify the candidate topologies. The end-to-end delay is evaluated by an improved gray wolf optimizer (IGWO)-based TSN scheduling algorithm. By integrating genetic and tabu search operators, the efficiency and scheduling effectiveness of the IGWO are significantly enhanced. The simulation verifies the superiority of the adaptive NSGA-II and IGWO on search capability. A design instance for a high-level autonomous vehicle is completed based on the proposed framework. The results demonstrate the effectiveness of the design framework and some design ideas are summarized.