response reality
Illustrative concept photograph of a robotics engineer interacting with a humanoid robot in a professional testing laboratory

Human movement.
Meaningful information.

A human-understanding layer for intelligent systems.

Understanding
the interaction.

We study how people and machines coordinate-and how that coordination changes.

See how it works

Observe interaction

Bring body movement, controller input and machine telemetry together on a shared timeline.

Measure coupling

Track how human actions and machine responses coordinate-and detect changes in that relationship.

Interpret changes

Use interaction patterns and context to assess whether a change comes from the person, interface, machine or task.

Inform assistance

Give the machine coupling estimates and confidence levels to guide assistance, timing and autonomy.

Human-machine coupling

One interaction.
A shared understanding.

A person acts. The machine responds. That response shapes the person’s next action. Human-machine coupling describes how well the two coordinate through this continuous loop.

The Human State Layer brings movement, control inputs and machine feedback onto a shared timeline. It measures changes in coordination, then uses the task context to assess where a change may come from: the person, interface or machine.

The result is useful information for the machine’s existing decision logic: what is changing, how certain the estimate is, and when support may help. Assistance can adapt its timing or intensity while the person retains control.

01

Observe the interaction

Three views of the same moment.

HumanMovement & gestures
InterfaceControl inputs & timing
MachineResponses & task feedback
02

Understand coordination

Human State Layer
Measure coupling

How well do actions and responses fit together?

Interpret change

What changed-and what is its likely source?

Estimates include confidence and task context.

03

Guide the response

The machine uses this information to choose appropriate support.

  • Adjust the pace or level of assistance
  • Give clearer feedback or ask for input
  • Support a handover when needed
The person retains control.Assistance stays within the system’s defined limits.
Every response becomes part of the next interaction.
From signals to shared understanding to context-aware assistance.
Use cases

Human understanding.
Real-world relevance.

Where human understanding can make a difference: four possible applications of the same interaction loop.

Illustrative concept photograph: Firefighters operating a reconnaissance drone at a fire-service training site
01

Firefighting drones

Coordinate under pressure.

A firefighter’s control inputs and the drone’s flight data reveal how well they are working together. Changes in coordination can guide clearer feedback and timely flight assistance.

Illustrative concept photograph: An older adult interacting with an assistive robot while a care professional observes
02

Robots in elderly care

Support at the person’s pace.

Movement, gestures and robot feedback help identify when an interaction slows or becomes difficult. The robot can pause, clarify its intention or adjust the pace of assistance.

Illustrative concept photograph: Driver at the wheel of a production vehicle with driver assistance
03

Driver assistance

Make handovers more informed.

Steering behavior, driver movement and vehicle responses provide a shared view of the interaction. This can guide the timing of prompts and transitions between driver and automated support.

Illustrative concept photograph: Supervised VR interpersonal training with a responsive virtual avatar visible on a monitor
04

Responsive virtual people

An interaction that responds to you.

In therapy and training scenarios, movement and conversational timing can help a virtual character respond to the participant-adapting its pace, prompts and behavior as the interaction unfolds.

Grounded in behavior.
Built for real interaction.

Our foundation is movement-dynamics research in VR. We are investigating how that principle extends to human-controller-machine systems.

Body keypoints, interaction signals and machine telemetry can provide complementary information. Meaning comes from how the signals relate to the person, the task and the system-not from a movement in isolation.

Explore Our Research

Building intelligent systems
that work with people?

We’re interested in exploring how human understanding can improve the interaction between people and intelligent machines.

Get in touch
Discuss your project

Prepares a draft in your email app. Nothing is sent by this website.
Privacy information