Fuzzy prediction of pedestrian steering behaviour with local environmental effects

Publication Type:

Conference Paper


Pedestrian and Evacuation Dynamics PED 2012 (2012)


This research focuses on prediction of pedestrian walking paths in indoor public environments during normal and non-panic situations. The aim is to incorporate imprecise and subjective aspects of pedestrian interaction with the environment to enhance steering behaviour modelling. The proposed model introduces a fuzzy logic framework to predict the impact of environmental stimuli, within a pedestrian’s field of view, on movement direction. Attractive and repulsive effects of the surrounding environment are quantified by social force method. A high flow corridor in an office area is considered as a case study. Stochastic simulation is used to generate walking trajectories, calculate a dynamic contour map of environmental stimuli in each step and recognise the high flow walkable areas in the corridor.

Wayfinding behavioural studies have revealed that the spatial ability of pedestrians allows them to find a path from the current location to a destination [1]. During wayfinding activities pedestrians are confront with environmental stimulations that change dynamically after each step. Environmental stimulations have important influence on visually directed walking tasks. However, variable factors such as trip intention and pedestrian’s attributes are contributing elements that make the prediction of pedestrian–environment interactions an imprecise and fuzzy problem.
Dynamical changes of environmental stimulations constantly update the pedestrians’ world view of their surroundings and affect their perception. Understanding a pedestrian’s perception of environmental stimuli is necessary to accurately estimate the pedestrian movement [2]. It is believed that information exchange within a dynamic environment contributes to the control of human locomotion tasks [3]. Golledge et al. [4] elevated the question of information integration of the route attributes. They highlighted that orientation and movement direction are more fuzzy and related to a wide range of elements. Recently, fuzzy rule-based systems have also been successfully applied in the field of robot navigation and path planning [5]. Fuzzy logic has the capability to model the imprecise and diverse nature of pedestrian perception and reaction towards the environmental impacts.
This study highlights two challenging issues: firstly, it addresses how to quantify the environmental effects, and secondly, how the environmental stimuli influence the spatial behaviour. In this context, we have proposed a fuzzy logic approach to model the local steering behaviour. The fuzzy system comprises of three inputs, one output and 216 rules as illustrated in Figure 1. Inputs are the agent’s perception from three possible future travel points, which are described as {Front Position, Right Position, Left Position}. The summation of attractive and repulsive stimulation is assessed for each future position. This perceived information is then recognised as {High attractive, Medium attractive, Low attractive, Low repulsive, Medium repulsive, High repulsive} to express level of perception. Therefore, each input consists of six membership functions and change of movement direction can be inferred by the output of fuzzy system.
To succinctly express pedestrian behaviour, we have made the following five assumptions: (i) all the objects induce attractive or repulsive effects to the surrounding environment, and the cumulative sum of attractive and repulsive stimulations is computed for each point in the terrain surface; (ii) the levels of stimulation in the front, right and left hand side of the agent are the inputs of the fuzzy system; (iii) in each step along the movement path, three alternatives exist (move forward, change direction to the right or to the left); (iv) the agent is able to change the direction between a continuous range from, -12 to +12 degrees, in subsequent steps; and (v) the angular change of direction is the final command for movement driven from fuzzy rules.
The three fundamental elements of the model investigated are psycho-sociological motivations, dynamic environmental information that is provided by objects located in the field of view to reflect local awareness, and the movement direction that is the output of fuzzy logic system.
Simulation of fuzzy steering model with local environmental effects Contradictory psycho-sociological forces motivate the agent to move towards a desired destination. The Helbing social force model is one of the more practical and reliable methods describing the behaviour of pedestrians, which considers the effects of attractive and repulsive forces [6]. We have adopted this approach to quantify the environmental influences. Figure 2 shows the trace of walking paths in the simulated corridor and the potential attractive, repulsive and total stimulation in that area due to the printer and the exit door. It is assumed that the printer has both attractive and repulsive effects, while the exit door provides an attractive influence. Further simulations have been completed to gain a better understanding of algorithm performance and the impact of dynamic changes of environmental stimuli.
The simulation results indicate that the proposed fuzzy-based approach is a promising method to model the pedestrian walking path under normal conditions. In this methodology, the individual-based representation of terrain employs the concept of field of view to capture visual stimuli that impact walking direction and acquire dynamic environmental information, which is a necessary characteristic for local awareness.
To verify the concept, a two dimensional space with walls, printer, entrance and exit was studied. As pedestrian interaction with the environment is an important feature of steering behaviour, we assessed the level of induced effects exerted by the environment on the pedestrian employing social force model to obtain the turning angle of direction for the next step using fuzzy logic framework. In this regard, fuzzy logic offers a framework to model the problem considering both imprecise nature of environmental effects and diversity in pedestrian perception and decision making.