Skip to content
KPT — Korea Positioning Technology

KAILOS NEURAL ENGINE

Inertial intelligence at the edge.

A compact AI-powered positioning module designed to turn motion sensor data into real-time trajectory and movement intelligence.

KAILOS Neural Engine module, back view with a board-to-board connector, pin headers and passive components
Back view
KAILOS Neural Engine module, front view: a compact blue circuit board with processor and memory packages, a pin header and four mounting holes
Front view

MOTION UNDERSTANDING

From sensor measurements to motion understanding

Raw inertial measurements describe acceleration and rotation. Useful positioning requires understanding what that motion means over time. The KAILOS Neural Engine is designed to apply real-time inference and sensor fusion to motion data so embedded systems can estimate trajectories and recognize movement.

Accelerometer

Gyroscope

Real-time inference

  • Trajectory estimation

  • Motion detection

Acceleration and rotation measurements are interpreted over time with real-time inference and sensor fusion to estimate trajectories and recognize movement.

EDGE INTEGRATION

Built for edge integration

The module is designed for systems where size, power consumption and real-time processing matter. Positioning intelligence can therefore live closer to the moving person or machine instead of depending entirely on remote processing.

Remote processing only

raw motion data

Intelligence at the edge

Trajectory and movement estimated on the module
  • Size
  • Power consumption
  • Real-time processing
Designed for systems where size, power consumption and real-time processing matter, so positioning intelligence can live on the moving person or machine.

LEARNED MOTION

Motion, understood.

Inertial sensors start to drift the moment they are left on their own. The KAILOS Neural Engine runs learned motion models on the module itself — it recognises how a person walks, a vehicle turns, a drone banks or a vessel glides, and keeps the trajectory on course where raw integration wanders off.

Illustration. The same motion data, integrated on its own (amber) and through learned motion models (cyan).

TECHNICAL OVERVIEW

KAILOS Neural Engine at a glance

Module size
24 × 20 mm
Power
< 0.25 W
Primary motion inputs
Accelerometer + gyroscope
Inference
Real-time inference
Integration
Designed for embedded integration
Sensor fusion
AI sensor fusion
Outputs
Trajectory estimation / Motion detection

APPLICATIONS

Motion intelligence for people and machines.

Pedestrian Tracking

Estimate human movement where direct satellite positioning is limited.

Drone Navigation

Add motion and trajectory intelligence for indoor, outdoor and transition environments.

Robot Autonomy

Provide inertial motion information as part of a broader autonomous positioning and perception stack.

SOLUTIONS

Defense & Mission-Critical

Extend positioning awareness into operational environments where relying on satellite positioning alone may not be sufficient.

Explore Defense & Mission-Critical

Emergency Response

Support responders with location information as they move from open outdoor areas into structures where satellite signals may become unreliable.

Explore Emergency Response

Add positioning intelligence to your product.

Tell us about the sensors, compute platform, dynamics and positioning requirements of your system.

Discuss Integration