Tech watch: We review the latest circular innovations

From the newsletter

Since September, several innovations have emerged seeking to advance the circular economy. Circular Rising highlights four that stand out for their anticipated impact. They include artificial intelligence-driven waste sorting technologies and an innovation that turns dirty plastic film into reusable material for making new products or for recycling.

• Two of the AI technologies have been developed by researchers at Indian and American institutions, while the third is a commercially available system that uses AI-driven robotic arms to recover metals from mixed waste. 

• These technologies aim to improve efficiency in waste sorting, reduce contamination, and support more sustainable recycling in urban and industrial settings.

More details

• Among the innovations is HybridSOMSpikeNet, developed by researchers at the Indian Institute of Technology. It is an AI system that automatically identifies and sorts different types of waste with 97 per cent accuracy. The system analyses visual patterns in waste, groups similar materials together, and learns from each sorting task to improve over time. This approach allows the system to handle diverse waste streams with precision, adjusting to different materials while improving recycling outcomes.

• What makes HybridSOMSpikeNet unique is its ability to combine fast pattern recognition with continuous learning, allowing it to handle complex waste streams more effectively than many current solutions. Its compact design and computational efficiency make it suitable for research-stage deployment in urban facilities or smaller waste centres, reducing contamination and supporting cleaner recycling

• Meanwhile, researchers at Cornell University in the United States have developed DWaste, a computer vision-based model designed to run on smartphones or small edge devices. The system identifies and classifies waste types in real time, enabling rapid sorting without the need for extensive infrastructure. This makes it particularly suitable for low-resource environments and decentralised recycling initiatives.

• DWaste uses lightweight classification and detection networks to achieve accurate sorting while maintaining low computational requirements. By enabling real-time identification of waste streams, DWaste helps reduce contamination, improves recycling efficiency and lowers operational costs. The system supports localised and community-level waste management, offering a practical solution for cities and regions looking to modernise their recycling processes.

• Meanwhile, Canada-based Waste Robotics has developed WR4, a system that focuses on automated metal recovery from mixed waste streams using AI-driven robotic arms. It is designed for industrial recycling facilities and can identify and pick metals or valuable materials with high precision.

• By automating the sorting process, WR4 increases material recovery efficiency, reduces reliance on manual labour and improves the quality of recyclable outputs. The system can handle complex and contaminated waste streams, making it suitable for modern, high-throughput recycling operations. Its deployment helps facilities streamline operations, lower operational costs and support circular economy goals by recovering valuable materials that would otherwise be lost.

• Away from the AI innovations, Erema, a German company, has developed Agglorema, a technology designed to process heavily contaminated plastic film, including moist or low-density scrap. The system homogenises, degasses, and preheats the waste before feeding it into an extruder, turning mixed and dirty plastic into high-density, pourable agglomerates ready for reuse.

• Agglorema stands out because it can process challenging, contaminated materials that conventional recycling systems struggle with. The resulting agglomerates can be used to produce simple plastic components or serve as feedstock for chemical recycling, helping to reduce waste, boost material recovery, and support a more circular plastics economy.

Our take

• The strong presence of technologies developed by universities shows that institutions of higher learning remain at the forefront of innovation, creating practical solutions to complex challenges in the circular economy.

• The dominance of AI technologies indicates that the sector is shifting from purely mechanical or manual recycling processes towards data-driven, automated solutions capable of handling complex and contaminated waste streams.

• For Africa, these innovations could improve waste sorting and increase recycling efficiency, but adoption may be constrained by infrastructure gaps and high implementation costs.