Practical Service Robotics

November 2010
by Bryan Bergeron

If you’re a robotics enthusiast, you’ve probably at least considered picking up one of the robot vacuum cleaners. I have. But I haven’t purchased one because I haven’t found a robot vacuum cleaner – the most popular form of service robot on the planet – that can replace a handheld vac. I also suspect that many consumers share my sentiments, as well.

My requirements for a home robot vacuum cleaner are simple enough – the vac simply has to work like the litter cleanup robots on Star Wars. If you recall, the small, black robots scurried about in search of litter. When a robot spotted something in the hallway, it would suck it up and then scurry away in search of the next object. Sounds simple enough, doesn’t it?

I want a robot that will recognize that I dropped a bowl of popcorn on the carpet, head to the sight of the disaster, suck up the popcorn, and then return to the charging station. As a bonus, if the vac happened to scoop up a coin, transistor, or earring, I’d like the valuable stored in a separate bin in the vacuum’s body.

However, in my recent survey of what’s available on the market, nothing came close to my ideal. The old standby — the iRobot Roomba — can swirl around a room for an hour before happening upon a pile of crumbs.

The more expensive robot vacs such as the Neato XV-11 and Samsung Navibot excel at navigation – meaning they can avoid obstacles and almost guarantee full floor coverage. However, even the $650 Navibot can’t spot a fresh spill, head directly to the area, clean it up, and then move out of sight. No, for quick spills, robot vacs still can’t hold their own against a $35 handheld vac.

So, what’s the holdup? Why can’t one of the big robotics manufacturers come up with software and a hardware platform that can effectively replicate a human-directed hand vac? Think about how you’d go about devising sensors and the control program for such a vac. One option would be to scan the rooms with ceiling-mounted cameras. You’d have to program the system to ignore people, pets, and toys. There’s also the issue of changes in lighting with time of day. Another potential problem is how to recognize – and ignore – spilled liquids. You don’t want your robot vac making a mess by rolling in spilled chocolate milk or red wine.

You might also consider a sensor that’s specific for organic material. Perhaps a sensor that’s sensitive to methane – produced by rotting organic material – could help a robot vac hone in on day old crumbs.

In the Star Wars movie, the dog-like litter robot worked on shiny, highly waxed floors. Perhaps a similar environment would make the work of a modern robot vac easier. Consider a laser radar that’s fired just above and parallel to the floor. Anything between the robot and floorboard that’s only a few millimeters in height or width is suspect and requires closer inspection.

Unfortunately, my hypothetical solutions to making a robot vac more useful aren’t necessarily economically feasible. What would you do to bridge the functionality gap? What are your must-haves for a robotic vac? Share your list and, if you have them, photos of your creation with SERVO readers. SV