Fuzzy inference is a very important aspect of computer science with numerous applications in expert system and computer vision. This paper has developed fuzzy inference models to reveal traffic conditions and behavior in public road locations. This approach employs some fuzzy logic systems with knowledge-based analysis to define useful linguistic variables and appropriate membership functions for mapping the input parameters to reduce the problem of high traffic congestion. Traffic Pedestrian movement and density have been used as input data to construct models while defuzzification process represents the output, which can be used at pedestrian detection levels with various range in membership values. The traffic movement, density and pedestrian level detection have been used to determine the congestion level or traffic condition of different locations considered in this paper. The results achieved have shown that the calculated traffic density and movement traffic level could be used to control the traffic movement and pedestrian levels. The closer the value of congestion level to 1, the higher the traffic level, however, when congestion level moves closer to zero the traffic level decreases. The experimental results demonstrated that fuzzy inference models could be effectively used for pedestrian detection in transportation systems.

Keywords— Fuzzy Detection, Inference, Pedestrian, Traffic and Defuzzification