Dual-layer formation control strategy for connected and mixed traffic flow utilizing fish swarm algorithm

Current research indicates that collaborative decision-making and control across several vehicles can markedly enhance traffic efficiency and driving safety. This study examines a vehicle formation control strategy in a multi-lane context, with the objective of improving vehicle efficiency near highway exit ramps and ensuring safety throughout the formation process. The study examines a mixed traffic flow comprising connected and autonomous vehicles (CAVs) and connected and human-driven vehicles (CHVs) and develops a framework for cooperative control strategies. The framework comprises two tiers: the upper tier enhances the formation scheme and suggests a vehicle-following relationship allocation strategy to optimize both the formation scheme and vehicle distribution, facilitating CAVs in directing CHVs to establish a mixed queue at minimal expense; the lower tier employs the Fish Swarm Algorithm to devise a dynamic path planning algorithm for vehicles, allowing for dynamic obstacle avoidance decision-making for CAVs and CHVs throughout the queue formation process. The efficacy of the suggested strategy was validated by SUMO simulations in a three-lane road segment scenario preceding a highway exit ramp. The numerical findings indicate that the suggested formation control strategy can achieve a maximum 55% reduction in Time Exposed Time-To-Collision (TET) and a 26% improvement in rear-end accidents, and enhance the real traffic volume of the road segment across all situations of CAV penetration rates, which mitigates congestion, optimizes traffic efficiency, and improves driving safety across diverse traffic flows.