Topological Data Analysis Machine Learning Algorithm (TDAML)

Integrating TDA with Artificial Intelligence/Machine Learning (AI/ML) for Data Fusion and Target Recognition
Download Tech Brief
Benefits & Applications
Contact Us to Partner
Return to T2 Portal

Relevant Publications:

  1. Interfacing topological data analysis (TDA) with AI/ML for multimodal data fusion and automatic target recognition (ATR)
  2. Highway slope monitoring using 3D laser scanning at different seasons
Download Tech Flyer

Problem

Hover/Click to Flip

Autonomous systems often struggle to integrate data from diverse sensors due to complexity, interference, and the limitations of slow, resource-intensive methods.

Solution

Hover/Click to Flip

The TDAML method transforms diverse data types into compact features, enabling more efficient fusion for target recognition and autonomy with less data.

Benefits of TDAML:

  • Creates compact "topological fingerprints" for data samples.
  • Transforms raw sensor data into deep learning-friendly formats
  • Uses TDA to encode and fuse data for AI/ML systems.
  • Enhances resilience against interference and adversarial threats.
  • Reduces data size while preserving detection accuracy.
Escape Data Collection: 30TB of sensor data from 10 modalities across 4 UAVs, 7 ground vehicles, and 5 dismounts at Griffiss Airfield & Stockbridge Test Site.

Applications for TDAML:

Collision Mitigation in Autonomous Vehicles

Autonomous Target Recognition (ATR)

Deployable Autonomy

Data Fusion in Contested Environments

Intelligence, Surveillance, and Reconnaissance (ISR)

Explore Partnership Opportunities for this Technology

Contact the Griffiss Institute Innovation & Partnerships Team to Learn More!

Tanya Weller
Senior Manager, Commercialization & Innovation

 
 

tweller@griffissinstitute.org

Linkedin Envelope

Contact Form

Interested in learning more about Innovation & Partnerships?

Return to T2 Portal
Scroll to Top

MVGives Now Accepting Donations Through 9/19 @ 9PM EST

Donate Today