Increasingly, when you find a trash can in a public place it has at least two openings—one for trash and one for recyclables. But according to Charles Yhap, humans don’t do a very good job of sorting what they throw in those cans.
“Americans typically achieve, when they have a 50-50 chance between two bins, it’s usually 30 percent,” said Yhap.
With that in mind, Yhap launched a company called CleanRobotics with a flagship product called Trashbot.
“You walk up to the machine and you throw an item of waste in, the hatch will close and it will take a reading from the sensor suite we have inside and then… it will make a classification and put it in the right bin,” Yhap said.
The machine reportedly has an 81 percent success rate based on tests done by the company and limited testing in the field.
CleanRobotics Vice President of Engineering Tanner Cook said he is hoping to improve that number through machine learning.
“It’s more or less us digging through trash bags, throwing objects in [the Trashbot] continuously,” cook said.
When the Trashbot is deployed, the camera in the machine will take a picture of any piece of trash that cannot be identified and send the image to a web service where humans will identify the object.
Among the robot’s other sensors is a metal detector, a sonar device and a scale that is used to determine if there is any left over food or drink in a recyclable container.
Clean Robotics currently has one trash bot and is making version two, which will be deployed in the Pittsburgh City County building in July.
The business model calls for CleanRobotics to rent the machines to building managers.
The production version of the trash-sorting machine will have a video display.
“We can put a screen on the back and just run passive ads,” Yhap said. “And in addition, because we can recognize logos on items of waste, we can have targeted ads. So imagine throwing away a can of Coke and Pepsi has paid for advertisements to be displayed whenever that happens.”
The system works best when a user throws each piece of refuse into the sensing bin separately. For example, a drink cup from a fast food restaurant should be deposited in three parts, first the straw, then the lid and finally the cup itself.
Yhap said their research has found the average user has three-and-a-half items to get rid of every time they use a public trash can.
The company is currently looking for seed investors to get it to the next level. Among the ideas still to be explored, according to Cook, is an in-home version that would allow you to drop anything through a hole in your countertop and have it sorted into separate bins.
In this week's Tech Headlines:
- An assistant professor at The University of Pittsburgh has been listed by The Journal of Materials Chemistry A, published by the Royal Society of Chemistry, as a "rising star"of materials chemistry research. John Keith of the Swanson School of Engineering came to the attention of the list makers for his computational work involving coating on tin that could improve the performance of tin electrocatalysts for CO2 reduction.
- The White House is urging technology CEOs to pitch in on President Donald Trump's effort to modernize government. Officials from Apple and Google were among those at a gathering in Washington Monday where the president said he wants to transform the federal government's technology in a way that will deliver better services to citizens and stronger protection from cyberattacks. Trump says such an effort could produce up to $1 trillion in savings over 10 years. The event included the formation of working groups on issues such as technology infrastructure, cyber security and visas for foreign workers.
The Associated Press contributed to this report.