State Estimation and Localization for an Autonomous car (Sensor Fusion)

  • Implemented Error-state Extended Kalman filter to localize the Vehicle using Python
  • The filter relied on IMU data to propagate the state forward in time, and position updates from GPS and LIDAR to correct the state estimate
  • It was simulated in the Unreal Engine’s Carla Simulator
  • Examined the effects of sensor miscalibration on the vehicle pose estimates & explored the effects of sensor dropout