Engineering Smarter E-Waste Disassembly with Robots and AI
The Fraunhofer Institute for Factory Operation and Automation IFF (Fraunhofer IFF) is working to overcome challenges to e-waste disposal.
05 May, 2025. 5 minutes read
Credit: Fraunhofer IFF
Electronics waste (e-waste) is piling up. With global e-waste surpassing 50 million metric tons each year,[1] the need for efficient and scalable recycling solutions is critical. In the US, where e-waste management is largely handled by state-level regulations and private recycling companies, automation could be the key to tackling the problem on a large scale. This blog explores current challenges to e-waste disposal and how organizations like the Fraunhofer Institute for Factory Operation and Automation IFF (Fraunhofer IFF) are working to overcome them.
The Problem: Handling Diverse, Outdated Electronics
One of the biggest challenges in electronics recycling is the wide variety of products that end up in recycling centers. Recyclers are faced with a muddled mix of components found in everything from cell phones and laptops to industrial equipment. Many of these components have no digitally available schematics, parts lists, assembly drawings, or standardized designs to help recyclers discern materials that can be reclaimed.
Traditional recycling practices involve manually dismantling devices to separate metals, plastics, and circuit boards. As you might imagine, this process is slow, labor-intensive, and sometimes even involves dealing directly with hazardous materials. In many cases, electronic devices are just shredded outright. While this method can recover materials like gold and silver, other critical raw materials, including rare earth elements, are often lost in the process.
"You take an old computer or a mobile phone, and it just gets shredded up into a big sludge. And then, with that, they're able to do some kinds of chemical processes to get the gold, silver, platinum, these kinds of things out," said Dr. José Saenz, Robotic Systems, Assistive, Service and Industrial Robots Group Manager, Fraunhofer IFF. “They're able to get out the metals, so that works; but there's no way to get out the plastic, so they basically just get incinerated.”
The inefficiency of material recovery is one of the biggest obstacles to creating a truly circular electronics economy where electronic products are designed for longevity and recyclability.
Revolutionizing Disassembly
To solve these issues, Saenz and his team have been using artificial intelligence (AI)-powered robotics to automate disassembly. Unlike traditional automation, which relies on predefined programs, modern robotics are adaptable and leverage technologies like computer vision, force-sensitive actuators, and machine learning to handle unpredictable materials.[2]
Saenz described his team's Intelligent Disassembly of Electronics for Remanufacturing and Recycling (iDEAR) project approach (Figure 1). The project currently focuses on one product type, but the team has plans to expand: "We are focusing right now on PCs, and we want to get the main board out so that it can be processed further. We want to do this automatically; but the idea is, if we set up this process in the right way, it'll be possible to do the same kind of setup for other products."
Rather than programming robots to follow specific X, Y, and Z positions, his team is developing a generic storyboard-based sequence for disassembly. This allows the robot to dynamically determine the required tools, reducing reliance on human intervention and making the process scalable for different electronics.
Sensors, AI-Powered Vision, and Force-Sensitive Robotics
The Fraunhofer team is training its robots to assess and disassemble devices using two methods: reinforcement learning and imitation learning.
In reinforcement learning, the robot tries different ways to remove components and gets rewarded for doing it correctly, improving over time. In imitation learning, a human operator guides the robot through the disassembly process, and the robot learns by mimicking those actions (Figure 2). These approaches allow the system to adapt to various devices without requiring pre-programmed instructions.
Their system includes the following technologies:
- Computer vision and AI: Advanced object detection algorithms (such as You Only Look Once, or YOLO, and object detectors) enable robots to recognize components in real time, even when dealing with a mix of unknown products.
- Force-sensitive actuators: Robots equipped with force-feedback mechanisms can adjust their approach based on resistance. This helps them handle stuck screws, warped casings, or delicate components without causing damage.
- Modular tooling: The system allows the robot to switch between tools, such as screwdrivers, grippers, and cutters, to perform the disassembly tasks.
Digital product passports—which provide “comprehensive information about each product’s origin, materials, environmental impact, and disposal recommendations”—could also play a role in future recycling efforts.[3]
The Impact of Automated Disassembly
The European Union has led the way with Right-to-Repair legislation,[4] but the US is starting to follow suit. All 50 US states have introduced legislation requiring manufacturers to make repair manuals and components available to consumers to encourage more repairable product designs.[5] This would also make disassembly easier for recyclers.
For engineers designing future products, these trends could mean a shift toward modular designs, standardized components, and accessible disassembly points. Future AI-driven disassembly systems could even integrate with manufacturer databases to allow robot access to schematics and automatically optimize recycling processes.
Automated disassembly would also impact the mining of materials like cobalt, lithium, and rare earth metals, which is environmentally destructive and geopolitically sensitive. Automating the disassembly and material recovery process could reduce reliance on newly mined resources by increasing recycling efficiency. For example, AI-powered systems could selectively remove valuable semiconductors, battery cells, and high-purity metals, making them available for reuse in new devices.
Challenges and Next Steps
It will not be easy to achieve widespread adoption of AI-driven electronics recycling, but Saenz and his team are focused on scalability and impact. So far, their progress shows that robotics and AI can speed disassembly, reduce waste, and improve materials recovery.
One of the most promising aspects of this technology is its ability to be adapted for electronic devices beyond PCs. The team is working to make the technology grow beyond PCs and become part of larger recycling operations that handle products like mobile phones and industrial electronics. The modular design will allow robots to switch between tools depending on the task, making the system adaptable and scalable.
The next research phase is even more ambitious: The team is looking to apply these AI-driven disassembly methods to jet turbines and other components currently being remanufactured.
Additionally, with advancements in digital product passports, the availability of product data will increase, improving the ability of AI systems to quickly assess and disassemble electronics. Right now, the lack of historical data is a hurdle; but Saenz believes that as more companies embrace circular economy principles, data accessibility will improve, and robotic disassembly will be even more effective.
Saenz doesn't see high upfront costs as a roadblock. He believes investment in automation will ultimately reduce operational expenses, as AI-driven disassembly minimizes labor-intensive processes and maximizes material recovery. As regulations push for more sustainable practices, industries will have more reasons to adopt these solutions, making them more cost-effective over time.
Saenz envisions a future where human-robot collaboration is optimized, with automation handling complex disassembly while humans oversee and refine the process. With continued research and industry collaboration, AI-driven electronics recycling has the potential to revolutionize waste management, making sustainability an achievable and profitable goal for manufacturers and recyclers.
Conclusion
For engineers, the rise of automated electronics recycling represents both challenges and opportunities. Designing products with AI-assisted disassembly in mind, leveraging Right-to-Repair principles, and integrating digital product passports could create a more sustainable electronics industry.
With advancements in robotic vision, machine learning, and modular product design, the future of electronics recycling looks smarter and more efficient. Saenz emphasized that his team is eager to collaborate with industry professionals to ensure their research is not only innovative but also practical.
"We don't want to build things that the world doesn't need. We want to hear from industry leaders about their specific challenges so we can create something truly useful and scalable," he said.
Resources:
[1] https://ewastemonitor.info/gem-2020/.
[2] https://www.iff.fraunhofer.de/en/press/2025/automated-electronics-disassembly-first-milestones-in-the-research-project.html/.
[3] https://data.europa.eu/en/news-events/news/eus-digital-product-passport-advancing-transparency-and-sustainability/.
[4] https://www.europarl.europa.eu/news/en/press-room/20240419IPR20590/right-to-repair-making-repair-easier-and-more-appealing-to-consumers/.
[5] https://www.ifixit.com/News/108371/right-to-repair-laws-have-now-been-introduced-in-all-50-us-states.