Probabilistic Robotics

Probabilistic Robotics introduces a paradigm shift in autonomous systems by using probability theory to handle the inherent uncertainty in robot perception and control. Rather than assuming perfect sensors and deterministic environments, this approach maintains probability distributions over possible states, enabling robots to achieve remarkable robustness in real-world conditions. The book provides a unified mathematical framework based on Bayes' theorem that underlies algorithms from Kalman filters to particle filters, from localization to SLAM.

9 sections ยท 36 lessons

Course outline

Foundations

  1. Why Uncertainty Matters
  2. Probability Basics
  3. Bayes' Theorem
  4. The Bayes Filter
  5. Prediction and Update
  6. Foundations

Gaussian Filters

  1. Kalman Filter Basics
  2. Kalman Filter Equations
  3. Extended Kalman Filter
  4. EKF Implementation
  5. Unscented Kalman Filter
  6. Information Filter
  7. Gaussian Filters

Nonparametric Filters

  1. Histogram Filters
  2. Particle Filter Concept
  3. Particle Filter Algorithm
  4. Resampling Strategies
  5. Adaptive Resampling
  6. Nonparametric Filters

Robot Models

  1. Motion Model Fundamentals
  2. Odometry Motion Model
  3. Beam Sensor Model
  4. Likelihood Field Model
  5. Robot Models

Localization

  1. Localization Problem Types
  2. EKF Localization
  3. Monte Carlo Localization
  4. Adaptive MCL
  5. Localization

Mapping

  1. Occupancy Grid Concept
  2. Binary Bayes Filter for Maps
  3. Mapping

SLAM Fundamentals

  1. The SLAM Problem
  2. Data Association
  3. SLAM Fundamentals

SLAM Algorithms

  1. EKF SLAM Concept
  2. EKF SLAM Limitations
  3. FastSLAM
  4. Graph-Based SLAM
  5. Graph SLAM Front-End and Back-End
  6. SLAM Algorithms

Planning

  1. POMDPs Introduction
  2. Exploration and Information Gain
  3. Practical POMDP Solutions
  4. Planning

Start learning with Wondering