Risky Trajectory Prediction for Safe Walkability under an Intuitionistic Fuzzy Environment

Trajectory prediction is a challenging problem in various fields such as road safety analysis, pedestrian tracking, and path planning. Road safety, in particular, is one of the most important services and challenges of smart cities. This is due to the huge loss it causes in lives and the resulting economic and social problems that directly affect the security of the state. While enormous research works have been devoted to addressing pedestrian road safety issues, further study on trajectory prediction is still essential to quantify this problem. In this chapter, we present a new model featuring a fusion between pedestrian trajectory prediction and pedestrian risk assessment. Specifically, we develop a fuzzy intuitionist model-based approach to extract spatiotemporally risk models from pedestrian risk trajectories. Then, we build a novel network among risk patterns for risky trajectory prediction using approaches based on deep neural networks, in particular, stochastic techniques, to deal with the large uncertainties on future trajectories, and the nonrigid aspect of the pedestrian.