Algorithm Design and Validation
Mosaic ATM designs complex, mathematical algorithms at the core of ATM concept alternatives. Our agile approach begins with algorithms appropriate for proof-of-concept activities and iterates to reach deployment-ready, validated algorithms that handle real-world data anomalies and off-nominal situations. We are experts in the development of algorithms that inherently address the uncertainty of input data.
Mosaic also has extensive experience developing and applying advanced analytical techniques across a wide variety of non-aviation business questions and use cases. Because of our depth of talent and expertise in machine learning, artificial intelligence, and other cutting-edge capabilities, Mosaic personnel will apply the right tools, techniques, and technology in collaboration with your team.
Our expertise in algorithm design and development includes:
- Optimization, including stochastic optimization
- Mixed integer linear programming
- Network optimization
- Various clustering techniques
- Genetic algorithms
- Neural networks and deep learning
- Predictive analytics and machine learning
- Principal components analysis
- Random forests, support vector machines, logistical regression
- Collaborative filtering
- Scheduling
- Trajectory prediction
- Taxi in/out prediction
- Computational geometry
- Algorithm parallelization and performance techniques
Mosaic is at the forefront of algorithm design and validation in aviation and other domains.
Mosaic is developing a real-time trajectory prediction model for the AEGIS combat system that uses artificial intelligence and machine learning techniques along with a modular and scalable software architecture that supports both its research and deployment
US Navy AEGIS Combat System Anomaly Detection
Mosaic has developed custom algorithms that predict the optimal amount of taxi time. Our scientific background gives this solution unparalleled predictive and optimization capabilities.
Taxi Time Prediction
Mosaic is providing an adaptable, modular, and automated labeling solution that evolves with the rapidly changing environments faced by modern warfighters. Our solution reduces the burden to label tactically relevant information in full motion video and still imagery, enabling automated high-quality data preparation at scale for the warfighter.