ROAMS robot works cheaply and effectively and can be very useful in 3D configurator projects.
Three-dimensional mapping makes sense not only for military purposes, but also in architecture. But so far such a conventional 3D lidar system, which consists of several lasers that measure in different directions, cost over 100.000 Dollars. Now researchers from New Jersey have developed the robot ROAMS, which provides adequate images for architectural or military mapping purposes – and only costs around 15.000 Dollar.
ROAMS stands for Remotely Operated Autonomous Mapping System“. The mobile machine uses various commercially available techniques to create three-dimensional maps of the environment. The robot was developed by the Stevens Institute of Technology in Hoboken, New Jersey, on behalf of the US-Army.
The system is mainly based on a radar-like method for optical distance and speed measurement. Lidar uses a laser that is deflected by a fast rotating mirror and measures the light reflected from the surrounding surfaces and objects. Lidar has long been used by autonomous aircraft used in space travel.
Instead of the expensive 3D lidar system, which measures in different directions with several lasers and costs over 100.000 Dollar., the Stevens researchers use the much cheaper 2D lidar system for about 6.000 Dollar. The 2D lidar system was mounted on a swivelling, rotating scaffold. Although the resolution of the images is lower, it is sufficient for most purposes.
In about 30 seconds, the system scans a 160-meter wide area. A camera, which was also mounted on the framework, complements the 3D map with missing colour information. The researchers have succeeded in keeping the resolution of the objects always adequate, irrespective of the distance. A human operator follows the robot at a maximum distance of 1.5 kilometres in a second vehicle to control ROAMs.
According to Kishore Pochiraju, professor and director of the university`s Design and Manufacturing Institute, the ultimate goal of the project is that the robots can work completely autonomously. The machines are then set down in the target area and independently create a complete map in 3D. The next challenge in autonomous mapping is to detect obstacles, recognize and exchange data between multiple robots.