ELEC-E8111 - Autonomous Mobile Robots P, 26.02.2019-03.04.2019
This course space end date is set to 03.04.2019 Search Courses: ELEC-E8111
Topic outline
-
-
Course introduction and arrangements
Introduction to mobile robots
• What is a service robot? Field Robot?
• Which kinds of service and field robots exist?
– Examples
• Mobile robot subsystems
• Control architectures
• More on Field robotics
• Autonomous vehicles -
Robot control system architectures
•Motivation and state-of-the-art issues
•Short introduction to control paradigms
•The Hierarchical Paradigm
•Biological foundations for the reactive paradigm
•The Reactive Paradigm
•The Hybrid Paradigm
•AuRA Architecture (example of Hybrid solutions)
-
Introduction to design cases and lab work arrangements
- Design of a case robot system
- Testing algorithms under ROS on Pioneer
robot platform
Introduction to ROS
- Overview
- Main components
- How does ROS communication work
- ROS Tools Helpful to know
Locomotion, Kinematics, and Low level Motion Control
- Locomotion – Principles and mechanisms to make the robot move.
- Kinematics – How to model motion of (rigid) bodies?
- Motion control – How to make a robot attain a goal or follow a trajectory?
-
Sensing and Perception Characterizing sensors
- Characterizing sensors
- Classification of sensors
- Sensor types for mobile robots
- Perception in mobile robotics
-
Localization, mapping and SLAM = Simultaneous Localization and Mapping, idea
- Main material as subset of ETH slides
- Kalman-filter based traditional vehicle localization in the tractor-trailer -case.Extended Kalman Filter SLAM (EKF SLAM)
Save videos for running. A video haehnel-RawOdometry-anim.avi demonstrates mapping with raw odometry localization - Just mapping, not SLAM. I odometry errors accumulate quickly too large.
A video haehnel-ScanMatching-anim.avi demonstrates first real SLAM, localization on the basis of laser scan matching, only one estimate for pose, Mapping works
file slam.pdf, EKF SLAM video be Nebot can be run in the 4_EKF_SLAM 2018.ppt -slide set. -
Particle filter
Particle filters are introduced in particle-filters 2018.pdf.
Fast SLAM
Particle distribution for robot pose, for each robot pose particle, separate EKFs for positions of landmarks. Own realization of grid type maps
Fast SLAM: You can run videos in the fastslam 2018.ppt -slide set. pdf available, too.
Some Fast SLAM videos added as separate files. In order to run videos, save them.
bruceton-one-loop.avi ; fastslam-dmb-fastslam.avi ; haehnel-ScanMatchingFastSLAM-Seattle-anim-2.avi
-
Graph SLAM
– Graph construction and optimization
– Case forest SLAM
– Advanced topics -
Satellite positioning – GNSS technologies
● GNSS (Global Navigation Satellite System)
– Applications
– Constellations: GPS, Galileo, GLONASS, Beidou, QZSS
● GPS architecture and system
● Multi-sensor data input
● Position determination
● Error sources
● GNSS performance
● GNSS shortcomings
● Practicals for coding (formats, libs, links) -
Graph-based Path Planning: Graphs to the Rescue!!
By Dr Kshitij Tiwari
Overview
Biography
Motivation
About
Graph-based Path Planners- Scenario
- Graphs
- Best-first Search Methods
Dijkstra Algorithm
A* Algorithm
D* AlgorithmDijkstra Algorithm Algorithm Sampling Based Methods
RRT
PRM
Cliff Hanger
Readings -
Informative Path Planning: Information to the Rescue!!
By Kshitij TiwariOverview
Recap
Informative Path Planning (IPP)
Information
Path Planning
Challenges
Summary
Readings
ROS TutorialCoverage Path Planning (CPP)
by Dr TiwariTravelling Salesman
Lawnmower
Piano Mover
Art Gallery
Watchman Route
Orienteering
Random Exploration
Frontier-based Exploration
Adaptive Voronoi ExplorationPose Estimation, Inertial Navigation Systems (INS), Inertial Measurement Unit (IMU)
By AV
• Introduction
• Mathematics of Inertial Navigation
• Errors and Aiding in Inertial Navigation
• Example: Simple Odometry Aided AHRS
• Use of cheap MEMS IMUs in robotics
• Summary -
Visual localization and object recognition
Perception for Mobile Robots
Machine Vision
Stereo Vision
Filtering, Edges, and Point-features
Corner Detection
Blob features
Place Recognition, Line Extraction
-
Outdoor Motion Planning and Control
Control of mobile robot
1. Robot Trajectory Following
2. Perception Based Control
3. Steering Trajectory Generation
4. Optimal and Model Predictive Control
5. Intelligent ControlExample: NMPC in autonomous driving of tractor and implement
The journal article is additional material and details are not required in the exam.
-