Accurate parking of mobile robots using visual servoing

• Main Author: Tan, Paul Qian Rui
• Other Authors: Teoh Eam Khwang
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/78307
• Institution: Nanyang Technological University

Traditionally, robotic navigation and auto-parking have been based on probabilistic methods that require fine-grained location grid maps and a high level of localization precision using expensive laser range finding sensors. In this project, we explore an alternative visual navigation approach based on the matching of images from low cost cameras. The inspiration behind this approach is because as humans, we do not navigate our way by moving along paths with precise coordinates every step we take, but rather, along paths guided by a mental model of the environment and the route we are taking. For a mobile robot to navigate from a known start and to the destination, the proposed approach adapts a hybrid visual servoing into the parking systems and attempts to navigate itself to the final position. The objective of this final year project is to focus on the implementation of the proposed visual servoing framework using MATLAB, which comprises of a few functions and the motion estimation algorithm. An existing parking controller used for asymptotic stabilization of robot is adopted in the testing and the proposed framework is implemented on a differential drive mobile robot model. Our proposed framework can achieve a practical exponential stability despite the constraints of parking a nonholonomic mobile robot. With its intuitiveness and efficiency, our implementation has the potential to be used in actuality for nonholonomic vehicles parking, such as the parking of wheelchairs and the recharging of vacuum cleaners at their parking docks. The report begins with an introduction to the parking of nonholonomic robots and traditional methods of visual servoing. Existing methods and limitations will be discussed along with the resolutions provisioned by the proposed visual servoing framework. The remaining chapters will present the simulation and implementation results on a mobile robot. The report will conclude with a summary and prospective plan for future work.