Feature Unified Scoring Engine
Data from multiple sensors is of little use if it cannot be related and joined together (fused) to increase the overall information content. MIKEL has been innovating multi-sensor and muli-data fusion for almost 15 years through novel Track-Before-Detect applications in highly complex sonar environments.
Classically multi-sensor and multi-data fusion is performed using individual measurements from each sensor or data source. However, as the resolution of our sensors or data sources has improved, the sheer volume of measurements produced is simply overwhelming. For almost two decades highly sophisticated measurement-based fusion approaches have been stymied by this volume of measurements, adequately characterizing measurement errors, and optimally performing measurement association. However, MIKEL has honed significantly better approaches that actually do not use measurements at all! We’ve developed a technology we call the Feature Unified Scoring Engine (or FUSE). FUSE leverages the pre-processing output of sensor information directly which both simplifies processing and eliminates the overwhelming burden of measurements.
The genesis of our approaches are early SBIR efforts such as N05-149 Combat Systems of the Future and has offered the first ever application of Track-Before-Detect methods to complex sonar problems* Our unique approaches to fusion have repeatedly proven to outperform measurement-based approaches in the presence of complex sonar environments often by a factor of two or more! Our approaches have applicability in a mulitidue of domains such as self-driving vehicles, unmanned undersea vehicles, and high-precision contact tracking. Our approach is indeed truly novel. Spend some time to check it out!
* T. Northardt, S. Nardone “Track-Before-Detect Bearings-Only Localization Performance in Complex Passive Sonar Scenarios: A Case Study,” IEEE Journal of Oceanic Engineering, April, 2018.