Lane-change-aware connected automated vehicle trajectory optimization at a signalized intersection with multi-lane roads

Trajectory smoothing is an effective concept to control connected automated vehicles (CAVs) in mixed traffic reduce oscillations and improve overall performance. However, smoother trajectories often lead greater gaps between vehicles, which may incentivize human driven (HVs) from adjacent lanes make cut-in lane changes. Such changes compromise the expected performance CAV trajectory smoothing. To figure out reasons behind issue, this paper designs a framework at signalized intersection with multi-lane roads considering detailed control, car following changing maneuvers all together. Based on framework, proposes decentralized lane-change-aware optimization model including discretionary change restraining mandatory yielding strategies. Riding comfort mobility are considered as joint objective. And complex non-linear constraints linearized convert proposed problem quadratic problem. The linearization allows investigated be easily fed into commercial solver. Numerical experiments conducted study of compare it other models (e.g., cooperative without mechanism) different scenarios. First, results show that HV cause reduction half or more benefits along segment intersection. Then, we find outperforms models. Especially, yields extra system objective (10–25%), riding travel time (1–8%), fuel consumption (3–15%) safety (5–25%) compared mechanism when market penetration rate not high. Sensitivity analyses road lengths, signal cycle saturation rates through-vehicle better under most scenarios, e.g., 20% benefit short length, 30% long 25% high rate, rate.