GPS/IMU Sensor Fusion: Vehicle Localization Drift

Smart mobility depends on a vehicle knowing exactly where it is, but no single sensor gets there alone. This simulator drives a car around a closed loop while three estimates of its position race each other in 3D: the ground-truth path, a raw IMU dead-reckoning trail that integrates noisy acceleration and drifts steadily off course, and a Kalman-filtered estimate that fuses the same noisy IMU with periodic, noisy GPS fixes to stay locked near the true path. Tune the GPS fix rate, GPS noise, IMU noise and vehicle speed, or trigger a simulated tunnel to cut GPS entirely and watch the filter's uncertainty ring inflate during the blackout and collapse the instant a fix returns — the same predict/update cycle running inside every autonomous-vehicle localization stack.