garbage-sorting robot gets its hands dirty
Industrial robots are usually good at completing precise tasks in controlled environments, such as assembling cars.
More chaotic and dangerous tasks fall on humans-for example, clean up piles of waste and look for precious raw materials that can be recycled.
But maybe it won\'t last long. Led by Tuomas J.
Lukka, a research team at ZenRobotics in Helsinki, Finland, wants robots to take over waste recycling.
The company\'s recycling robots use data from visual sensors, metal detectors, weight measurements, and robot arm tactile feedback to pick and classify possible garbage.
Through repeated trials, its machine learning software has been taught to identify about a dozen materials, including different plastics.
As it moves along the conveyor belt, it can remove concrete, metals and wood from the waste logistics.
\"I \'ve never heard of anyone really trying to do this in such an unstructured environment,\" said Edwin Olsen, a computer scientist at Ann Arbor University of Michigan.
For more obscure types of waste, such as a piece of plywood with nails, the robot uses a spectrometer to identify objects based on the unique light patterns they reflect.
This means that the robot can distinguish the type of waste according to its color and put it in the appropriate bin.
Since the launch of the testing phase in February, the robot has learned to correctly identify half of the building debris it was fed.
This is far from perfect.
But according to the Building Materials Recycling Association, construction waste accounts for half of all landfill materials in the United States.
Recycling a small part of it will mean significant savings in resources and landfill fees.
Although the robot is designed for construction waste, one day it will also sort out household waste, says Luke.