Welcome to the Autonomous Robotic Systems page of the Swarmlab, Maastricht University.
In this new course (based on the previous Situated Agents, and Autonomous Systems courses), we will be using the Robot Operating System (www.ros.org) from Willow Garage, and our Turtlebots with Kinect RGBD cameras and Hokuyo laser range finders in simulation. ROS is supported by a large and growing robotics community.
Course Description
This course introduces the students to the foundation of situated autonomous robots. The course will start with an introduction to the field of mobile robots. Concepts such as recursive state estimation and Kalman filtering will be introduced. Additionally basic questions of how to effectively control a mobile robot will be addressed. The core of this course will address the problems of localization, planning and control, perception and robot motion and navigation. Planning and control will be approached from a probabilistic perspective in the form of Markov Decision Processes, and Partial Observable Markov Decision processes. Robot navigation on the other hand will be approached from an evolutionary perspective where the focus will be on swarm intelligence, genetic algorithms, and neural networks. This course will be accompanied by a large practical part in which students have the opportunity to apply the learned material in practice. After completing this course, students will have a good understanding of the major concepts in autonomous systems such as localization, planning and control. The student will be able to apply the learned concepts to real autonomous systems.
Knowledge and insight:
Key questions addressed are: “What is robotics?”, “How does a robot knows where it is?”, “How does a robot knows where it is going?”, “How does a robot knows how to get to its goal?”, “How can robots evolve and learn under extreme uncertain conditions?”, “How can robots cooperate to solve a complex task?”.
Skills: The course will be accompanied by hands-on exercises necessary to attain a good understanding of the introduced concepts. Available platforms to work with are Turtlebots, e-puck, and quadcopter.
Study material:
a) Thrun et al. (2005), Probabilistic Robotics, The MIT press. (Course textbook)
b) Floreano and Nolfi (2000), Evolutionary Robotics, The MIT press.
c) Floreano and Mattussi (2008), Bio-Inspired Artificial Intelligence, The MIT press.
d) Bekey (2005), Autonomous Robots, The MIT press.
Recommended literature: Additional papers can be found on Student Portal, on the Syllabus page, and will be announced during lectures.
Examination: Written exam.
ECTS: 6.
Prerequisites
Discrete Mathematics (KEN1130), Linear Algebra (KEN1410), Probability and Statistics (KEN2130), Data Structures and Algorithms (KEN1420).
Teaching Staff
Dr. Rico Möckel
MSc. Kirill Tumanov
Assignments
There will be several robot programming assignments. They will be aimed to test your ability to take material presented in the lectures and translate it into practice on a simulated mobile robot. The robot control architecture we will use is called ROS (www.ros.org), which is an open-source system. A separate talk will be devoted to the explanation of the tools related to the course assignments.
We try to specify the tasks a way which is as clear as possible. Usually there is an example code that illustrates part of the task. If there is any doubt about the intent of an assignment, you are urged to communicate your concerns as soon as possible.
Grading
Grades are based on performance in the projects, intermediate test results and on the final exam.