• 1. 시간: 2024년 7월 26일 (금요일) 2:00 ~ 4:00
• 2. 장소: 연구협력관 103호
• 3. 연사: 한경석 교수(경북대학교 기계공학부)

Title: Interaction and Energy-aware Control Design for Vehicle Autonomy and Electrification

Over the past decade, advancements in vehicle autonomy have captured significant attention from both automakers and academia. This increasing attention has led to a lot of work to make autonomous vehicles safer and more efficient. In the near future, it is clear that roads will be shared by both human-driven and autonomous vehicles. In such a landscape, ensuring the safe operation of autonomous vehicles becomes a critical concern, especially considering the need to interpret and respond to the behaviors of surrounding vehicles. Additionally, the integration of vehicle connectivity technologies presents new opportunities for achieving more energy-efficient and secure vehicle path and speed planning. This integration not only improves individual vehicle operations but also enhances overall traffic efficiency.

In this talk, we will first explore the application of game-theoretic and interaction-aware decision-making and control design for autonomous vehicles. This approach is crucial for developing systems that can dynamically adapt to the changing behaviors of other traffic participants. Furthermore, we will discuss the methods that leverage preview traffic information to improve the efficiency of electrified vehicles. These systems predict changes in traffic flow and adjust accordingly to use energy more effectively and decrease stops and starts, which typically waste a lot of energy.

A significant portion of the talk will be dedicated to discussing the challenges and solutions related to the design of model predictive control (MPC) systems. MPC is renowned for its ability to handle multivariable control problems with constraints; however, it often involves complex computations that can be computationally intensive. Towards the conclusion of the talk, we will introduce our recent efforts aimed at reducing the computational complexity associated with MPC. These developments are pivotal in designing MPC more practical and applicable. Also, the data-driven approach to design the MPC will be briefly touched, resulting in improving the model accuracy.