Research

RF & Electromagnetic Perception

Research areas include:

• radio-frequency sensing

• software-defined radio

• cellular signals

• Wi-Fi and Bluetooth

• passive RF detection

• spectrum monitoring

• electromagnetic phenomena

• signal classification

Key question:

What information exists in the electromagnetic environment that autonomous systems could perceive and interpret?

Optical & Machine Vision

Research areas include:

• visible-light imaging

• low-light imaging

• machine vision

• multispectral imaging

• hyperspectral imaging

Key question:

What conditions can machines recognize visually that humans might miss, overlook, or be unable to safely observe?

Thermal & Infrared

Research areas include:

• infrared sensing

• thermal imaging

• heat signatures

• temperature anomalies

Potential applications include inspection, safety, search, environmental awareness, and anomaly detection.

Acoustic & Ultrasonic Perception

Research areas include:

• conventional acoustic sensing

• directional audio

• ultrasonic sensing

• machine-based acoustic classification

• anomaly detection

Machines may be able to recognize signatures in sound that humans cannot reliably detect or interpret.

Chemical & Environmental Sensing

Research areas include:

• toxic gases

• volatile organic compounds

• carbon monoxide

• particulates

• air quality

• radiation

• temperature

• humidity

• pressure

• hazardous environmental conditions

A central use case is allowing autonomous systems to characterize potentially dangerous environments before human entry.

Spatial & Motion Perception

Research areas include:

• LiDAR

• radar

• ultrasonic ranging

• inertial measurement

• magnetometers

• positioning

• proximity sensing

• motion sensing

These capabilities help autonomous systems understand where objects are, how environments are changing, and how the sensing platform itself is moving.

Sensor Fusion

Individual sensors frequently provide incomplete or ambiguous information.

Trusted Perception investigates whether combinations of sensing modalities can produce substantially better understanding than individual sensors alone.

Questions include:

• Which sensor combinations provide complementary evidence?

• How should conflicting observations be handled?

• How should sensor reliability affect a fused conclusion?

• Can one sensing modality help validate another?

Edge AI

Sending every sensor observation to a remote cloud system may be impractical or undesirable.

Trusted Perception investigates local processing that can:

• reduce communications requirements

• support low-latency decisions

• preserve sensitive information

• continue functioning when disconnected

• combine multiple sensor streams locally

• recognize conditions or anomalies near the point of collection

Trust, Security & Assurance

A trusted perception system should be able to provide more than a prediction.

Research questions include:

• How confident is the conclusion?

• What evidence produced it?

• Which sensors agreed?

• Which sensors disagreed?

• Could a sensor have failed?

• Could the input have been spoofed?

• Has the operating environment changed?

• Is the model outside its validated conditions?

• What uncertainty should be communicated to the operator?

This research connects perception with cybersecurity, trustworthy AI, resilient communications, assured autonomy, and human-machine teaming.