Generic dynamic modelling and model-based trajectory tracking control of wheeled mobile robots (WMRs)
This thesis focusses on the dynamic modelling and on the systematic evaluation of model-based control schemes for trajectory tracking of Wheeled Mobile Robots (WMRs) for industrial applications. The research extends the theoretical and experimental work at the Centre for Industrial Control, Departme...
Main Author: | |
---|---|
Format: | Others |
Published: |
1997
|
Online Access: | http://spectrum.library.concordia.ca/392/1/MQ40219.pdf Thanjavur, Karunananth G <http://spectrum.library.concordia.ca/view/creators/Thanjavur=3AKarunananth_G=3A=3A.html> (1997) Generic dynamic modelling and model-based trajectory tracking control of wheeled mobile robots (WMRs). Masters thesis, Concordia University. |
Summary: | This thesis focusses on the dynamic modelling and on the systematic evaluation of model-based control schemes for trajectory tracking of Wheeled Mobile Robots (WMRs) for industrial applications. The research extends the theoretical and experimental work at the Centre for Industrial Control, Department of Mechanical Engineering, Concordia University on mobile robotics since 1982. The methodology applies the Kane's approach to model WMRs. This approach uses the natural degrees of freedom and emphasises the system's non-holonomic nature, unlike schemes based on the Newton-Euler and Lagrangian methods. The novel 5-step procedure is applied to model WMRs with common wheel types and configurations for various loading and operating conditions. The dynamic loads on the wheels are obtained using 'virtual' speeds. The methodology is illustrated by deducing the dynamic equations of common WMRs. The insight into the mobile robot's dynamic behaviour provided by the model is used for model-based trajectory tracking control. The performances of the Feed Forward and the Computed Torque schemes have been evaluated through simulations and with experiments using a differentially driven WMR. The experimental results indicate that the inclusion of the dynamics in the closed control loop reduces tracking errors by an order of 50% in comparison with a conventional PD controller. Results also indicate the importance of accounting for the dynamic nature of wheel loads and of dynamic model parameters in the Feed Forward controller performance |
---|