Home / Tech News / Turd vs. tech: Inside the bizarre quest to build a poop-scooping drone

Turd vs. tech: Inside the bizarre quest to build a poop-scooping drone

[ad_1]

People in tech like to dream big. Steve Jobs wanted to take personal computers and transform them from geeky gadgets for a few hobbyists into beautiful consumer-friendly items found in every home. Google’s founders wish for a world in which they can make information universally accessible and useful.

Gerben Lievers wants to use drones and ground-based robots to clean up dog poop.

More: Halt! A new home security system deploys a drone to patrol your property

The 30-year old founder and strategy director of Tinki.nl, a Netherlands-based canine price comparison website (comparing accessory prices; not the dogs themselves), Lievers said that the one thing that upsets him about dog ownership is the 220 million pounds of dog droppings that are not properly disposed of in the Netherlands alone each year.

“We hear a lot of people complaining about dog poo,” he told Digital Trends. “One day I decided to try and think of a way to help solve the problem.”

“On paper it was perfect. Then we built a prototype, and it actually worked!”

While most of us would probably write an upset letter to the newspaper, or our local congressman, Lievers stumbled upon another possible solution.

“I was at an entrepreneur evening, and after the show I got talking to someone who was working in the drone business,” he continued. “We ended up talking about dog poo, and I asked him how he would consider solving it with drones. We agreed to meet up again, and to go to the drawing table to try and come up with a solution for this big problem. On paper it was perfect. Then we built a prototype, and it actually worked!”

The first prototype the pair — Lievers and the brains behind drone company Space53 — came up with represented a two-part attack on dog mess. The aerial component consists of a drone called Watchdog 1, which sports thermal imaging for spotting errant canine coils, based on their warm temperatures compared to the surrounding area. This data is then translated into GPS coordinates and transmitted to a ground-based robot called Patroldog 1, which trundles off to collect its foul-smelling prize.

More: Piqapoo is the Kickstarter accessory that will collect your dog’s poop

“We tried [a version] that was more like a vacuum cleaner to suck it up, and one with arms that could [pick up the poo],” Lievers said. “In the end, the vacuum cleaner was better because of the different consistencies. That way the poo is also completely gone so the entire area is left clean.”

So engineering problem solved, right? Not quite. “The prototype we’re now working on deals with the problem that some dog poo is older and no longer warm,” he noted. “To get around this challenge, we use recognition software to determine that it’s dog poo.”

The team is Developing these tools with some help from researchers at the nearby University of Twente. “When I first contacted them on the phone, they laughed, but then I explained how serious I was,” he said.

As far as recognition tasks go, Lievers said that dog deposits are just another object to detect. “It’s a bit like facial recognition,” he explained, making the faces/faeces comparison we never wanted to hear. “Every face looks different, but they’re recognized as faces. That’s the same with the technology we’re using here with poo. The form and shape may be different with each one, but when you have a database with enough training images it’s possible. There may be some false positives in it, but most will be recognized.”

“To get around this challenge, we use recognition software to determine that it’s dog poo.”

Improving the database of poop-related pictures, he said, could end up relying on crowdsourcing. “It may be that people will need to send us pictures of the poo of their dog in order to train our machines to be better,” he suggested.

So what’s next? “We hope to trial the prototypes in a few neighborhoods where the dog poo is really a big challenge as early as the end of this year, or the beginning of next year,” he explained. “A lot of this depends on how the technology can be developed, however, as well as how willing the government is to take part.”