Waste Classification via Friction Sound Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing waste classification methods are complex, costly, and not suitable for devices with limited resources, requiring improved, cost-effective, and accurate waste detection systems.

Innovation Solution

A method using a friction element in a waste container to generate a sound signal captured by a sensor, combined with machine-learning classifiers like neural networks, particularly utilizing Mel Frequency Spectral Coefficients (MFSC) for audio signal analysis, reduces computational costs and enhances classification accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If complex classification methods are used to detect and separate waste, then classification accuracy is improved, but device complexity and manufacturing cost increase

Engineering Contradiction:
Improveclassification accuracyVSAvoiddevice complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces complex mechanical classification systems with an acoustic-based system. A friction element (such as a ramp or inclined plane) is introduced to generate characteristic friction sounds when waste items slide along it during insertion. These acoustic signals are captured by a microphone and processed through a neural network classifier to identify waste types, thereby substituting mechanical complexity with acoustic sensing and computational processing.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The friction element acts as an intermediary between the waste item and the classification system. Instead of directly analyzing the waste item's physical properties through complex mechanical means, the system uses the friction element to convert the insertion action into a characteristic acoustic signal that carries information about the waste material type, simplifying the overall classification mechanism.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If high computational cost methods are used for waste classification, then classification accuracy is improved, but suitability for embedded devices with limited resources decreases

Engineering Contradiction:
Improveclassification accuracyVSAvoidcomputational energy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent extracts only the essential acoustic features from the friction sounds generated during waste insertion. By focusing on specific acoustic characteristics (frequency, amplitude, temporal patterns) rather than processing complete raw audio signals or using complex multi-sensor data, the system reduces computational requirements while maintaining classification accuracy, making it suitable for embedded devices with limited processing power and energy resources.

Inventive Principle:
Principle #2Taking out (Extraction)

3Device complexity

If simple classification methods are used, then device complexity is reduced, but classification accuracy decreases

Engineering Contradiction:
Improvedevice complexityVSAvoidclassification accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent changes the parameter being measured from complex physical or visual properties of waste to acoustic parameters (frequency, amplitude, temporal characteristics) of friction sounds. This parameter transformation enables the use of relatively simple hardware (microphone, friction element) while achieving accurate classification through neural network processing of the acoustic signals, thus maintaining low device complexity while improving classification accuracy.

Inventive Principle:
Principle #35Parameter changes

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 provides reliable, efficient, and low-cost waste classification with reduced computational requirements, suitable for devices with limited resources, by leveraging stable sound signals from waste interaction with friction elements.

Implementation Method 1

the friction of the waste item with the friction element. A device may be provided in the container, in correspondence with the waste inlet, on which the waste can bounce or which the waste can hit, and produce a sound to be captured by the sound sensor

Methodology Applied
Scientific EffectFriction: Friction

Data Source

PatentEP4157747B1A method for classifying waste
Publication Date: 2025.07.23 CANDAM TECH SL
  • EP4157747B1 patent drawingFigure 1
  • EP4157747B1 patent drawingFigure 2A~2B
  • EP4157747B1 patent drawingFigure 3

AI summary

A method for classifying waste when inserted in a waste container is presented, the method comprising capturing audio signals corresponding to the interaction of a waste item with a friction element in the waste container, when the waste item is inserted in the container and slides on the friction element, and performing a classification of the captured audio signal as corresponding to a type of waste item detected as a predetermined type of waste item, the classification being performed based on MFSC.