sc 21(16): e1

Research Article

Transfer learning-based method for automated e-waste recycling in smart cities

Download1447 downloads
  • @ARTICLE{10.4108/eai.16-4-2021.169337,
        author={Nermeen Abou Baker and Paul Szabo-M\'{y}ller and Uwe Handmann},
        title={Transfer learning-based method for automated e-waste recycling in smart cities},
        journal={EAI Endorsed Transactions on Smart Cities},
        volume={5},
        number={16},
        publisher={EAI},
        journal_a={SC},
        year={2021},
        month={4},
        keywords={Artificial Intelligence, Transfer Learning, Circular Economy, Automated E-Waste Recycling, Smart Cities},
        doi={10.4108/eai.16-4-2021.169337}
    }
    
  • Nermeen Abou Baker
    Paul Szabo-Müller
    Uwe Handmann
    Year: 2021
    Transfer learning-based method for automated e-waste recycling in smart cities
    SC
    EAI
    DOI: 10.4108/eai.16-4-2021.169337
Nermeen Abou Baker1,*, Paul Szabo-Müller1, Uwe Handmann1
  • 1: Ruhr West University of Applied Sciences, Lützowstr. 5, 46236 Bottrop, Germany
*Contact email: nermeen.baker@hs-ruhrwest.de

Abstract

INTRODUCTION: Sorting a huge stream of waste accurately within a short period can be done with the support of digitalization, particularly Artificial Intelligence, instead of traditional methods. The overlap of Artificial Intelligence and Circular Economy can flourish many services in the environmental technology domain, in particular smart e-waste recycling, resulting in enabling circular smart cities.

OBJECTIVES: We analyse the growing need for automated e-waste recycling as an essential requirement to cope with the fast-growing e-waste stream and we shed the light on the impact of Artificial Intelligence in supporting the recycling process through smart classification of devices, where the smartphone is our case study.

METHODS: Our study applies transfer learning as a special technique of Artificial Intelligence by fine-tuning the output layers of AlexNet as a pre-trained model and perform the implementation on a small-size dataset that contains 12 classes from 6 smartphone brands.

RESULTS: We evaluate the performance of our model by tuning the learning rate, choosing the best optimizer, and augmenting the original dataset to avoid overfitting. We found that the optimizer of Stochastic Gradient Descent with Momentum and 3 𝑒𝑒−4 as a learning rate brings almost 98% model accuracy with generalization.

CONCLUSION: Our study supports automated e-waste recycling in decreasing the error-rate of e-waste sorting and investigates the advantages of applying transfer learning as the best scenario to overcome the rising challenges.