Navigation and Control of an Autonomous Vehicle
The navigation and control of an autonomous vehicle is a highly complex task. Making a vehicle intelligent and able to operate â unmannedâ requires extensive theoretical as well as practical knowledge. An autonomous vehicle must be able to make decisions and respond to situations completely on...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Others |
| Published: |
Virginia Tech
2014
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10919/32634 http://scholar.lib.vt.edu/theses/available/etd-05122005-034726/ |
| Summary: | The navigation and control of an autonomous vehicle is a highly complex task. Making a vehicle intelligent and able to operate â unmannedâ requires extensive theoretical as well as practical knowledge. An autonomous vehicle must be able to make decisions and respond to situations completely on its own. Navigation and control serves as the major limitation of the overall performance, accuracy and robustness of an autonomous vehicle. This thesis will address this problem and propose a unique navigation and control scheme for an autonomous lawn mower (ALM).
Navigation is a key aspect when designing an autonomous vehicle. An autonomous vehicle must be able to sense its location, navigate its way toward its destination, and avoid obstacles it encounters. Since this thesis attempts to automate the lawn mowing process, it will present a navigational algorithm that covers a bounded region in a systematic way, while avoiding obstacles. This algorithm has many applications including search and rescue, floor cleaning, and lawn mowing. Furthermore, the robustness and utility of this algorithm is demonstrated in a 3D simulation.
This thesis will specifically study the dynamics of a two-wheeled differential drive vehicle. Using this dynamic model, various control techniques can then be applied to control the movement of the vehicle. This thesis will consider both open loop and closed loop control schemes. Optimal control, path following, and trajectory tracking are all considered, simulated, and evaluated as practical solutions for control of an ALM.
To design and build an autonomous vehicle requires the integration of many sensors, actuators, and controllers. Software serves as the glue to fuse all these devices together. This thesis will suggest various sensors and actuators that could be used to physically implement an ALM. This thesis will also describe the operation of each sensor and actuator, present the software used to control the system, and discuss physical limitations and constraints that might be encountered while building an ALM. === Master of Science |
|---|