Project: MPPS - Multipurpose Plastic Sorting

A new generation of plastic sorting is needed to reach Europe’s circularity goals. This project explored radical solutions to sort plastic waste smarter, faster, and with higher quality — from household packaging to hospital waste.

In short:

Europe recycles less than 20% of post‑consumer plastics; new solutions are urgently needed.
Three novel sorting concepts were developed and demonstrated with industry partners.
Results show promising pathways for AI‑based detection, sorting of medical waste for pyrolysis, and swarm‑based and multi‑fraction sorting.
Prototypes demonstrate proof‑of‑principle potential for future scaling.

From highest throughput to high-quality plastics

Europe produces almost 30 million tonnes of post‑consumer plastic waste each year. Only 19% becomes recycled plastics. Most is still incinerated, landfilled, or lost along the way. In the Netherlands the situation is better, yet still insufficient to reach the national goal of recycling over 50% of plastic waste by 2025.

In current plastic sorting facilities, the focus is on achieving the highest throughput. They don’t prioritise high‑quality or specialised output streams needed for advanced recycling. New forms of mechanical sorting — more accurate, more flexible, and tailored to specific end‑uses — are essential to reduce dependence on fossil resources and close the loop for plastics.

Three use cases on mechanical sorting

The MPPS – Multipurpose Plastic Sorting Project explored radically different approaches to mechanical sorting. The project aimed to:

Develop innovative, cross‑sector technologies for sorting post‑consumer plastics
Design new concepts that combine existing but not used techniques for plastics sorting
Use real industrial waste streams as test cases
Build and demonstrate modular prototypes to prove feasibility

In this project we worked with three use cases from industrial partners, each targeting a plastic waste challenge not addressed by current sorting systems.

Results

1. Waste transfer station rejects (Midwaste)

The objective of this use case was to develop a technology based on AI vision that can detect and classify contamination from correctly collected LWP.

Show the results

Waste transfer stations rely on manual visual inspections to identify contaminants in source‑separated lightweight packaging waste (LWP). This is subjective and inconsistent.

We assessed whether existing AI object‑classification systems could automate this step. Tests with several technology providers showed that off‑the‑shelf models cannot handle the wide variability of contaminants without huge datasets and complex classifications.

*Outcome of the anomaly detection model*

A custom anomaly‑detection model — identifying “non‑LWP” objects rather than classifying every contaminant — showed promising proof‑of‑principle results. A prototype system successfully highlighted contamination using this approach, signalling potential for future development.

2. Unlocking hospital waste for chemical recycling (SABIC)

The objective of this use case was to develop a sorting process for a medical waste stream that provides feedstock for thermo-chemical recycling via pyrolysis.

Show the results

Most non‑specific hospital waste (NSHW) is incinerated, despite containing 30–40% valuable polyolefins. For pyrolysis feedstock, disturbing objects must be removed reliably.

The team designed several sorting concepts and tested the most promising route: shredding the waste to homogenise size, followed by optical sorting. Trials at NTCP reached 90% purity — sufficient for pyrolysis.

Small‑batch pyrolysis experiments showed that sorted NSHW can indeed yield usable oil. Economically, the process is challenging at current waste‑handling costs, but improved collection practices in hospitals may support future feasibility without burdening medical staff.

A tour guided by Lieuwe-Lei Hendriks at the Nationaal Testcentrum Circulaire Plastics.

Curious? Read more in this interview with NTCP and Sabic.

3. Sorting higher quality with innovative waste sorting solutions (Omrin/Danone)

The state-of-the art technology for waste sorting is a binary sorting approach. That means that a sorting machine only sorts into two fractions. Alternative sorting principles, existing in other sorting domains, are swarm-based sorting and multi-fraction sorting. The objective of this use case was to assess the feasibility of swarm-based sorting and multi-fraction sorting on household packaging waste. It is intended to translate the principle into a tangible prototype that demonstrate the technology up to a proof-of-principle level (TRL 3-4).

Swarm‑based sorting

Swarm‑based sorting is a bold concept inspired by robotics domains outside waste management. The project developed a tabletop robot platform to demonstrate the fundamentals: manoeuvring, object detection, and manipulation using a swarm of five robots.

In parallel, TU/e and Antfarm built their own robot platforms based on the same principles. All three prototypes show technological and economic potential worth further exploration.

Multi‑fraction sorting

Current sorting machines typically split waste into only two fractions. Multi‑fraction sorting is based on the principle that an object is detected once after which it can be separated into multiple fractions without loosening its detection information and localisation. This reduces many detection steps and allows for more sorting fractions.

A prototype system using electromagnetic actuators below a conveyor belt was built and demonstrated. Early tests show potential, but a dedicated test campaign is needed to validate feasibility at higher throughput.

*CAD design of the test-set-up of multi-fraction sorting using voice-coil actuation to release objects.*

Conclusion

The MPPS project delivered the concepts, prototypes, and proof‑of‑principle demonstrations envisioned at the start. Each work package resulted in tangible designs evaluated with partners and stakeholders. Together, the outcomes show real potential for next‑generation sorting solutions that can increase the circularity of plastic waste in Europe.

This is a project of the Circular Plastics Initiative, powered by ISPT and DPI.