In this project we modeled an aerial target moving in space in a defined motion. in addition, we modeled simple ground sensors that measure range, then we simulated range measurements and learned to estimate target location and speed using a linear Kalman Filter. We then implemented an Extended Kalman Filter (EKF) that receives raw range measurements – Tightly coupled (TC), and also one that uses a separate position estimation and then uses the linear Kalman filter – loosely coupled (LC).
We compared performance between the methods using Monte Carlo simulations and found that for the model we chose adding multiple sensors, a wide deployment of sensors and a high target would lead to a more accurate solution. In general, the Extended Kalman Filter solution is more accurate than the other solutions we reviewed, especially for faulty measurements.