Plastic Sorting via Multi-Sensor Fusion and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current plastic recycling technologies face challenges in efficiently sorting and recycling plastics beyond type #1 and #2, particularly struggling with black plastics and multilayered polymers, due to limitations in existing sensor technologies and the inability to accurately identify these materials, leading to contamination and reduced recycling rates.
Innovation Solution
A system combining multiple sensor technologies, including spectral imaging and machine learning, to classify and sort plastics based on both organic and inorganic compositions, enabling the separation of #3 through #7 plastics and mixed polymer types into novel fractions, overcoming the limitations of single-sensor systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If single-sensor systems are used for plastic sorting, then device complexity is reduced, but measurement precision and sorting accuracy deteriorate due to inability to accurately identify black plastics and multilayered polymers
Solution Approach 1:
The patent combines multiple sensor technologies (NIR, VIS, SWIR, XRF) into an integrated sorting system. Each sensor type detects different material properties: NIR for polymer identification, VIS for color detection, SWIR for black plastic detection, and XRF for inorganic composition analysis. This multi-sensor approach resolves the contradiction by achieving high measurement precision through sensor fusion while managing device complexity through systematic integration and centralized control.
Solution Approach 2:
The sorting system is designed with multi-functional capabilities to handle diverse plastic types including black plastics, multilayered polymers, and mixed materials. The system performs multiple functions simultaneously: identification, classification, and sorting of various plastic categories. This universality allows a single integrated system to accurately detect and sort all plastic types without requiring separate specialized systems for each material category.
2Reliability
If advanced multi-sensor systems are implemented, then sorting accuracy and recycling quality improve, but device complexity and capital costs increase
Solution Approach 1:
The system segments the sorting process into distinct functional modules: NIR sensing for polymer type detection, VIS sensing for color identification, SWIR sensing for black plastic detection, and XRF sensing for inorganic content analysis. Each sensor module operates independently but contributes to the overall sorting decision. This segmentation allows for modular deployment and maintenance while achieving high sorting accuracy through coordinated operation of specialized subsystems.
Solution Approach 2:
The patent introduces a centralized control system that acts as an intermediary between multiple sensors and the sorting mechanism. This mediator integrates data from all sensor types, processes the information through machine learning algorithms, and generates sorting commands. The intermediary layer manages the complexity of coordinating multiple sensors and translating their outputs into actionable sorting decisions, thereby achieving high reliability without proportionally increasing operational complexity.
3Productivity
If traditional sorting methods are used, then device complexity remains low, but productivity and recycling rates remain limited due to inability to process diverse plastic types
Solution Approach 1:
The system utilizes different sensing parameters (wavelength ranges, detection methods) to identify various plastic types. NIR detects polymer-specific absorption patterns, VIS detects color characteristics, SWIR detects black pigment absorption, and XRF detects inorganic elements. By changing and combining multiple detection parameters simultaneously, the system achieves high productivity across diverse plastic types rather than being limited to single-parameter detection methods.
Solution Approach 2:
The patent replaces manual or simple mechanical sorting methods with an automated optical and electronic detection system. Machine learning algorithms process sensor data to automatically classify plastics, eliminating the need for manual inspection and sorting. This substitution dramatically increases productivity and enables processing of complex plastic types that would be impossible to sort manually, justifying the increased device complexity through substantial gains in throughput and capability.
4Manufacturing precision
If comprehensive material identification is performed, then sorting quality and recycled material purity improve, but loss of time and processing speed increase
Solution Approach 1:
The sorting system operates continuously with all sensors detecting materials simultaneously as they move along the conveyor belt. The NIR, VIS, SWIR, and XRF sensors operate in parallel without sequential processing delays. This continuous simultaneous detection maintains high sorting precision for all material types while minimizing processing time, as no single detection method needs to complete before the next begins.
Solution Approach 2:
The system performs preliminary classification using faster detection methods (NIR for polymer type, VIS for color) to quickly identify broad material categories. Based on this preliminary information, the system then applies more specialized detection (SWIR for black plastics, XRF for inorganic content) only when needed. This staged approach achieves comprehensive material identification and high sorting precision while minimizing overall processing time by avoiding unnecessary detailed analysis for all materials.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the accuracy and efficiency of plastic sorting, allowing for the creation of higher-quality recycled materials and increased recycling rates by accurately identifying and separating a broader range of plastic types, including previously difficult-to-sort materials like black plastics and multilayered polymers.
Implementation Method 1
these sensors rely on the reflection of light from an external source and can only view the surface of the material
Implementation Method 2
NIR spectroscopy can identify #1 type plastics that are clear and light blue PET and #2 HDPE
Data Source
AI summary
Systems and methods for classifying and sorting of plastic materials utilizing a vision system and one or more sensor systems, which may implement a machine learning system in order to identify or classify each of the materials, which may then be sorted into separate groups based on such an identification or classification.