Control of Energy Efficient Powertrain for Autonomous and Connected Vehicles in a Mixed Autonomous and Human Driving Environment

在不久的将来，人类驾驶员将与自动驾驶汽车共享道路。 如果车辆控制系统集成了人类驾驶员行为的模型，自动驾驶车辆可能会实现显著的节能。这种节省将有益地影响运输经济，从而增加国家繁荣。该项目考虑在自动驾驶和人类驾驶车辆存在的道路环境中控制自动驾驶和联网车辆。该项目将把人类驾驶员行为的新模型集成到自动驾驶汽车运动和动力系统控制中，从而实现能源效率和驾驶员安全的同时优化。实验还研究了人类驾驶员和自动驾驶汽车在封闭试验场上的自适应交互，其中可以安全地测试和验证系统性能。 该项目将包括一个教育部分，为研究生、本科生和高中生提供开展研究的培训。该项目还开发动画软件，帮助学生了解联网和自动驾驶汽车的工作原理。这项研究调查了新的控制方法，通过整合人类驾驶行为的新模型，能够预测人类驾驶车辆和自动驾驶车辆在很长一段时间内的运动，从而提高自动驾驶车辆的动力系统效率。方法包括：基于危险的建模框架和滚动水平机制，确保在考虑驾驶员和车辆异质性的同时对车辆运动进行准确建模;人类驾驶行为如何影响自动驾驶车辆控制的能源效率的实验评估，以及自动驾驶汽车控制策略的变化将如何影响人类驾驶。该奖项反映了NSF的法定使命，并通过使用基金会的学术价值和更广泛的影响审查标准。

In the near future, human drivers will share the roadways with fleets of autonomous vehicles. Significant energy savings might be achieved in autonomous vehicles if vehicle control systems integrate models of human driver behavior. Such savings would beneficially impact the economics of transportation, thereby increasing national prosperity. This project considers control of autonomous and connected vehicles in an on-road environment where both autonomous and human driving vehicles exist. The project will integrate novel models of human driver behavior into autonomous vehicle motion and powertrain control, thereby enabling simultaneous optimization of energy efficiency and driver safety. Experiments also examine the co-adaptive interaction between human drivers and autonomous vehicles on a closed-course proving ground, wherein system performance can be tested and verified safely. The project will involve an educational component that provides training to graduate, undergraduate, and high school students in conducting research. The project also develops animation software to help students understand the working principle of connected and autonomous vehicles. This research investigates new control methodologies that promise improved autonomous vehicle powertrain efficiency by integrating a novel model of human driving behavior, which is capable to predict the movement of human-driven vehicles and autonomous vehicles over long time periods. Methods include: a hazard-based modeling framework and rolling horizon mechanism that ensures accurate modeling of vehicle movements while considering driver and vehicle heterogeneity; experimental evaluations of how human driving behavior will impact the energy efficiency of autonomous vehicle control, and how changes in autonomous vehicle control strategy will impact human driving.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.