State Estimation and Filtering for Robotic Perception

How does an autonomous aircraft determine its velocity vector and position while taking into consideration GPS, airspeed, and IMU measurements? How does an autonomous car map its surrounding environment and determine its position relative to that environment with noisy data including camera and Lidar measurements? These are examples of a fundamental problem in engineering: state estimation. In this course you will study algorithms that are used to determine the state of a dynamical system over time while filtering out erroneous measurements. By exploring examples from robotics, including state estimation for drones, SLAM for autonomous cars and mobile robots, and attitude estimation for autonomous spacecraft, you will learn how to use filters to solve mathematical problems. You will develop an understanding of state estimation in the larger context of Bayesian estimation, which is relevant to a range of topics in machine learning, artificial intelligence, and signal processing.

Topics Include

  • Dynamical systems and probability review
  • Linear-Gaussian systems
  • Kalman and Bayesian filtering
  • Nonlinear architectures
  • Iterated filters and optimization
  • Rigid body dynamics

Course Page
$4,200.00 Subject to change
Online, instructor-led
10 weeks, 10-20 hrs/week
Robotics and Autonomous Systems Graduate Certificate
Stanford School of Engineering