Bus arrival time is very
important to passengers, not only at the origin terminal but also at every
stop. If there is no predicted arrival time at a stop, the headway designed
should match the bus frequency at the stop. The uncertainty in bus arrival time
can hinder the headway from matching the bus frequency at the stops. Moreover,
lack of information on the service route and actual arrival times at stops
leads to difficulty for passengers in planning their trips. Observation surveys
were conducted to collect data on the problems of bus arrival frequency and
uncertain arrival times at a selected stop with multiple routes during off-peak
hours in Putrajaya Malaysia. This paper proposes a method to estimate arrival
times at bus stops using the adaptive neuro fuzzy inference system (ANFIS) and
several models are proposed to predict arrival times using MATLAB Curve Fitting
Tool. All the proposed models exhibited RMSE close to 0 and R2 close
to 1.
