Robotics in solid waste handling: From science fiction to reality

It is currently estimated that approximately 55% of the world population lives in urban centers (World Bank, 2020). This is partly caused by phenomena such as the migration of people from rural areas toward cities. A direct consequence of this situation is the increase in solid waste (SW) generation. In 2018, the World Bank stated the amount of waste generated annually is approximately 2 trillion tons (World Bank, 2018). Particularly, in low-income countries, 90% of waste is not appropriately handled. With-out proper SW handling, the environment becomes polluted, causing a global problem. In support of these arguments, we can assert that ad-equately recovering SW has become a serious and widespread problem in urban areas. Therefore, an appropriate treatment of SW is necessary to reduce the environmental impact generated by this kind of residual. There are different processes associated with the task of SW handling, as shown in Fig. 1. As evidenced in the scientific literature, authors have approached SW handling from different perspectives, especially to improve the processes associated with the recycling chain. These approaches have included the use of techniques from different areas, such as robotic methods [mechanical arms for waste handling in recycling plants (Chen et al., 2019)], machine learning [for the separation and categorization of materials located on the ground (Liao et al., 2019)], deep learning [for the recognition of construction waste located in work areas (Ku et al., 2021)], and neural networks [for the separation of recyclable items from landfill trash objects (Sultana et al., 2020)], among others. This article examines how diverse techniques provided through science fiction (sci-fi) some decades ago have inspired the development of solutions applied to procedures for SW handling in the real world, notably, for recovery; separation; and classification, transit, and storage operations.