Overview / Usage

The year 2017 is on track to be the hottest year ever recorded by a wide margin – a whopping 0.17°C hotter than 2014, which previously held that record. Records show global temperatures rising dramatically over the past 20 years and society is starting to see the negative impacts of such climate change. Increased storm severity, as was witnessed with Cyclone Ockhi, is only one of the negative environmental issues that are forcing scientist to take a closer look at our oceans. (Government officials have said that 33,000 people in Kerala and 2,800 in Tamil Nadu have been affected by the cyclone so far. The Centre said 39 people have lost their lives and another 167 are missing after Cyclone Ockhi engulfed parts of Tamil Nadu and Kerala.) The world’s oceans hold a great amount of data that is valuable in understanding environmental conditions; yet difficulties such as cost, time, safety, and range of current data collection methods are limiting the amount of data that we can collect from the environment. We believe the use of Autonomous Surface Vehicles (ASVs) can help to overcome some of these challenges and assist scientists and researchers in their efforts to understand and mitigate the impacts of a rapidly changing environment.

Methodology / Approach

In order to train my DL network, I need huge amount of data. In my case I need to know how a normal designed boat behave to different conditions in the ocean. Collecting those data directly is practically impossible. So I'm modelling those conditions using modelling tools(StarCCM+). And I'm applying reinforcement learning to those simulations. Also I'm currently developing a prototype of my boat for testing purpose, and I'm testing the performance of communication and pathfinder modules. (APM is my controller)