This website uses cookies primarily for visitor analytics. Certain pages will ask you to fill in contact details to receive additional information. On these pages you have the option of having the site log your details for future visits. Indicating you want the site to remember your details will place a cookie on your device. To view our full cookie policy, please click here. You can also view it at any time by going to our Contact Us page.

'ATM' takes old phones for cash, charitable donation or recycling

18 September 2012

With mobile devices, yesterday's hot technology can quickly become obsolete, but they still have value as an affordable alternative, or even as spare parts.

With support from the US National Science Foundation (NSF), ecoATM of San Diego, California, has developed a novel automated system that lets consumers trade in those devices for reimbursement or recycling.

Using artificial intelligence, ecoATM kiosks can differentiate varied consumer electronics products and determine a market value. If the value is acceptable, users have the option of receiving cash or store credit for their trade, or donating all or part of the compensation to one of several charities.

ecoATM finds second homes for 75 percent of the phones it collects, sending the remaining ones to environmentally responsible recycling channels to reclaim any  rare earth elements and keep toxic components from landfills. ecoATM is certified to the eWaste environmental standards of Responsible Recycling (R2) and ISO 14001.

"The basic technologies of machine vision, artificial intelligence and robotics that we use have existed for many years, but none have been applied to the particular problem of consumer recycling," says ecoATM co-founder and NSF principal investigator Mark Bowles. "But we've done much more than just apply existing technology to an old problem - we developed significant innovations for each of those basic elements to make the system commercially viable."

The initial development funding delivered 97.5 percent accuracy for device recognition, removing human oversight and making the system viable for broad use. A follow-on grant is helping ecoATM close that final 2.5 percent accuracy gap.

According to Bowles, traditional machine vision generally relies on pattern matching, pairing a new image to a known one. Pattern matching is a binary approach that cannot handle the complexity of ecoATM's evaluation process, which includes eight separate grades based on a device's level of damage.

"We are now able to tell the difference between cracked glass on a phone, which is an inexpensive fix, versus a broken display or bleeding pixels, which is generally fatal for the device," says Bowles. "We were warned by leading machine-vision experts that solving the inspecting/grading problem-with an infinite variety of possible flaws was an impossible problem to solve. Yet with our NSF support, we solved it through several years of research and development, trial and error, use of artificial intelligence and neural network techniques."

The company's databases are now trained with images of more than 4,000 devices, and when an identification mistake occurs, the system learns from that mistake.

When a user places their device into an ecoATM kiosk, the artifical intelligence system conducts a visual inspection, identifies the device model and then robotically provides one of 23 possible connector cables for linking it to the ecoATM network (the company warns consumers to erase all personal data before recycling).

Using proprietary algorithms, the system then determines a value for the device based on the company's real-time, worldwide, pre-auction system. Within that system, a broad network of buyers have already bid in advance on the 4,000 different models in eight possible grades, so the kiosk can immediately provide compensation.

A number of robotic elements enable the kiosk to safely collect, evaluate and then store each device in a process that only takes a few minutes.

Print this page | E-mail this page