Autonomous navigation of a partially unknown environment

Pont, Jason (2019) Autonomous navigation of a partially unknown environment. [USQ Project]

Text (Project)

Download (1MB) | Preview


The aim of this project was to investigate navigation methods for supporting autonomous operations on Mars. To date there have been several robotic rover missions to Mars for the purpose of scientific exploration. These missions have relied heavily on human input for navigation due to the limited confidence in computer decision-making and the difficulty in localising an unknown environment with limited supporting infrastructure, such as satellite navigation. By increasing the confidence in the performance of an autonomous rover on Mars, this project will contribute to increasing the efficiency of future missions by reducing or removing humans from the control loop.

Due to the signal propagation delay between Earth and Mars, a certain level of autonomy is required to ensure a rover can continue operating while awaiting instructions from a human on Earth. However, due to the level of risk in relying solely on automation, there is still considerable human intervention. This can result in significant downtime when awaiting a decision by a human operator on Earth. While acceptable for scientific missions, greater autonomy will be required for routine Mars operations.

The project reviewed systems and sensors that have been used on previous robotic missions to Mars and other experiments on Earth. The most appropriate systems were assembled into a simulated test environment consisting of a small rover, an overhead camera that might be carried by a drone or balloon and wireless communications between the systems. A machine vision algorithm was developed to test the concept of an overhead camera mounted on a drone or balloon, while evaluating different path-planning algorithms for speed in navigating a previously unknown environment. An experimental system was built consisting of a rover, fixed overhead camera and communications between them. The machine vision algorithm was used to send instructions to the rover which could then follow a path through a test environment with different obstacle densities. Two different path-finding algorithms were tested with the system.

The key outcomes of the project were the construction and testing of the system. The rover could navigate, rotate towards and travel to a target location, after receiving instructions via serial radio communications. The rover could also detect obstacles using an ultrasonic sensor and send this information back to the machine vision algorithm. The algorithm would then update the path with the new information received on obstacle locations and the rover would then follow the new path to the target location. By successfully testing the concept, the project showed that this system could be used to support future scientific missions, resource gathering and preparation for human exploration of Mars.

Statistics for USQ ePrint 43183
Statistics for this ePrint Item
Item Type: USQ Project
Item Status: Live Archive
Faculty/School / Institute/Centre: Historic - Faculty of Health, Engineering and Sciences - School of Mechanical and Electrical Engineering (1 Jul 2013 - 31 Dec 2021)
Supervisors: Lobsey, Craig
Qualification: Bachelor of Engineering (Honours) (Mechatronic)
Date Deposited: 25 Aug 2021 00:31
Last Modified: 26 Jun 2023 23:03

Actions (login required)

View Item Archive Repository Staff Only