What makes this different from a lab testbed?

The environment is not controlled. Robots operate in venues with customers, staff and furniture that moves, and tasks are interrupted by people who did not read any instructions. Findings from that setting transfer differently than findings from a cleared laboratory floor.

What research areas fit best?

Human-robot interaction in public space, multi-robot task allocation, navigation among unpredictable pedestrians, and the operational side of fleet management. The assignment problem in particular is a live one: priority, distance and battery weighed together under constant re-evaluation.

Is data available for research?

Any data sharing involves customer venues and their guests, so it is a conversation with defined terms rather than a download. We will not publish a dataset here or promise access we have not negotiated with the sites that would be affected.

Can students use the platform?

Tell us the project. There is a mock robot type in the integration layer that allows the system to run without hardware, which is often what a student project actually needs, and it avoids the question of physical access entirely.

We have no published papers to point you at, and we are not going to imply otherwise. If a collaboration produces something publishable, authorship and data terms should be agreed at the start rather than negotiated once results exist.

Government or academic researcher

Hybot for researchers and academics

Hybot runs service robots in working venues, which makes it a field environment rather than a laboratory one. Tasks are interrupted, floors are crowded and people behave unpredictably. For researchers in human-robot interaction, multi-robot coordination or field robotics, that setting is the interesting part, and collaboration is something we will discuss case by case.

Field conditions, not lab conditions

The uncomfortable thing about service robotics is that almost everything works on a cleared floor. The interesting failures need a real room: a child standing still in a doorway, a chair pushed out two inches, a tray that nobody collects for four minutes, a second robot coming the other way.

Hybot's robots run in working venues, so that is the default condition rather than an edge case someone remembered to test.

Where the open problems are

Multi-robot task allocation. Hybot re-scores the whole queue every two seconds, weighing task priority, distance to pickup and remaining battery together rather than in sequence. That is a defensible engineering choice; it is not a solved problem. Weighting under changing fleet size, partial information and battery heterogeneity is exactly the sort of thing that deserves proper study. How task assignment works has the specifics.

Human-robot interaction in public space. Confirmation waypoints mean a robot sometimes waits for a person to acknowledge a handover. How long people take, what makes them ignore it, and what the robot should do when they never come, are behavioural questions with operational consequences.

Navigation among pedestrians who are not participants. Guests did not consent to a study. They are just trying to get to a table.

What we can and cannot offer

We are a company running customer sites, not a research institute, and the honest constraints follow from that:

Data touches customer venues and their guests. Any sharing is a negotiated arrangement with defined terms, not a link on this page.
Physical access depends on the venue's willingness, not only on ours.
We have no published papers, and we are not going to imply a research pedigree we have not earned.

What is straightforwardly available to discuss: the mock robot type, which lets the system run with no hardware attached and is often what a student project actually needs.

If you want to talk

Tell us about the project — the question, the timeframe, what you would need from us. We would rather say "not that, but possibly this" early than have a vague conversation for six weeks.

Where to go next

Multi-robot coordination, explained
How service robots navigate
The AI-in-robotics cluster — including an explicit account of which techniques Hybot does and does not use

Other audiences: developers, operators, general readers.

Frequently asked questions

What makes this different from a lab testbed?: The environment is not controlled. Robots operate in venues with customers, staff and furniture that moves, and tasks are interrupted by people who did not read any instructions. Findings from that setting transfer differently than findings from a cleared laboratory floor.
What research areas fit best?: Human-robot interaction in public space, multi-robot task allocation, navigation among unpredictable pedestrians, and the operational side of fleet management. The assignment problem in particular is a live one: priority, distance and battery weighed together under constant re-evaluation.
Is data available for research?: Any data sharing involves customer venues and their guests, so it is a conversation with defined terms rather than a download. We will not publish a dataset here or promise access we have not negotiated with the sites that would be affected.
Can students use the platform?: Tell us the project. There is a mock robot type in the integration layer that allows the system to run without hardware, which is often what a student project actually needs, and it avoids the question of physical access entirely.
Do you publish?: We have no published papers to point you at, and we are not going to imply otherwise. If a collaboration produces something publishable, authorship and data terms should be agreed at the start rather than negotiated once results exist.

Where this fits

This page is part of Hybot for every audience. If you are working through the topic in order, these are the neighbouring pages.

Build on Hybot: the developer's view
A REST and WebSocket API over a real robot fleet: dispatch to a point of interest, create tasks, stream live state, and integrate with a vendor-agnostic robot layer.
Service robots for business operators
For the person who has to justify the spend: what a Hybot deployment actually involves, what we can measure, and the questions to ask before you sign anything.
What is a service robot, and what do they do?
A plain explanation for the curious: what these robots do in restaurants and hotels, how they find their way, what they are bad at, and where to read more.

The next step from here

One conversation, routed to the person who can answer it rather than to a general inbox.

Start a research conversation