Artificial Intelligence (AI) and machine learning (ML) are increasingly becoming a larger part of our everyday business and personal lives. Data is everywhere and growing exponentially. AI and its underlying algorithms, machine learning and deep learning, combine large amounts of data from disparate sources to find patterns, produce insights and generate valuable information that can solve complex problems and automate operations.
In addition to our services listed below the bullets, some examples of our work with artificial intelligence, machine learning, and deep learning have involved the following platforms and use cases.
TekMark has extensive experience with many programs and use cases involving AI and ML. Some examples include:
- Combining statistical and ML models to asses air-to-ground signal strength, providing a more robust advanced air mobility (AAM) communications system that supports the integration of autonomous and unmanned aircraft into the National Airspace.
- A ML prediction explainability decision support system that quantifies the measure of uncertainty, calibrates trust, and grants trust in the autonomous operation of space base robotic systems.
- Development of a digital twin of a flight management system that resides in the cloud with only non-safety critical functions “twinned”. The digital twin has access to unlimited computer resources and information, much of which is not available elsewhere (aircraft weight, top of climb, top of descent, remaining fuel and burn rate, etc.) along with access to airline AOC and ATC information. This enables experimenting safely with ATC algorithms using actual data to avoid turbulence, avoid contrails, calculate wake vortex for space reduction, and noise calculation for AAM aircraft.
- An ML based service that analyses alternate flight routes between city pairs, filters trajectory options sets by removing routes negatively impacted by weather and aviation constraints (e.g., Traffic Management Initiatives) providing air traffic controllers and other flight management personnel with optimal alternative routes. Registered on NASA’s Digital Information Platform (DIP) and integrated with their Collaborative Digital Departure Re-routing (CDDR) Tool.
- Development of a library of robust ML-based surrogate models trained on the inputs/outputs of NASA’s/NOAA’s physics based atmospheric models that generate global weather forecasts at the same resolution and skill level as their physics-based counterpart, but at a fraction of the processing time.
- Development of an automated, real-time cybersecurity and vulnerability monitoring, assessment and risk mitigation platform that supports autonomous flight operations
