Control of Truck and Multiple Trailer Systems

Abstract

A truck and multiple trailer system consists of a wheeled mobile robot that tows multiple trailers. These systems are widely used across a variety of industries for both indoor and outdoor services. Based on the kinematic model, the truck and multiple trailer system is considered to be a typical nonlinear system. Thus, controlling a truck and multiple trailer system is a challenging research topic that has received increasing attention from both academics and engineering practitioners.

The use of a truck and multiple trailer system is important as it can boost the capacity of land transportation vehicles. However, an insufficient degree-of-freedoms in the available controls means that the steering of such systems remains problematic. Indeed, as the numbers of trailers in any system increases the issues become increasingly complicated. This research studies three intelligent control approaches that progressively improve system performance, such that the requirements of an exact and sophisticated system model can be relaxed.

For the first approach, a virtual-robot tracking strategy was employed whereby a truck was directed to follow a specified trajectory. The generation of drive speed and turn-rate was viewed as an intelligent optimisation problem and a particle swarm optimisation algorithm was used for performance and implementation simplicity. Simulation studies in which a truck-and-trailer system moved across irregular paths were conducted using a developed intelligent controller. Satisfactory results supported the feasibility of this approach.

For the second intelligent control approach, a hybrid force field controller was developed. Previously, the design of controllers for truck and multiple trailer systems have, for most cases, assumed collision free workspaces; however, for practical applications the linear velocity of the truck had to be adjusted so that it followed an arbitrary reference trajectory and avoided obstacles. A controller was designed hybridising virtual vehicle control strategy, force field approach and the particle swarm optimisation algorithm. A linear velocity was obtained that provided the desired positioning of the truck and multiple trailer system in a practical working space.

Finally, a fuzzy logic controller was designed for a truck and multiple trailer system whereby the controller generated the drive velocity and turn-rate for the truck alone. An automatic fuzzy rule and membership tuning approach was developed that eradicated the need to rely solely on human expert inputs. The problem was approached using a particle swarm optimisation algorithm that removed the need for human expert knowledge input while searching for a near optimal solution. Thus, the truck and- multiple trailer system could be driven to a directed position following a trajectory on a two-dimensional terrain. The simulation results of both a one-trailer system and multi-trailer system showed the effectiveness of the derived fuzzy controller.

The outcomes reported in this thesis have contributed to the development of intelligent motion controllers for the truck and multiple trailer system. In particular, intelligent design approaches were adopted that mitigated the need for an exact knowledge of a complicated system model.