The CircuBot will exploit and further develop recent advances in emerging technologies: robotics – pick-and-tossing for increasing working space & soft gripers for versatile material handling; and AI and computer vision – for multi-object detection and shape analysis; and cloud computing for optimized waste management. Combining those methods and expertise of two promising Serbian teams in robotics and AI, complemented by waste management and environmental protection experts, CircuBot will result in a novel prototype of versatile (applicable for different waste materials – PET, can, electronic), modular (different robots, grippers and settings can improve efficiency), cooperative (can work side by side with human operators), efficient (with improved robot capabilities for material handling – pick-and-toss) and intelligent (for optimal use of resources in selection and sorting) robotic system for waste sorting and management.

The CircuBot provides technologies that make impact to key Program areas – Circular economy and Environment

Improved efficiency and digitization of waste sorting management – Waste sorting plants in Serbia are relying on human factors so the occurrence of flaws in waste sorting and treating commonly results in expensive claims – while the source of flaws commonly remains not possible to find and prevent in the future. Due to the lack of digitalization, there is a high pressure on top-management to be constantly present in the waste sorting halls – whose time could be better invested to enable a company’s business growth.  

The digitalized robotic workstation proposed within this project  opens new possibilities to significantly improve waste management by automatic collection and analysis of large amounts of data extracted with AI from processed waste. The CircuBot cloud-based waste management platform will enable centralized monitoring, maintenance, and control of workstations. Smart data analytics will provide tools for waste counting, per-class analysis, reporting, and forecasting.

Contribution to the specific scientific field. Computer science, Industrial engineering, Waste management.

Short-term impact (CircuBot outcome). In this way, waste management will be more objective and predictive, managers will be released from the constant presence and the quality of the recycled product will be easily tracked. 

Long-term impact. In the long run, systematic analysis of waste will lead to the agile management of recycled products, such that the quality of the recycled material is always determined. In this way, companies will remain competitive and easily spread to a European market (foreseen 10% larger amount of exported plastic and electronic waste). Waste analytics will improve recycling protocol through the reduction of waste sorting errors by at least 50% and the reduction of the amount of final waste that would eventually end up in the landfill by at least 30% in Serbia.

Target group: Direct and intended impact on waste management plants. 

 

Improved efficiency of waste handling – Waste management plants in Serbia and beyond adopt different mechanical procedures to sort the waste, such as magnets, rotary sieves, and air separators. While each method is effective per se, it still cannot deal with the significant variability of processed waste. A human worker is still needed at different stages of the recycling process, such as preselection and final sorting and inspection. Thus, humans work in a dusty, dirty environment being in touch with hazardous materials.

Project result: Versatile, safe, and efficient robotic system for waste sorting.

Short-term: The proposed system will be able to perform sorting with a significant efficiency enabled by AI and pick-and-toss solution  and improve conditions for a human worker. At least 10 waste management plants will be exposed to cutting-edge technology. Reduced health problems and infections of the human workers since they don’t have to touch waste with their hands (reduced number of sick leaves by 20%). The research will lead to an increased number of publications on waste management topics (refer to section 2.2).

Long-term: Reducing future investments due to the scalability (arbitrary number of robots) and versatility (three types of grippers) of the solution. It is foreseen that more than 30% of companies that learned about this technology will decide to install it in the next 5 years. Increased number of sustainable, high-quality jobs: new job opportunities for robotics/AI engineers. Reduced musculoskeletal injuries at plants. Plants increase their profit by at least 10%. Scientific field: artificial intelligence, robotics and automation.

Target group: Direct and intended impact on academy and waste management plants. Indirect impact on logistics and manufacturing factories, which might profit from the pick-and-toss strategy.

 

Improved waste detection and sorting – The number of waste management plants in Europe that implement visual waste detection systems is steadily growing. Contrarily, Serbian waste management plants are still not exploiting the benefits of this technology.

Project result: A state-of-the-art system based on deep learning will be developed for waste detection and labeling and a database of communal waste images will be made.

Short-term: Using this system will lead to increased speed and enhanced accuracy of the waste sorting process.
Long-term: Increased capacities of waste management plants by at least 20%. Increased number of high-quality jobs: new job opportunities for computer vision engineers. The database of communal waste images will have a world-wide impact on the research society: an increased number of high-ranking journal papers that exploit the database for developing artificial intelligence algorithms (at least 30 new papers in the next 5 years).

Scientific field: computer vision, artificial intelligence.

Target group: Direct and intended impact on academy and waste management plants.

 

CircuBot’s impact on the environment

According to Serbian Law on Waste Management (Official Gazette of the Republic of Serbia 36/2009-95/2018) waste management hierarchy includes 5 steps: reducing waste at the source, reuse of materials, recycling, energy recovery, and landfilling (Figure 5, right). It gives top priority to waste prevention, which is usually impossible in non-food industries. In the middle of the hierarchy is reuse, recycle and energy recovery. Since energy recovery of plastic waste for example leads to great emissions of hazardous gases, reuse and recycling remain the only environmentally friendly options.

Higher reuse of waste and efficiency of recycling process – In line with the national policy on waste management, CircuBot will contribute to reducing environmental pollution by efficient waste sorting and monitoring.

Project result. The digitalized robotic workstation proposed in this project will have multifold benefits for society and the environment. Waste data analytics will lead to zero waste since monitoring the sorting process will provide excellent insight into improving the existing recycling technology. More efficient recycling technology will lead to reduced emissions – littering of the city garbage container and municipal landfills will be significantly reduced, as well as hazardous substances emissions from the landfills, pollution of the air, surface and groundwater, and soil. The proposed research will directly lead to a cleaner environment.

Short-term impact. The digitalized robotic workstation will result in faster sorting. Faster sorting will ensure minimizing the energy consumption of the recycling plant and the production of waste over the life cycle of the product.

Long-term impact. Reduced landfilling by at least 30%. Reduced disposal cost by at least 30%. Improved alignment with the European Green Deal. A 5% reduction in greenhouse gas emissions from mixed waste sorting could save 0.73 billion tons of CO2 per year. Data analytics will improve the official statistics of the produced waste which will increase the environmental awareness of people and impact their carbon “footprint”. More realistic data on waste generated by people will affect their attitudes toward waste and make them more eco-friendly minded and careful while buying non-recyclable products or sorting waste in their houses (at least 23.000 visitors of fairs will learn about waste management, our posts on social media will reach at least 1000 people, at least 20 Serbian entrepreneurs will learn about the project activities).

Target groups. Residents living close to landfills, city residents, recycling plant workers