Toward smarter management and recovery of municipal solid waste: A critical review on deep learning approaches

Increasing generation of municipal solid waste, heterogeneity of waste composition, and complex processes of waste management and recovery have limited the performance of traditional treatment approaches. It is urgent to innovate waste management toward smarter and more efficient modes and break up the bottlenecks of the current system. Recently, deep learning has emerged as a powerful method for revealing hidden patterns or deducing correlations for which traditional treatment approaches face limitations or challenges. However, deep learning concepts and practices have not been widely utilized by researches in municipal solid waste management (MSWM). Herein, this research provides a critical review for deep learning and its application in MSWM. The framework and algorithms of a variety of deep learning methods have been compared and assessed. A body of deep learning applications have been reviewed according to their engagement in waste collection, transportation, and final disposal. Application of deep learning in MSWM stays in its infancy and requires great efforts for further development. The challenges and futures opportunities in the application of deep learning in the MSWM have been discussed to highlight the potential of deep learning in this field.