Recycle Your Old Cell Phone for Cash at New “ecoATM”

ecoATM finds sec­ond homes for three-fourths 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.”

ecoATM received its first NSF sup­port in 2010, then received follow-on fund­ing from Coinstar, Claremont Creek Ventures and Silicon Valley Bank to launch the first kiosks in 2011. The com­pany expanded to the Washington, D.C. met­ro­pol­i­tan area and other areas along the east coast this month, and plans to have more than 300 kiosks deployed by the end of 2012 in shop­ping malls and large stores across the country.

The sys­tem began as a wooden-box pro­to­type that required the pres­ence of an ecoATM rep­re­sen­ta­tive to ensure that users were being hon­est with their trades. While that setup proved con­sumers would be com­fort­able with the device-exchange con­cept, it was lim­ited by the need for human intervention.

The first NSF SBIR grant allowed ecoATM to develop arti­fi­cial intel­li­gence and diag­nos­tics that deliv­ered 97.5 per­cent accu­racy for device recog­ni­tion, remov­ing human over­sight and mak­ing the sys­tem viable for broad use. A follow-on NSF SBIR 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 AI 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.

“The ecoATM project is an extremely inno­v­a­tive way to moti­vate the pub­lic with an incen­tive to ‘do the right thing’ with dis­carded elec­tron­ics, both socially and envi­ron­men­tally,” says Glenn Larsen, the NSF SBIR pro­gram offi­cer over­see­ing the ecoATM grants. “This may change behav­ior from sim­ply dump­ing unwanted elec­tron­ics to a focus on recy­cling, while help­ing put more hi-tech devices in the hands of oth­ers that might not oth­er­wise be able to afford or acquire them.”

The com­pany is part­nered with San Diego-based D&K Engineering to help design and build the kiosks domes­ti­cally, and has expanded from an orig­i­nal work­force of less than 10 in 2010 to a team of more than 150 employ­ees and con­trac­tors today.

Since its found­ing, ecoATM has filed over 20 patents, been awarded seven patents to date, and won numer­ous awards. The com­pany is cur­rently one of three final­ists for a Consumer Electronics Association (CEA) Inaugural Innovation Entrepreneur Award.

“ecoATM meets the required thresh­olds of both con­ve­nience and imme­di­ate finan­cial incen­tive nec­es­sary to inspire mass con­sumer par­tic­i­pa­tion in elec­tron­ics recy­cling,” adds Bowles. “We believe we are the first sys­tem to achieve those thresholds.”

Source: National Science Foundation release