Browsing by Author "Hashem Ali, Mohammed Abdo"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Autonomous mobile robot navigation and control in the road following and roundabout environments incorporating laser range finder and vision system
    (Universiti Teknologi Malaysia, 2014) Hashem Ali, Mohammed Abdo
    This research focuses on the development of a wheeled mobile robot (WMR) model that is able to navigate autonomously on the road and open space area in a road roundabout using natural landmarks. A novel search graph approach for path planning has been developed to find the mobile robot path in a complicated unknown environment. This approach gives the possibility to consider some constraints to the robot movement as in the real world application such as on the road. A simulation study of the WMR employing a laser simulator (LS) approach is applied to the roundabout grid map settings for determining a free collision path of the WMR from the start to goal positions. The resulted path navigation is implemented with a control system that enables the robot to track robustly a pre-planned path when passing through a roundabout using MATLAB/Simulink, considering a number of scenarios. Thus, a new WMR platform for autonomous on-the-road navigation both in simulation and experimental works is designed, developed and tested in the laboratory. The realized WMR was equipped with a number of sensors including a camera, laser range finder (LRF) and odometry device to determine the free collision path during navigation. An embedded controller system employing a number of feedback control schemes and navigation algorithms has been developed to integrate the mechanical components, electronics and computer programs via suitable interfacing configuration. The experimental work investigates a complete autonomous WMR path planning and motion control in the road following and roundabout environments. The sensor fusion technique is also utilized to gather all the sensors data to perform the path planning and localization of the robot within the surrounding terrain based on the proposed navigation algorithms. The local maps were built using data from both camera and LRF to estimate the road border parameters such as road width, curbs and roundabout diameter. The algorithm for the path generation is fully derived from the local maps and subsequently implemented into WMR for indoor and outdoor environments. A localization algorithm is used to remove the noises and uncertainties of the data from the sensors and provide the optimum measurements to the path planner by applying the Kalman filter. The path execution is controlled by the external and internal feedback control loops using resolved acceleration control (RAC) and active force control (AFC)- based strategies. As a practical case study and for validation, an automatic road mark painting system is developed and incorporated into the WMR to perform an autonomous detection of the road layout and a painting task. Experimental results show the capability of the algorithms and control schemes to robustly drive the WMR in the road following and roundabout environments, thereby verifying the simulation counterpart