| Popis: |
During the past decades, there has been increasing research in the field of automated vehicle driving. Autonomous vehicle steering control techniques plays a key role and many researchers are investigating different steering control methods to improve autonomous vehicle path following performance. In this dissertation, parameter space approach based robust PID controller design is studied, which takes variable vehicle parameters into account such as vehicle mass, vehicle velocity and road-tire friction coefficient. The designed multi-objective robust PID controller satisfies D-stability, phase margin constraint and mixed sensitivity constraint at the same time. Model predictive control based autonomous vehicle path following is also investigated. It has the advantage in dealing with a wide variety of plant based control constraints systematically. A validated model of our Ford Fusion Hybrid research autonomous vehicle is used for all simulation analyses in this dissertation. In order to further improve vehicle path following performance in the existence of model uncertainty and external disturbance, disturbance observer (DOB) is embedded into the steering to lateral deviation automated driving loop to handle uncertain parameters including vehicle mass, vehicle velocity and road-tire friction coefficient and to reject external disturbances such as yaw moment disturbances. Time delay in the “steer-by-wire” system in the actual vehicle can also lead to stability issues since it adds large negative phase angle to the plant frequency response and tends to destabilize it. Communication disturbance observer (CDOB) which does not require the accurate knowledge of the delay time is injected into the steering actuation loop to compensate this time delay. Motivated by the limitation of disturbance observer and double disturbance observer, a double disturbance observer (DDOB) which is a combined structure of DOB and CDOB is applied to autonomous vehicle path following control system. The DDOB compensated control system realizes external disturbance rejection and time delay compensation simultaneously. The effectiveness of the proposed DOB, CDOB and DDOB compensated control system is validated by improved path following performance compared with the ones of feedback control system only. Furthermore, this dissertation develops control system design from continuous time domain to discrete time domain. Multi-objective robust digital PID controller is developed based on parameter space approach, where phase margin constraint, gain margin constraint and mixed sensitivity constraint are considered as design requirement. The designed digital robust PID controller is used for the autonomous vehicle path following and simulation results validate the effectiveness of the proposed control algorithms. In addition, structures of DOB and CDOB in discrete time domain are also proposed, which are embedded into the digital robust PD controlled system to improve autonomous vehicle path following performance. Safety is another significant consideration during autonomous vehicle path following. When operating the vehicle in urban area where the Vulnerable Road Users (VRUs) consisting of pedestrians and cyclists, it is necessary to respect the social space of VRUs. This dissertation proposes elastic band theory based socially acceptable collision avoidance algorithm. Results demonstrate that the vehicle tries to avoid both stationary and moving VRUs and follows the socially acceptable collision-free path successfully. |
