The Macroscopic Fundamental diagram (MFD) has proven to be a powerful concept in understanding and managing vehicular network dynamics, both from a theoretical angle and from a more application-oriented perspective. In this contribution, we explore the existence and the characteristics of the pedestrian Macroscopic Fundamental Diagram (p-MFD). From a theoretical perspective, the main contribution of this research shows how we can derive the p-MFD from assumed local fundamental diagrams (FDs). We show that we can relate the average (out-) flow from a pedestrian network as a function of the average spatial density and the density spatial variation . We show that the latter is essential to provide a reasonable description of the overall network conditions. For simple linear relations between density and speed, we derive analytical results; for more commonly used FDs in pedestrian flow theory we show the resulting relation using a straightforward simulation approach. As a secondary contribution of the paper, we show how the p-MFD can be constructed from pedestrian trajectory data stemming from either microsimulation or from experimental studies. The results found are in line with the theoretical result, providing further evidence for the validity of the p-MFD concept. We furthermore discuss concepts of hysteresis, due to the differences in the queue build up and recuperation phases. We end with applications of the presented concepts, e.g. in crowd management.
AbstractThe Macroscopic Fundamental diagram (MFD) has proven to be a powerful concept in understanding and managing vehicular network dynamics, both from a theoretical angle and from a more application-oriented perspective. In this contribution, we explore the existence and the characteristics of the pedestrian Macroscopic Fundamental Diagram (p-MFD). From a theoretical perspective, the main contribution of this research shows how we can derive the p-MFD from assumed local fundamental diagrams (FDs). We show that we can relate the average (out-) flow from a pedestrian network as a function of the average spatial density and the density spatial variation . We show that the latter is essential to provide a reasonable description of the overall network conditions. For simple linear relations between density and speed, we derive analytical results; for more commonly used FDs in pedestrian flow theory we show the resulting relation using a straightforward simulation approach. As a secondary contribution of the paper, we show how the p-MFD can be constructed from pedestrian trajectory data stemming from either microsimulation or from experimental studies. The results found are in line with the theoretical result, providing further evidence for the validity of the p-MFD concept. We furthermore discuss concepts of hysteresis, due to the differences in the queue build up and recuperation phases. We end with applications of the presented concepts, e.g. in crowd management.