Large sprawling airports in unfamiliar cities can be difficult places to find your way around. While it helps if someone can point you in the right direction, what's best is if they can actually take you where you want to go. Well, that's where Spencer comes in. "He" is a multi-lingual robot that's designed to guide travellers through airports.
The project was initiated by Dutch airline KLM, as a way of reducing the costs incurred when passengers miss their flights due to getting lost. Researchers and businesses in five European countries are now involved in the development of the robot.
Spencer is loaded with map data of the airport in question, and is able to ascertain its current location via a combination of sensors such as accelerometers and range-finding lasers. It uses the latter both to avoid objects in its path, and to gauge its proximity to known landmarks – by comparing detected landmarks to ones on its maps, it's able to confirm that it's where it should be.
While quickly-moving obstacles (i.e: people) don't pose much in the way of problems for it, a larger challenge is "temporarily permanent" objects such as parked luggage carts. These can't just be discounted as fleeting obstacles to avoid, yet they also aren't permanent structures that help the robot to orient itself within the airport.
It's an issue that is still being addressed, along with the task of making sure that the robot is able to understand human behaviour and act accordingly. This would include things like knowing to go around groups of people instead of trying to cut through them, and keeping track of passengers to make sure that they're still following it.
Plans call for Spencer to be assessed over a one-week period starting on Nov. 30th in Amsterdam's Schiphol international airport, with a larger test scheduled to take place in March. The technology could ultimately be applied to all robots that are required to interact with humans, including robots that care for seniors in their homes.
Spencer can be seen in action, in the video below.