Keywords: pedestrian modelling, cellular automata, self driven particles, corners.

For the planning of large pedestrian facilities, the movement of pedestrians in various situations has to be modelled. This paper investigates the characteristics of some common models of pedestrian behaviour for the simple cases of 90° and 180° bends in a constant width corridor. The 90° bend is quite common in office buildings and schools, while the 180° bend simulates the landings of a staircase. While traditional hand calculation methods [1] have no special treatment for the bends, in cellular automata [2] they may come out as a strong obstacle, depending on the details.

For cellular automata, the static floor field determines the guiding of pedestrians through the rooms. The standard guiding field is based on the shortest distance to the exit either in the Manhattan metric or in the Euclidian metric. We have investigated the effect of the bends for both, as well as for newly designed floor fields that are based mostly on keeping in lane and have only a week tendency to switch to a shorter lane. With shortest distance guiding any corner is an obstacle and only the innermost two or three lanes get used behind the corner. Before the first corner, the inner lanes are blocked by persons coming in from the outer lanes. With this behaviour, wide floors seem to be useless. The egress capacity does not grow much with width beyond ~2m. The lane keeping floor field, on the other hand, guides people smoothly round the corner using the full width of the corridor with only a mild tendency towards the inner lane, and the egress capacity grows almost linear with corridor width.

With the social force model [3], the most widely used example of a self driven particle model, the walkers are guided by intermediate destination lines. It is shown that many such lines in the corner with small angles between them give a smoother walking round the corner and a higher calculated capacity than two lines with a 90° angle. However, the calculated value is unrealistically low in all situations, so that the model in the present form is not capable of predicting egress rates.

1

P.J. DiNenno, "SFPE Handbook of Fire Protection Engineering", National Fire Protection Association, Washington D.C., 2002.

2

A. Kirchner, A. Schadschneider, "Simulation of evacuation processes using a bionics-inspired cellular automaton model for pedestrian dynamics", Physica A, 312, 260-276, 2002. doi:10.1016/S0378-4371(02)00857-9

3

D. Helbing, "Verkehrsdynamik. Neue physikalische Modellierungskonzepte", Springer, Berlin, 1977.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £78 +P&P)