Skip to Main Content
Particle filter is a Monte Carlo simulation method that designed to approximate nonlinear problem and tracks the state of the dynamic system. Over the past decade, particle filters have been successfully applied in wide variety of applications. By using geographical information system (GIS) database, particle filter applications involve tracking of underwater (UW) vessels, surface ships, cars and aircraft. Particle filter also contributed in robotics community, where the algorithm can solve the simultaneous localization and mapping (SLAM) problem. Since multiple targets can be tracked from a video stream using particle filter, visual tracking is another interesting application of particle filter. However, particle filter is rarely used in the flood water level prediction and tracking applications. This paper proposes flood water level prediction and tracking using Sampling Importance Resampling (SIR) particle filter which is one of particle filter variations. As the effectiveness of the particle filter depends on the parameters stated in the algorithm, the parameters are varied to analyze the performance result. The parameters are number of particles and number of time step. From the simulation, the performance result of particle filter is quite impressive to look into.