Which filter is commonly used to fuse GNSS and INS measurements in navigation?

• A. Extended Kalman Filter (EKF).
• B. Particle Filter
• C. Finite Difference Method
• D. Fast Fourier Transform

Answer: (D)

In sensor fusion for navigation, the most practical and widely used approach is a state estimator that can handle nonlinear motion and measurement models. The Extended Kalman Filter is the go-to for fusing INS data with GNSS. It keeps a state that includes position, velocity, and sensor biases, then uses the high-rate INS measurements to predict how the state evolves. When GNSS data arrive, it performs an update step that adjusts the predicted state toward the GNSS measurements, weighting by the uncertainties of both sensors. This combination corrects the INS drift while preserving the high-frequency information from the inertial sensors, giving a real-time, robust estimate.

The EKF is favored because it effectively handles the nonlinear relationships between motion and measurements by linearizing around the current estimate, which is computationally efficient for real-time navigation. While particle filters can handle more complex non-Gaussian scenarios, they’re typically more computationally intensive. Finite Difference Methods and Fast Fourier Transforms aren’t designed for fusing time-series navigation measurements, so they don’t serve as the standard fusion tool here.