Abstract
This study explores enhanced brush modelling and control techniques with respect to facilitating the development of a new generation road-sweeping vehicle. Achieving automatic road sweeping was difficult in the past, partly because brush characteristics were not fully understood thus it was not known how to properly use a brush to sweep debris more efficiently. A number of results to overcome such a constraint are presented in this work. In the first part of this thesis, a Finite Element (FE) model for analysing brush characteristics is presented. The FE model calculates three-dimensional brush tine deformation to give a general prediction of brush mechanical characteristics, such as force-deformation relationship, brush shape, rotational torque, etc. This model is an extension work of a previous mathematical model developed by Peel (2002), with the advantages that it provides better accuracy and that it can be used to analyse different brush types. The modelling results can also be utilized for a wide range of brush design applications. Secondly, road sweeping effectiveness has been assessed using both theoretical and experimental methods. In order to deduce the fundamental debris removal mechanisms, a FE model has been developed to analyse the interaction between brush tines and debris. By implementing sweeping tests on different classes of debris, optimized criteria for configuring brushing force, tilt angle and rotational speed have been suggested. These criteria can be used to design a predictive brush control system. This study finally addresses several considerations for designing an automatic brush control system. It is shown that the performance of such a system not only depends on the optimized brush configuration criteria, but also depends on reliable servo controllers, which should be able to deal with brush nonlinear characteristics. It is highlighted that the FE modelling results should be utilized to predict the brush stiffness, so that an auto-tuning PD controller can be developed to improve the control performance.