Particle Filters (and Navigation)

As the final course in the Applied Kalman Filtering specialization, you will learn how to develop the particle filter for solving strongly nonlinear state-estimation problems. You will learn about the Monte-Carlo integration and the importance density. You will see how to derive the sequential importance sampling method to estimate the posterior probability density function of a system’s state. You will encounter the degeneracy problem for this method and learn how to solve it via resampling. You will learn how to implement a robust particle-filter in Octave code and will apply it to an indoor-navigation problem.

• A brute-force solution for highly nonlinear systems
• This week, you will learn a computationally intensive method to estimate the state of highly nonlinear systems, where the pdfs do not need to be Gaussian.
• How to approximate multidimensional integrals efficiently
• This week, you will learn the tricks we will use to approximate the brute-force solution.
• Developing and refining the particle-filter algorithm
• This week, you will put all of the tricks from week two together to implement (and then refine) the particle-filter method.
• Navigation application using a particle filter
• This week, you will learn how to apply the particle filter to an indoor navigation problem.