Skip to main content

Applied Research Portfolio

Applied AI and Edge Intelligence

Active SBIR proposals applying Physics-Informed Neural Networks to Army, NASA, and Navy programs — alongside core capabilities in edge ML, battery diagnostics, and physics-constrained modeling.

Proposed programs

Active SBIR Proposals

The following programs are proposed SBIR Phase I efforts currently under agency evaluation. Technical details are presented at the capability level.

SBIR Phase I · ProposedU.S. Army

Depot-Doctor

Battery Diagnostics

State-of-health estimation and remaining-useful-life prediction for Army 6T battery systems. Physics-Informed Neural Networks embed electrochemical constraints directly into the model — providing physics-grounded predictions for depot-level diagnostic workflows where black-box ML is not acceptable.

Physics-Informed Neural Networks · 6T Battery SoH

SBIR Phase I · ProposedNASA

Swarm-Doctor

Spacecraft Health

State-of-health monitoring across spacecraft constellations, applying Physics-Informed Neural Networks to satellite power and battery systems. Designed for fleet-level prognostics where individual sensor data is sparse and physics constraints must compensate for incomplete observability.

Physics-Informed Neural Networks · Constellation SoH

SBIR Phase I · ProposedU.S. Navy

SHADOW

Hypersonic Analysis

SPIKAN-Accelerated Hypersonic Analysis, Design, Optimization Workflow. A PINN-based framework for the analysis, design, and optimization of hypersonic systems — reducing the high-fidelity simulation burden by embedding physical constraints directly into the learning process.

Physics-Informed Neural Networks · SPIKAN · Hypersonics

Capability foundations

Underlying Technical Domains

Diagnostics·Physics-constrained, field-deployable

Battery health diagnostics

State-of-health and remaining-useful-life estimation for battery systems, with safety-first modeling assumptions and maintenance decision support.

Edge / Embedded·Disconnected ops, constrained hardware

Edge ML and sustainment intelligence

Deployment-aware inference patterns for constrained environments, disconnected operations, and field-level prognostics.

Modeling·Hybrid first-principles + ML

Physics-informed modeling

Hybrid approaches that combine first-principles domain constraints with machine learning for systems where black-box predictions are not acceptable.

Documentation·Validation evidence, tech transfer ready

Secure transition documentation

Architecture notes, validation evidence, and operational guidance written for technical reviewers, safety assessors, and integration partners.

Collaboration model

Work described at capability level, not contract level

Engagements are scoped around capability gaps, not retrofitted to existing program structures. Technical depth is available for teaming partners who need architecture discussions under NDA.