Interaction-aware eco-driving control under complex traffic environment: A comprehensive review

The rapid development of eco-driving technologies is conducive to improving the energy economy of connected and automated vehicles (CAV). However, most eco-driving control methods have ineffective practical application in complex traffic environments because existing methods cannot sufficiently consider the complex interaction behaviors and factors. To reveal the limitations and improvement direction of current researches, this paper delivers a comprehensive review of interaction-aware eco-driving control under a complex traffic environment. The features and modeling process of the interaction-aware eco-driving control problem are first elaborated in detail. Influencing factors of eco-driving control in the interaction scenario are then innovatively summarized from the people-vehicle-road interactive environment. Furthermore, interactive scenario-oriented eco-driving studies in complex traffic environments are generalized and divided into five typical types: car-following interaction scenarios, lane-changing interaction scenarios, intersection interaction scenarios, ramp merging interaction scenarios, and multi-vehicle platoon interaction scenarios. Finally, according to the above-stated overview, technical challenges and future research trends for interaction-aware eco-driving are suggested from five aspects: scenario cognition, personalization, decision-making, algorithm self-evolution, and control robustness.