In order for autonomous delivery robots to find your front door, an entire mapping system takes place behind the scenes. Usually, this requires an area to be mapped out in advance. Then the robots are programmed to follow a specific route.
This is becoming more and more common in places such as universities or specific office locations, for example.
Pushing out these types of door to door deliveries on a larger scale, however, creates security and privacy issues. A team of MIT engineers may have found a solution to this with their new mapping system.
MIT's map of clues
What are these clues? They are, in fact, points on a map such as front doors, or garage doors. With MIT's new system, autonomous robots are trained to understand that a driveway usually leads to a front door.
Using a light-to-dark coloring system, the robots learn that the lighter the color, the nearer they are to their final destination. If, for instance, a point on the map is dark, the robot understands that it is far away from its ultimate destination. Pretty nifty.
"Even if a robot is delivering a package to an environment it's never been to, there might be clues that will be the same as other places it's seen," said Michael Everett, a graduate student in MIT's Department of Mechanical Engineering. "So the world may be laid out a little differently, but there's probably some things in common."
It's not just new and shiny. It's new and improved.
The MIT team's technique utilizes pre-existing algorithms. They look at data and labels it with clues — such as a sidewalk or a driveway.
By using these algorithms, the team created a map as their robots drove along. The main difference now is that the robots can determine their own path to the final destination, thanks to these built-in clues.
The mapping system will be showcased this week in Macau at the International Conference on Intelligent Robots and Systems.