Modeling and control of single-track vehicles: A human-machine-environment interactions perspective

Single-track vehicles, such as motorcycles and bicycles, not only provide an everyday transportation means and recreational sport, but also offer an excellent platform to study physical human-machine-environment (HME) interactions. The main goal of this dissertation is to present a modeling and control system design framework for HME interactions in single-track vehicle systems. The dissertation focuses on three aspects: autonomous vehicle design, vehicle-environment interaction, and human-vehicle interaction. First, we propose novel modeling and control designs for riderless single-track vehicle to achieve agile maneuver navigation and stationary balancing. To achieve agile maneuver, the zero lateral velocity nonholonomic constraint at the tire contact point is relaxed. An empirical tire-road friction model is explicitly considered in the dynamic model. An external/internal convertible (EIC) model-based controller is designed for both trajectory tracking and path following strategies. Two different control designs are then presented to balance the stationary bicycle through steering control and gyroscopic actuator control, respectively. To capture the vehicle-environment interaction, the second part of the dissertation focuses on the study of the tire-road interaction. A high-fidelity tire model is proposed and built on the calculation of the deformation and friction force distributions in stick-slip transition. An in-situ sensing technique is also developed to directly measure the friction force distribution. The model and the sensing development can be further used for facilitating real-time friction parameter estimation and vehicle safety control. The third part of the dissertation mainly discusses the human-vehicle interaction. A dynamic model is first proposed to capture the physical rider-bicycle interaction. A novel pose estimation approach is developed to integrate the wearable inertial sensors with on-board force sensors. A balancing design is finally presented to control the stationary rider-bicycle interaction. All the modelings and control designs in the dissertations are validated through extensive simulations and experiments. The outcomes of the dissertation provide not only a modeling and control framework but also a physical experimental platform to study the unstable HME interactions. We discuss the future research direction at the end of the dissertation.