Roomba
The Robot Vacuum Cleaner
iRobot Manufacturer
“iRobot AWARE Intelligence System” has been deployed in Roomba which gives it the capability to navigate, move, manipulate and have artificial intelligence.
,
Howstuffswork
When Roomba knocks into something or detects an obstacle, it would “perform (and repeat) the sequential actions of backing up, rotating and moving forward until it finds a clear path.
SELF-NAVIGATION
IN REAL ACTION
Once Roomba is powered up for cleaning, it will first send out infrared signals to estimate the size of the room. This is not for Roomba to do any mapping but rather to provide Roomba with a very “rough picture” of the room. Using this information, Roomba would be able to estimate the amount of time it needs for cleaning.
When Roomba knocks into something or detects an obstacle, it would “perform (and repeat) the sequential actions of backing up, rotating and moving forward until it finds a clear path. It also believes that it has reached the perimeter of the room. It then cleans along the ‘perimeter’ until it hits another obstacle”.
In fact, a closer examination of the “iRobot AWARE Intelligence System” reveals that the manufacturer iRobot is using a reactive architecture, also known as “subsumption architecture”.Subsumption architecture is an artificial intelligence (AI) concept connected with behaviour-based robotics and was first developed by Rodney Brooks in 1986. It is a “design paradigm for mobile robot control that emphasises re-use of modules, decentralisation and concurrent, communicating processes”
It opposed from the traditional AI that made used of “Sense-Plan-Act” paradigm. In this old paradigm, the robot would first sense its environment. After getting a sense of it, it would plan its next action. Finally, it would act, in accordance to the plan that it has designed, using available resources installed on the robot. However, this approach requires a significant amount of computational power and could only run very slowly within a controlled lab setting. Any slight change in the environment would force the robot to stop and calculate on the changes before it could take any further action.
In the new paradigm, there is “no specific place in the architecture where cognition is done. Instead, cognition emerges as the result of different layers of the architecture acting together. This idea is called emergent behaviour”. In other words, the robot no longer works in a sequential arrangement and planning can take places anytime. The robot will react accordingly to the situation, making it “smarter”. Under this new paradigm, it emphasises a lot on learning and adapting to the environment.