About Antarctic Intelligence & Navigation System (AINDS)
AI-Enabled Antarctic Sea-Ice, Iceberg & Navigation Decision Support System
Problem Statement: Develop an AI/ML-enabled decision support platform capable of forecasting Antarctic sea-ice concentration, predicting iceberg trajectories, and identifying safe and fuel-efficient navigation routes for research vessels using satellite, oceanographic and meteorological datasets.
This platform is an academic prototype engineered for the Smart India Hackathon. It does not claim official Government of India ownership or active maritime licensing. Predictions are intended strictly for research and mission simulation, and must never replace certified polar ice navigator command.
System Architecture & Core Pillars
Hybrid 5-layer bidirectional ConvLSTM paired with a 2D U-Net residual decoder. Ingests 5-day sequential histories of AMSR2 passive microwave brightness temperatures, ERA5 wind barbs, and CMEMS sea-surface temperatures to output 6h to 72h Antarctic ice pack evolution.
Coupled hydrodynamic-atmospheric drift physics incorporating surface and depth-averaged currents (CMEMS 1/12°), 10m wind stress on iceberg sails (ERA5), Coriolis acceleration, and sea-ice resistance with bivariate Gaussian confidence ellipses.
Continuous tensor formulation R(φ, λ, t, v) evaluating space, time, vessel ice class, and composite hazards: sea ice concentration, iceberg proximity buffers, wave squalls, and GEBCO under-keel clearance.
Constrained Pareto-A* graph search integrating the Lindqvist continuous ice resistance formulation. Yields three distinct non-dominated strategies: Safest, Balanced, and Fastest.