Recycling Robots: How They Work & Are They Worth the Investment?

Robot Tech

Recycling robots observed by two workers

Recycling Robots: How They Work & Are They Worth the Investment?

Mark Fairchild
Freelance Technical Copywriter
Fairchild Copywriting

Recycling robots transform the economics of Material Recovery Facilities by automating sortation. Explore how, and discover ways to source the robots.

In this article, we explore the subject of recycling robots through the following topics:

The Importance of Recycling

Infographic illustrating a comparison between the linear economy (take-make-waste) and the circular economy (make, use, recycle.)


In 2014, the European Union introduced its circular economy vision, although the idea had been proposed earlier. To understand this concept better, let’s consider the linear economy first. It’s a one-way process. Raw materials are taken, usually from the ground. These resources are then made into products. When the products have been used, they are considered waste. The waste is disposed of by putting it into a landfill or dumping it in the ocean.

In contrast, a circular economy considers a used product as a resource. The materials used in making the product can be recovered and used as raw materials for making new products. Hence the idea of the circle.

Industrial society, until recent decades, has followed the linear model. Now, the circular economy is the goal. The expected benefits include:

  • Reducing the cost of many raw materials
  • Reducing waste and pollution
  • Regenerating nature

The Barrier to Waste Recycling

The biggest reason only a small percentage of waste material is recycled is the economics of sorting the waste. It is often cheaper for a municipality to put garbage into a landfill than to recycle it. According to an article from CitizenSustainable, to recycle one ton of Municipal Solid Waste (MSW) in New York City costs $686, whereas the landfill cost is only $126. In Boston, it’s $160 per ton to recycle and $80 per ton to put the waste into a landfill. The recycling cost is at least partially offset by selling the sorted and recycled output.

Recycling is so much more expensive than putting it in a landfill because although the paper, plastic, glass, and metal contained in waste products is valuable, it must be sorted. Separation is the key to purity. And the cost of sortation is high.

One early solution to the sortation problem was to ask consumers to sort by providing different bins for each of the distinct types of material. However, sorting your garbage into multiple bins is a bother. As a result, many people wouldn’t recycle at all. Other people were careless and put things in the wrong container, making the purity of the material poor. In addition, this “multi-stream” approach often meant different trucks to pick up the waste – a separate truck for each kind of material. That dramatically increased the cost of collection.

Most cities have opted for the “single-stream” recycling approach to encourage recycling. A single stream means one recycle bin for all kinds of recyclable material. However, the result is the sortation must be done by the Material Recovery Facilities (MRFs).

Bulk processing and sorting methods exist for some kinds of material. For example, magnets can separate metals from the rest of the waste. But magnets do not attract aluminum. And there are many materials for which easy bulk sortation is not possible. 

Waste Sorting Robot Solutions

With improvements in artificial intelligence-aided vision, it is now possible to use robots to do the sortation. The AI-vision system detects the kind of material of which an object is made. A robot arm picks the items from a conveyor belt and deposits them into different chutes, depending on the kind of material.

Best Waste Fractions to be Sorted with a Robot

Certain products are more recyclable and suitable for robotic sorting than others.

Construction Waste

Metal, wood, plastic, stone and concrete are well suited to recycling and robotic sorting.

According to the U.S. Environmental Protection Agency (EPA), only 34% of metal in U.S. municipal waste facilities was recycled. Yet, most kinds of metal can be reused again and again without losing their essential qualities.

Wood can be shredded and reused as particle-board if it is not too contaminated with non-wood substances like paint and preservatives.

Manufacturers can use recycled plastic to create a wide variety of construction materials, including:

  • roofing tiles
  • stronger concrete
  • indoor insulation
  • structural lumber
  • PVC Windows
  • Bricks
  • Fences
  • Floor tiles
  • Carpeting
  • Ceiling tiles

Electronic Waste

In 2021, an estimated 57.4 million tons of electronic waste was created worldwide, according to an article in the WEEE Forum. The EPA estimates that only 15-20% of e-waste is recycled. The rest goes into landfills and incinerators.

Rare earth elements found in electronics are valuable. Elements such as neodymium are used to create strong magnets used in hard drives and in-ear headphones. Mineral deposits containing neodymium are found in only a few places on earth.

Dry Household Waste

Metal, plastic, glass, and paper products discarded by households can be recycled and reused in various ways. Over half of recycled cardboard is made into new cardboard boxes. Lower quality used cardboard can be downcycled into cereal boxes and shoeboxes, according to Earth911.

The Glass Packaging Institute says glass is 100% recyclable and can be reused endlessly without losing its quality or purity.

The Critical Technologies behind Waste Sorting Robots

The main components of a robotic waste sorting system are the following:

  • Robotic arm
  • Gripper
  • AI-vision system
  • Conveyor system

A variety of robot types are used for sorting recyclable waste. Each has its benefits and limitations. Some vendors use a Cartesian robot suspended from an overhead gantry. The advantage of a Cartesian or XYZ robot is it can be made very strong and can lift heavy weights. However, Cartesian robots are not as fast as some other types. Another approach is to use a Delta-type arm, which has the advantage of being very fast. But Delta robots cannot handle items as heavy as the Cartesian bots.

Robotic waste sortation systems employ diverse gripping strategies, each with advantages and constraints. Some systems use a two-fingered gripper. These grippers pick objects up by their edges. Other systems employ a vacuum-cup gripper. A suction cup type of End-of-Arm-Tooling grabs things usually from their top surface. Suction cups might have difficulty with crumpled cans or objects with many edges and creases. Two-fingered grippers would not do as well on large pieces of cardboard when the edges are farther apart than the spread of the gripper.

A crucial part of the challenge of waste-sorting technology is object recognition. How to reliably tell if something is a plastic bottle when it might be crushed almost beyond recognition? The robotics systems use deep learning, in which images are collected and stored in a database. The more images, the better the recognition accuracy becomes.

Most waste sortation systems use a combination of sensors to distinguish between varied materials.

Near InfraRed (NIR) sensors use the wavelength signature of different resins to tell one kind of plastic from another. According to Tom Outerbridge, General Manager of the SIMS Municipal Recycling facility in Brooklyn, a disadvantage to NIR sensors is that they cannot “see” black plastic.

Robotic waste sortation systems often use visible light cameras. Combined with deep learning and A.I., such systems can adapt their behavior because of their learning. Some systems recognize logos like Starbucks and Amazon. This information can be used as feedback to the relevant companies, letting them know what percentage of their packaging is recycled.

Conclusions and Current Technology Readiness

The promise of robotics is to change the waste recycling economic equation. By reducing the cost and increasing the speed of sortation, recycling becomes more viable. The value of the waste material is “unlocked” in the economic calculations.

Artificial Intelligence and computer vision advances have made it possible for robots to do this task. According to an article in Engineering and Technology, robotic systems have picking rates of 2,000 to 4,200 picks per hour (pph), compared to a typical human waste-picker rate of 200 pph. A ten-to-twenty-fold increase in speed transforms the entire economics picture. Lowering the cost of sortation is crucial to a lower price per ton to recycle.

Another benefit to robotic waste sortation is an increase in the quality of the output. By reducing errors in the sortation, the purity of the fraction is increased. A higher-quality product commands a higher price. Better uniformity of recycled waste can bring double or triple the price of lower-quality output.

Multiple start-up companies are now addressing this need. There are many systems already in use. The technology is ready to scale up.

Yet, waste recycling robotics are costly. An article in Forbes reported a cost from one vendor of $300,000 per system (or $6000 per month to lease). To justify the initial investment, an MRF must have a sufficient volume.

How to Source the Ideal Waste Sorting Robot for Your Organization

HowToRobot is a global platform connecting end-users with robot and automation suppliers worldwide. We have the world’s largest directory of robotics companies. Using our guide, you can find the type of robot you need, ideally suited for your application.

If you want to automate a task within recycling, you can get tailored solution proposals from various suppliers. Simply describe your project and start receiving answers.

You can also get quotes and receive product information for specific recycling robots, parts, components, and consultancy services. You will receive product information and pricing from multiple vendors.

Please note that there are impartial HowToRobot experts who can help you navigate through the process. Click here to set up a consultation with an expert advisor.