A Fuzzy Logic Modeling Approach to Assess the Speed Limit Suitability in Urban Street Networks - Agents and Artificial Intelligence - page 60

Similarly, Figure 8 illustrates the effects of the 'Posted Speed Limit' and the
'Hourly Pedestrian Volume' (as input variables) on the '85 th Percentile Speed' for
weekday. As shown, the posted speed limit is positively correlated and hourly pede-
strian volume is negatively correlated with the 85 th percentile speed. As can also be
seen, the effect of the posted speed is not quite noticeable if it exceeds 60 km/hr in
cases of high pedestrian volumes.
As shown, the posted speed limit is positively correlated and hourly pedestrian vo-
lume is negatively correlated with the 85 th percentile speed. As can also be seen, the
effect of the posted speed is not quite noticeable if it exceeds 60 km/hr in cases of
high pedestrian volumes.
It can be said that regardless limited number of data, fuzzy logic shows the rela-
tionship between the input and output variables realistically. As fuzzy logic handles
linguistic terms (for a range of numeric values), it is less sensitive to each individual
numeric value. This replicates true human nature about perceiving factors on the
roads. For example, it is clear that drivers' choice of operating speed (represented by
85 th percentile speed) is influenced by the length of the road segment or pedestrian
volume. With larger length, the operating speed tends to be higher. Such changes do
not occur for every one km change of length. In reality, the decision of choosing any
particular range of operating speed tend to be stable for range of length (say between
0 to 1 km). Fuzzy logic predicts such relationship very realistically.
Fig. 8. Effects of 'Posted Speed Limit and Hourly Pedestrian Volume' on the '85 th Percentile
Speed' (weekday model)