The Need: Less than 10% of plastics are recycled in the United States, according to the EPA, and major consumer brands produce much of the material that winds up in landfills. In order for materials to flow successfully through the recycling stream, they must be accurately sorted and separated at materials recovery facilities (MRFs) to be valuable to a buyer who wants to turn them into new packaging.
However, there are hundreds of recycling facilities throughout the U.S., many with outdated equipment and/or relying on workers to manually sort through and separate materials. When facilities want to sort for uniformly shaped items—such as beverage bottles labeled #1 and bottles that store milk or laundry detergent labeled #2—their common shape makes it easier to find them. But in order to sort non-uniformly shaped items—such as containers using polypropylene #5 plastic, one of the most widely used materials in packaging for consumer goods and the material used in yogurt containers, butter tubs, and a whole host of other everyday products, including K-Cup pods—it gets more complex. If MRFs invest and upgrade, they will be able to create a greater supply of recycled plastic from a wider range of products, reducing the need for new plastic.
The Solution: The capacity to process more recyclable material more accurately, especially with advanced technology, does exist. The AI platform that guides AMP’s robotic sorting systems can differentiate objects found in the waste stream by color, size, shape, opacity, brand, and more, contextualizing and storing information about each item it perceives. AI and machine learning enable the robotic sorting of material as granular as a type of plastic at a pick rate of upwards of 80 items per minute—more than two times as fast as human sorters, and with greater accuracy and consistency. Testing has validated that this process produces a higher quantity and quality of plastic—some of which hasn’t been as commonly recycled in the past—that suppliers of recycled resin can sell back to manufacturers to create new materials versus using virgin plastic. As an example, AMP partnered with beverage company Keurig Dr Pepper to test and refine the capability of AMP’s robotics systems to properly identify and sort the company’s polypropylene plastic K-Cup pods within recycling facilities. Polypropylene is used in an array of food and non-food packaging and is a valuable commodity to be continually remanufactured, especially as more and more consumer packaged goods companies are making commitments to incorporate more recycled content into their packaging.
What makes it particularly smart or circular? AMP’s AI platform becomes smarter and more effective over time as the company deploys more robots; AMP can add limitless subcategories of brand-level material to meet market demand and distribute the functionality to identify and sort it across its fleet. What this means is that the solution is replicable with many types of packaging. The results also validate the importance of collaborative efforts to strengthen recycling, and how technology can accelerate progress. In this case, MRFs, where recyclables are processed also need to be connected to the initiative to move the needle.
Results, Benefits, and Outcomes to Date: Collaboration with producers demonstrates the strength, precision, and flexibility of AI to learn new packaging to the specificity of a manufacturer or brand. It enables manufacturers to directly influence what’s recoverable in recycling facilities and take advantage of the ability to capture more of their specific packaging. Recyclers across the world with AMP’s robotics systems gain these sorting capabilities, and because the company’s AI is continuously learning, AMP is only improving the recovery of such materials over time.
AMP’s AI platform encompasses the largest known real-world dataset of recyclable materials for machine learning. The company can classify more than 100 different categories and characteristics of recyclables across single-stream recycling, e-scrap, and construction and demolition debris. AMP has extended its object recognition run rate to more than 10 billion items annually thanks to its approximately 200 deployments across North America, Europe, and Japan. In the U.S. alone, AMP has robotic systems deployed in more than 25 states, recovering all types of plastics, along with paper, metals, and more.
Which of the Pact’s 4 Targets does your work help achieve? Target 2, 3
How are you communicating your success? We’re employing a variety of tactics to promote the success and benefits of our technology, engaging with the industry through media relations, webinars, and speaking opportunities.
What’s Next? We’re continuing our market expansion with brand owners and producers given the level of customization to specific packaging our AI can provide. To further increase recycling rates, we’ve launched an automated facility design for advanced secondary sortation, allowing us to aggregate small volumes of difficult-to-recycle mixed plastics, paper, and metals sourced from residue supplied by primary MRFs. The ability to recover recyclables from residual waste streams represent a major opportunity to increase national recycling rates, helping meet the growing demand for recycled content and protecting our environment.
Quote: “Intelligent plastics sortation, powered by AI, robotics, and advanced data analytics, can have cross-value chain impact and direct benefits to plastic waste generators, sorting facilities, recyclers, and consumer packaged goods companies. AI-guided sortation ensures a higher-quality end product that isn’t contaminated by other materials, and a larger volume of recycled material. Collaboration across the recycling value chain will turn product and packaging waste back into the inputs for future manufacturing while growing and strengthening our recycling system.”—Rob Writz, director of commercial partnerships
Company/Organization Information: AMP was founded in 2014. The company is headquartered in Louisville, Colorado, and has approximately 250 full-time employees.