From the road to the care home
Machines that understand people.
We spent a decade teaching vehicles to see, understand and anticipate people. We are bringing that to care, together with the people who do the caring.
Anonymous by design. It keeps a figure, never a picture.
In care homes
One small camera. A model of the room. Help where the shift needs it.
The room is measured once. From then on the system follows what happens in it in metres and degrees, as a figure and never as a picture. Pick a situation and drag the timeline.
How it works
It sees a picture, keeps a figure, and sends a sentence.
Each step keeps less than the one before. That is how it understands a room without watching the people in it.
An ordinary camera on the wall. Each frame is read on a small computer in the room and discarded.
The room in three dimensions, built once. The person as a figure placed in it: where, how upright, how fast.
What happened, since when, and how it knows. Written so that a nurse can check it at a glance.
Figures reported by German care operators, surveys and the 2026 care reform draft. Our own numbers will come from pilot rooms.
Privacy
Nobody is watched. Nobody is recorded.
A camera in a room where someone lives is a serious thing. So the system is built to need the picture for a fraction of a second, and then to work without it.
No video, anywhere
Frames are read in the room and deleted at once. There is no recording to watch, leak or request.
A figure without a face
What remains has no face, no skin, no clothing and no name. It is a posture in a room, in metres and degrees.
It stays in the building
The computing happens on a small device in the room. Only the sentence for the staff leaves it.
Silent unless needed
While nothing is wrong, nothing is sent and nobody looks. Staff see a message only when someone needs them.
Decided with the people it concerns
What is kept, for how long and who may see it is agreed with residents, relatives and staff before anything is switched on. It can be switched off in any room at any time.
The engine underneath
Four things we have spent years getting right.
Everything above runs on the same four abilities. None of them is specific to care.
Reading a room
From a single camera to a room in metres: floor, furniture, free space, the way to the door. It knows where a person can walk and where a machine can pass.
Reading a person
Every body is different. Our body model describes the individual: build, proportions, posture, and how they move. It can be measured, and followed over months.
Knowing what happens next
On the road a system has two seconds to decide whether someone will step out. Anticipating people, and acting safely around them, is what our team has worked on for more than ten years.
Building the machine
We build the hardware too. Our research vehicle carries its own sensors and computers through Munich traffic. The same engineering goes into what we put in a room.
Beyond care
Wherever people and machines share a space.
Care is where we start, because it is where understanding people matters most. The engine does not change when the room does.
Therapy and sport
Technique, load and progress, measured without markers or wearables.
Workplaces
People and machines on the same floor, each aware of the other.
Public space
Crossings, platforms and entrances that notice who needs more time.
Mobility and robotics
Vehicles and robots that read intent, by day and at night.
Safety and security
Knowing where people are when conditions are hardest: darkness, smoke, crowds, large sites.
People
A small team from automated-driving research in Munich.
We have built perception and planning systems for vehicles, together and for years. Sensai is where we take that work next.
Prof. Dr. Fabian Flohr
Chief Scientific Officer
Mark Azer
Chief Technology Officer
PhD researcher
Katharina Winter
Co-founder
PhD researcher
Mohan Ramesh
Chief Executive Officer
PhD researcher
Work with us
Help us build this the way care needs it.
We are looking for care organisations who want a say before it is finished. There is nothing to buy.
- Now
A 30-minute conversation
Tell us what a shift is really like, and where this page is wrong.
- Next
A visit
We come and watch, without equipment.
- Then
A letter of interest
Non-binding. It tells our funders that care wants this built.
- Later
Pilot rooms
In rooms you choose, on terms you set with residents and staff.