The Problem With Robots
An Original Online Book By Frank Fogg
One high-tech growth industry of the 21st century undoubtedly will be robots. From a plethora of robot pets and bugs, to humanoid walkers, to garbage can style utility 'bots, they just seem to be popping up everywhere nowadays. Yet, they still don't seem to do very many useful things at home.
One problem holding back the advent of robots from general consumer utility use is their concept and design. Most low cost home robots perform few useful functions other than entertainment. Sure, they may be hooked up via a wireless network to read e-mail, or perform security-monitoring functions, but there are cheaper and more efficient ways to accomplish these goals. How about one that picks up all the dirty clothes we throw around, loads the dishwasher, or actually gets a snack from the refrigerator without human intervention?
The problem I see with design is that the robots are expected to manipulate tools and appliances that were initially designed for humans. Let's just take a refrigerator for an example. The door must be opened, and the height and design of the mechanism to accomplish this is just right for a seven year old, but probably not even close for a seven year old robot. If the droid can successfully locate the refrigerator and get the door open, now it has to find the item you want. This is going to be fairly difficult if your refrigerator is organized anything like mine.
Another common household chore is cleaning dishes. Just having the robot successfully load the dishwasher, press any number of buttons or knobs to get it going, and getting the proper amount of detergent in it to do the job might take years of programming. Then, once the dishes are clean, they have to be unloaded and placed in a cabinet that is usually at or above the eye level of most adults.
I'm sure you can already see that any number of examples could be proposed that would equal these I've listed above. I believe that with a simple change in several concepts, robots could perform a number of common household duties easily and effectively. From my perspective, the household robot function should focus on what it primary does easily and best - provide a mobile platform for performing work. So how do we get from the present to the future within these constraints?
Uniform Appliance Design and Control
In addition to having controls that humans can operate, appliances need to have the ability to function without mechanical manipulation. This can be readily accomplished using today's technologies with infrared and wireless networking. Instead of mechanically opening the refrigerator door, the robot simply issues a data command and the door opens on its own power. This interface could be standardized in any number of ways, from self-programming security codes to user selectable switches, and be available in infrared technology at a minimum of cost on every appliance including stoves, dishwashers, refrigerators, toasters, microwaves, washers, dryers and coffee makers.
Additionally, the "business end" of the major appliance needs to be at a relatively uniform height so that a mechanical robot arm can easily load or unload food, laundry, and garbage without a large vertical reach requirement.
Why spend years trying to teach a robot every possible movement necessary to load and unload a refrigerator or dishwasher? Think about this. Using a uniform set of glasses, dishes, silverware and pots, the robot now must merely transport to and from the dishwasher. Once it arrives at the dishwasher, it has several slots to insert dirty dishes, one for each type (using a uniform size and design makes optical identification easier.) Once the dish is placed in the slot, an optical sensor inside the washer detects that a new item has been loaded, and an internal specialized loader takes over from there. When the cycle is completed, it calls the robot to take the now clean dishes from the appropriate slots and move them to the proper bin in our robotic friendly shelving for their next use. Now all we need to do is ask the robot to set four places at the table - it can select four services from their proper bin and place them at the dinner table, I think pretty much autonomously.
Shelving and Storage
While this may not seem necessary, having uniform height for storage of cleaning supplies, canned goods and other non-refrigerated households in a comfortable area for the robot to reach increases the amount of weight that can be safely transported and stored. Perhaps storage bins could be used with bar coded labels so that the infrared sensor of the robot could "read" what was in the bin, even if it was not originally placed there by the robot, and without requiring any other previous programming. Interfaced to a wireless network with DSL capability, a robot with optical sensors could even read the UPC from a can or bottle and access a web database to identify what it is holding, attempting to retrieve, or trying to organize.
Most robots seem pretty dumb about how to maneuver about. There seems to be several popular concepts regarding this issue. One is to place wires, strips, or painted lines that the robot must follow - not very versatile or adaptable. Another is the bump and avoid technique- useful but not predictable. Another method involves training, not bad until something goes haywire. Finally, there's the remote control. If you have to spend all night behind the joystick, how much free time are you actually gaining?
I think another cheap solution is available to solve this problem. What I propose is to mount infrared emitters in the top corners of each room. These would be digitally modulated so as to provide a unique code, and would be stationary from the robot's point of view. Mounted high in the room as they are, they would be visible from most points in a common room, and the robot could check it's position in the building based on the computing the geometric horizontal and/or vertical angles of each emitter from the robot's current location. This would assist in verifying its location based on pre-programmed and learned routes throughout a building at a minimum of expense. Another advantage is that should the robot lose power or be picked up and moved, it could immediately re-orient itself from known codes and geometries.
As I see it, the most important necessity for the usefulness of home robots is to have standardized items, sizes, controls and navigational guideposts for the robot to use. As many things as possible are at or near the same height when mechanical interface from the robot to the chore is accomplished. Uniform shelf placement and markings assist the robot in properly selecting the item desired. Uniform wireless controls allow it to operate many common appliances. The two most successful applications for home utility robots today are the vacuum cleaner and lawn mowing robots. They succeed because they take advantage of the robot's most powerful feature - mobility. The manufacturers don't attempt to design the robot to use a tool designed for humans, but instead adapt the tool in a way that it can be most effectively used by the robot. I firmly believe these principles will guide future robot development in successful household applications.
Press control + D now to bookmark this site.