Dataset "Bicocca (indoor)"

Bicocca location: library The location is set into a pair of building belonging to the Università di Milano-Bicocca, in Milan (Italy). The part of the location explored by the robot during the gathering of the datasets includes one floor of an office building and the corresponding floor of a nearby building. The two buildings are connected by roofed, glass-walled bridges. They sport a wide range of different of architectural features, which are summarized below.

- Corridors (mainly fairly narrow ones) with doors on their sides, mostly leading to offices. Some of the doors are deeply recessed within the walls, at different depths, so corridor boundaries are far from planar.

- A wide hallway where the starting point for the explorations - and the ground truth equipment - were located. This is a very wide space, with features such as columns, staircases, elevators and escalators.

- Smaller hallways of various types.

- Spaces, of different shapes and dimensions, with tables and chairs.

- Two narrow bridges between buildings, already described above.

- A rather wide and architecturally varied library. This occupies most of a floor of one of the explored buildings. Its main features are: open bookcases (many of these of course), desks, PCs and photocopiers, spaces with tables and chairs - both very open (halls) and more constrained such as rooms - and so on.

- Various kinds of doors and passageways, including security doors for the library.

Terrain is very smooth, with occasional small ridges where junctions between different floor sections are present. Floors are horizontal, the only exception being two slanted planes leading to the bridges (necessary due to the different heights of the corresponding floors of the interconnected buildings). One of these ramps is rather wide and has a moderate slope, while the other is narrower and more sloped.

A description of the Bicocca location, complete with pictures, can be found in Rawseeds' Deliverable D1.1. Deliverable D2.2, instead, describes the elements of the Bovisa dataset.