The problem OilGuard solves

Overview of the Project: AI-Driven Solution for Oil Spill Detection and Cleanup

• Problem: Oil spills in marine environments harm marine life, coastal habitats, and industries. Traditional methods of detection and cleanup are slow and heavily manual.
• Solution: Our project uses AI technology to provide a faster, safer, and more efficient way to detect and clean up oil spills.

Core Innovation

• AIS Data and Satellite Imagery: The system combines AIS data with satellite images for effective monitoring of oil spills.
• Machine Learning Models: These models predict anomalies in AIS data and use real-time satellite images to identify potential spills.
• Reduced False Alarms: The system reduces false alarms, allowing for faster response and minimizing ecological damage.

Autonomous Response

• ASV Deployment: Autonomous Surface Vehicles (ASVs) with booms, skimmers, and tankers are deployed to collect and remove oil from the water.
• Autonomous Operation: ASVs operate autonomously in hazardous or hard-to-reach areas, reducing the need for human involvement.

Economic and Preventive

• Cost-Effective: The technology reduces the need for manual deployment, lowering overall costs.
• Predictive Capabilities: It provides early warnings to maritime authorities and oil companies about high-risk zones.
• Investments: One-time investment and further maintenance.
• Real-Time Response: The system quickly moves from detection to containment, minimizing response time.

Target Audience and Impact

• Audience: Governments, oil companies, and coastal authorities responsible for marine safety and environmental compliance.
• Impact: Provides automated, continuous monitoring, simplifying oil spill detection and cleanup. This protects marine biodiversity and coastal economies, reducing environmental damage.

Challenges we ran into

Challenges Faced and Solutions

Data Inconsistency and Cleaning

• Problem: Data inconsistency was a major issue, needing extensive cleaning.
• AIS Data Issues: Missing or incorrect values in key fields like ‘Navigational Status’ and ‘ROT’ (Rate of Turn) affected anomaly detection.
• Preprocessing: We developed scripts to clean and preprocess the data for our machine learning models.
• Steps Taken:
  • Identifying Outliers: Pinpointed outliers to prevent skewed results.
  • Handling Missing Values: Addressed missing data through imputation or removal.
  • Normalization: Standardized the dataset to ensure consistency in training.

ASV Prototype Development

Vessel Stability

• Challenge: The initial monohull design was unstable, especially in rough waters.
• Solution: Switched to a catamaran hull design, which improved stability by balancing weight between two hulls, reducing the risk of capsizing in challenging conditions.

Motor System and Control

• Problem: A single motor caused maneuverability and control issues.

• Solution: Adopted a dual-motor system, with one motor per hull, enhancing control and navigation precision.

• Relay Synchronization:

  • Challenge: Synchronizing two relays for the motors was complex.
  • Solution: Simplified the design by using one relay to control both motors, reducing complexity and improving reliability.

• Outcome: These solutions refined the ASV prototype, making it more effective for oil spill recovery operations.