Robotics & Physical AI

Machines that perceive, decide and act

Motors and actuators, sensors, embedded computing, computer vision and machine learning: we build robots and intelligent machines that work in the physical world.

Engineer in a cleanroom suit using a tablet next to a large white industrial robot arm

Robotics

Mechatronics with a brain

Robotics brings every discipline together: mechanics, power electronics, motor control, sensors, real-time software and perception. We integrate them into prototypes that can be tested and improved quickly.

For perception and decision, we use ROS, OpenCV and machine-learning tools such as Keras, on embedded platforms or FPGA when real-time processing is required.

Physical AI

When artificial intelligence gets a body

Physical AI is artificial intelligence that perceives the real world through sensors and acts on it through motors and actuators. It powers robots, autonomous machines, smart instruments and assistive devices — and it must be fast, robust and safe.

Perception

Cameras, depth and distance sensors, IMUs and force sensors interpreted by vision and AI models.

Sensor fusion

Combining several sensors to estimate position, motion and context reliably.

Decision & control

Embedded models and control loops that turn perception into precise, real-time action.

Learning & simulation

Data collection, training, simulation and tests on the real machine to close the gap between lab and field.

Human–robot interaction

Collaborative behaviour, intuitive interfaces and assistance to the operator.

Safety & reliability

Safety functions, fail-safe behaviour and validation, considered from the design stage.

Capabilities

What we design

Motion control

Electric motors, drivers, actuators and closed-loop control.

Sensing

Accelerometers, encoders, distance sensors and cameras.

Computer vision

Image processing and recognition with OpenCV, on embedded targets.

AI & machine learning

Model prototyping with Keras and Python, deployment on edge hardware.

ROS integration

Robot software architecture, nodes, simulation and tooling.

Safety

Safety functions and procedures considered from the design stage.

How we work

From need to validated design

Concept

Tasks, environment, safety and performance targets.

Architecture

Mechatronic architecture, actuators, sensors, computing platform.

Demonstrator

A first robot that moves and senses, to validate the concept.

Iteration

Tests, measurements and improvements towards the next version.

3D printer print head and motion system in the middle of a print

Project examples

Related projects

  • Engineer in a cleanroom suit using a tablet next to a large white industrial robot arm

    Robotics

    Robot prototype: motion, sensing and computer vision

    Development of a robotic prototype combining motor control, sensors, embedded computing and a vision pipeline, from the electronic boards to the control software.

    A working demonstrator used to validate the concept and prepare the next design iteration.

    • ROS
    • OpenCV
    • Electric motors
    • Actuators
    • Raspberry Pi
    • C++
    • Python
  • FPGA system-on-module carrier board with a cooling fan, high-density connectors and a PCIe edge connector

    AI · Embedded vision

    AI inference on an FPGA system-on-chip

    Embedded and digital electronics design to run an AI processing chain on a Xilinx Zynq SoC, sharing the work between the ARM cores and the programmable logic.

    Real-time processing on a compact, low-power platform, without a cloud connection.

    • Xilinx Zynq
    • Vivado
    • VHDL
    • ARM
    • Keras
    • Python
  • Public playground with slides and a climbing frame on a loose-fill gravel surface, surrounded by trees

    SportTech · Playground safety

    Head-impact test equipment for playground safety (EN 1177)

    Technical expertise and improvement of the electronics of test and measurement equipment used to check the safety of playgrounds and sports facilities, in particular instrumented impact heads used to measure the critical fall height according to EN 1177.

    Detailed product specifications (50+ pages) written for HIC (Head Injury Criterion) test equipment compliant with EN 1177:2008, and test procedures defined and corrected.

    • EN 1177:2008
    • HIC
    • Accelerometers
    • ADC
    • Signal processing
    • Test procedures

FAQ

Good to know

Do you design the mechanics too?

We define the mechatronic architecture and work with mechanical partners and 3D printing for parts and structures.

Can AI run without the cloud?

Yes. We deploy models on embedded processors or FPGA to process data locally, in real time and without sending it outside.

What is the difference between AI and Physical AI?

Classic AI works on data and text. Physical AI closes the loop with the real world: it senses, decides and moves, under real-time, energy and safety constraints. That is why it needs electronics, embedded software and robotics expertise as much as data science.

Let’s talk about your project

Describe your need in a few lines: we reply with a first technical analysis and a proposal.