The platform draws from every aspect of robotics, incorporating state-of-the-art computer vision, path planning, and classification, yielding an exceptionally low cost, fast, and extensible system. In particular, the system does not rely on any fixed characteristics of the waste, and is instead trained by showing it thousands of images of its target waste. The system naturally learns to be robust to problems like dust, while being able to identify all classes of waste, from Construction and Demolition (C\&D) to Municipal Solid Waste (MSW). The use of cutting-edge techniques enables the system to use inexpensive, commodity robotic hardware and sensors and compensate for inaccuracies with intelligence. The results include a reduction in capital costs and an increase in plant efficiency, changing the economics of sorting and allowing for a large increase in diversion volume.
From the beginning of our design, we have focused on attacking the fundamental economics of the recycling process. Key to the value proposition for the recycling facilities are the switching, capital, and operating costs of the equipment. We have therefore designed our system to tackle and ameliorate these three issues. In particular, our system can drop into existing recycling lines with virtually no change to existing operations. The installation process will be quick – we believe it can be completed over a weekend – and our upfront cost will be a fraction of existing recycling equipment, in many cases a fifth or less. The result is a low-risk, high-reward path of migration for these facilities.
Login or become a member to access more detailed company information
Already a Member?
Not a Member Yet?