Intelligent trajectory prediction algorithm design for dynamic obstacles under factory environments

Abstract

In modern days, many factories have incorporated smart technologies, such as mobile robots and automation, into their environments to improve workflow efficiency. However, moving through factory floors poses significant challenges due to dynamic obstacles such as other moving machinery and human workers. In particular, human movements are the hardest to predict. Predicting human movement in crowded environments is a complex task due to many factors, one being the intricate social interactions among people. Traditional models often fail to account for these factors effectively. To devise an algorithm enabling robots to move safely and efficiently on factory floors, we must first develop an algorithm that can predict human trajectory with high precision and accuracy. This paper aims to explore and study the existing human trajectory prediction algorithm based on machine learning and determine the most suitable model to be used for factory floor navigation.