Skip to content

Latest commit

 

History

History
41 lines (25 loc) · 1.35 KB

README.md

File metadata and controls

41 lines (25 loc) · 1.35 KB
Raw

Sensor-Fusion-and-Nonlinear-Filtering

Repository for the course "Sensor Fusion and Nonlinear Filtering"

Final Assignment: Orientation estimation of a cellphone using EKF and Sensor Fusion

In this project accelerometer, magnetometer and gyroscope measurements were used in a sensor fusion fashion to estimate the orientation of a cellphone.

This project includes:

  • Derivation of the motion model using quaternions representation
  • Derivation of measurement models
  • Implementation of EKF subject to outlier rejection and sensor failure
  • Detailed analysis of the implemented EKF

Assignment 1: Gaussian distribution analysis

  • Transformation of Gaussian random variables
  • Analysis of joint distribution and posterior densities
  • MMSE and MAP estimates for Gaussian mixture posteriors

Assingment 2: Linear Kalman filter analysis, Motion and Measurement models

  • A first Kalman filter and its properties
  • Study of different Motion and Measurement models
  • Kalman filter and its tuning

Assignment 3: Nonlinear Kalman filter models (EKF, UKF, CKF)

  • Approximations of mean and covariance
  • Non-linear Kalman filtering
  • Tuning non-linear filters

Assignment 4: Smoothing, Particle Filter and Localization

  • Non-linear RTS smoothers
  • Particle filters for linear and Gaussian systems
  • Localization: bicycle tracking in a village using Particle Filters