Artificial Intelligence systems for AUVs
Improving the performance of AUVs using AI methods.
Project leader
Dr Rosemary Norman
Dr Maryam Haroutunian
Jeffrey Neasham
Dr Alan J Murphy
Dates
2018 to 2021
Project staff
Mr Zhizun Xu (PhDcandidate)
Description
Oceanography has traditionally relied on ship-based observations. Recently, robotic platforms such as Autonomous Underwater Vehicles (AUV) have augmented these observations. AUVs are untethered powered mobile robots able to carry a range of payloads over large distances in the deep ocean.
Past research highlights the limitations of navigation and communications. It also shows that the intelligence of AUVs is a crucial challenge. AUVs need planning and fault-tolerance abilities to achieve mission goals. Moreover, AUVs travel underwater without communicating with support ships or shore for hours or days. They execute more and more complex missions. Thus, Artificial Intelligent (AI) for AUVs with decision-making ability will become a trend in future.
Techniques of navigation and acoustic communication have undergone rapid development. But the intelligence of AUVs appears to have remained at a low level for many years. For example, the AUV system is unable to respond to abrupt changes in the external environment, system damage, and uncertain or indeterminate data input.
We will improve in-situ decision making and fault tolerance for AUVs. To do this, we will use deliberative architecture and AI methodologies.