rlabbe/kalman-and-bayesian-filters-in-python

This is an interactive textbook that teaches you how Kalman filters and Bayesian filters work, using Python code you can run right in your browser. Kalman filters solve a very practical problem: sensors are noisy, and the real world is unpredictable, so how do you get the best possible estimate of what's actually happening? Think of a GPS that gives slightly different altitude readings each time you pass the same point, or a drone that needs to know its exact position despite wind gusts and sensor errors. Kalman filters blend imperfect sensor readings with a model of how a system behaves to produce the most accurate estimate possible. The book covers the standard Kalman filter plus more advanced variants, extended Kalman filters, unscented Kalman filters, and particle filters, all explained with plain language and working code rather than dense academic math. The goal is to build intuition through hands-on experimentation, not just formal proofs. You can tweak the code, run the simulations, and see exactly how changing parameters changes the results. You would use this when you're working on any project that involves tracking something over time using noisy sensor data: robotics, drones, computer vision tracking, IoT devices, or anything dealing with GPS or motion. It's written in Python, presented as Jupyter Notebooks (interactive code documents), and all exercises include solutions. Used internally at SpaceX to teach state estimation concepts.