Why Vtr Exists
Vtr is a small independent blog about machines that see, decide and move around inside buildings. It is written for hobbyists building their first indoor rover, students meeting machine learning for the first time, and working engineers who want a plain-spoken second look at a specific problem.
Three subjects that keep running into each other
A robot that drives down a corridor depends on all three of our topic areas at once. It needs eyes, it needs some way of making sense of what those eyes report, and it needs a body that does something sensible with the answer. Vtr treats them as one conversation.
- Indoor autonomous robots looks at vacuum and delivery-style rovers, warehouse and hospital carriers, mapping, docking, door and elevator handling, and why floors, glass and clutter cause so much trouble.
- Artificial intelligence explains planning, learning and decision-making in terms a builder can use, along with the limits that marketing language tends to skip.
- Computer vision covers cameras, depth sensing, object detection, tracking and the lighting problems that break a model that looked perfect on a test set.
How we work
The goal is advice that a reader can test on a bench or a laptop. Technical claims are checked against published papers, open documentation, dataset descriptions and the maker’s own specifications when a device or tool is discussed. When results depend on the room, the sensor or the dataset, the article says so rather than implying a single right answer.
This field moves quickly. Libraries change, models are superseded and products are discontinued, so older pages are revisited and revised when they stop being accurate.
Who benefits most
You do not need a lab or a large budget to follow along. Articles aim to be useful to someone with a single camera and a small robot kit as well as to someone evaluating a fleet.
Spotted an error?
If a fact is wrong, a reference no longer holds or a method failed in your setup, write to [email protected]. Fixes are made on the page itself, in the open.