Artificial Intelligence to Support Autonomous Seabed Mapping Operations

Presently the Ocean Mapping Community is spending a lot of time manually cleaning the raw sonar survey files. If this cleaning process is done by an Artificial Intelligence system as clean as a human would do, this could be transforming the entire Ocean Mapping community. Therefore, we propose a novel solution of using an Artificial Intelligence Algorithm – Reinforcement Learning to clean the sonar survey files. The proposed algorithm was used by Google’s DeepMind to beat GO Grandmasters. We believe that by tapping this algorithm we would be able to clean the raw sonar survey files as efficient as the manual cleaning of the raw sonar survey files.

Faculty Supervisor:

Ian Church

Student:

Vishwa Barathy Gandhi Kalidasan

Partner:

Springboard Atlantic

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

University of New Brunswick