CSE 398/498 Introduction to Arm-Type Robots, Professor Jeff Trinkle, Spring 2021
This course will cover fundamental concepts, mathematics, and algorithms most relevant to serial chain robots, i.e., legs, arms, spines, fingers, and toes. The main topics to be studied are position and orientation representations, forward and inverse kinematic maps, coordinated linkage control, basic legged locomotion and grasping. Semester-long projects will be done in interdisciplinary teams of three students. These projects will incorporate most of the basic topics above, but will also offer the opportunity to explore topics of interest to students that are beyond the main topics. Assignments and labs will be done using virtual robots in simulation, but the possibility of using physical robots is be examined.
The 498 (graduate) version of this class will differ from the 398 (undergraduate) version by one or more of several possible extensions designed by the student with the approval of the instructor. These extensions include, but are not limited to, extra homework assignments, a more extensive project, an in-class presentation of a topic of useful results from the literature.
Prerequisites: basic programming ability (e.g., CSE 007 or CSE 012) and permission of the instructor. Students who are interested in taking the course should send a paragraph of what they’d hope to get from the course.
ECE 350/450 – Introduction to Robotic Perception and Computer Vision, Professor Yahong Rosa Zheng, Spring 2021
This course will introduce the basic framework of modern robotics and then focus on the sensing and perception aspect. We will study IMU, camera, lidar sensors and their roles in robotic perception. We will cover sensing feature extraction, sensor information fusion, computer vision, Artificial Intelligence (AI) or Deep Learning (DL) algorithms. If time allows, the course will also cover simultaneous localization and mapping (SLAM), as well as exploration and learning aspects. This course will utilize an f1/10 model car (see f1tenth.org for details) and the NVIDIA JETSON TX2 embedded GPU which allow the students to work on 5-6 hardware and programming projects and multiple car races.
CSE 298 – Foundations of Robotics, TR 4:00-6:50, Professor Corey Montella, SUMMER SESSION II (6/29/20-8/6/20)
Introduces students to the field of robotics, covering foundational mathematics and physics as well as important algorithms and tools. Topics include simulation, kinematics, control, machine learning, and probabilistic inference. The mathematical basis of each area will be covered, followed by practical application to common robotics tasks. This course is designed to be taught remotely using simulated robot platforms and sensors. Pre-requisites: (CSE 002 and (CSE 001 or CSE 012 or ENGR 010)
CSE 360/460-010 – Introduction to Mobile Robotics, TR 2:05-3:20, Professor David Saldaña, FALL 2020
Algorithms employed in mobile robotics for navigation, sensing, and estimation. Common sensor systems, motion planning, robust estimation, bayesian estimation techniques, Kalman and Particle filters, localization and mapping.
CSE 371/CSE 471 – Principles of Mobile Computing, MW 10:45-12:00, Professor Mooi Choo Chuah, FALL 2020
Lecture/seminar course covering the fundamental concepts and technology underlying mobile computing and its application as well as current research in these areas. Examples drawn from a variety of application domains such as health monitoring, energy management, commerce, and travel. Issues of system efficiency will be studied, including efficient handling of large data such as images and effective use of cloud storage. Research coverage will be drawn from the best publications in the recent research conferences. Deep learning methods will be covered. Students will do Android programming and possibly develop Alexa/Google home skills for homework assignments and final class projects. Prerequisites: CSE 109 and (CSE 202 or ECE 201).
ME 450 – Special Topics: Formal Methods in Robotics, MW 3:00-4:15, Professor Cristian-Ioan Vasile, FALL 2020
The course is an introduction to formal methods with emphasis on robotics applications. The fields of formal methods has its roots in computer science, where problems of correctness of programs, circuits, and processors was the initial driver. Safety and expressive specifications are important in the design of controllers for dynamical systems (such as robots) as well. The aim is to develop computational frameworks that take rich temporal and logic specifications and automatically construct or certify robot behaviors (controllers). The course covers formal specification languages (regular expressions, linear temporal logic, signal temporal logic, and time window temporal logic), automatic controller synthesis and formal verification methods. Methods based on abstractions, simulation/bisimulation, refinement, automata, mathematical programming, and satisfiability modulo theories are presented. These are complemented and combined with algorithms from control theory and motion planning commonly employed in robotics. Recently, these frameworks have been employed to ensure the safety and to increase the capabilities of learning algorithms. Some topics on inference of temporal logic properties, and reinforcement learning under safety and temporal logic constraints may be explored based on student interest and available time. The material is grounded in examples involving aerial and ground vehicles, manipulators, and self-driving cars. Course evaluation is based homework that involves theoretical exercises and work in Robotic Operating System (ROS), car simulator CARLA, and traffic simulator SUMO, and a final project. Final projects may span the students’ topics of research and interest in robotics, but must integrate some aspect of safety, formal specification, automatic synthesis, and explainability. Progress is tracked and marked based on technical reports associated with the projects.
Required background: Python, C, C++, or Matlab; Linux; basic notions of graph theory; basic notions of dynamical systems and control; differential equations;
Recommended background: control theory; graph theory; build toolchain; ROS; git; latex;
Primary Textbook: “Formal Methods for Discrete-Time Dynamical Systems” by Calin Belta, Boyan Yordanov, and Ebru Aydin Gol (Springer, doi: 10.1007/978-3-319-50763-7).
Reference Books: “Introduction to Automata Theory, Languages, and Computation” by Hopcroft, John E.; Motwani, Rajeev; Ullman, Jeffrey D. (Pearson, ISBN 978-1292039053), “Principles of Model Checking” by Christel Baier and Joost-Pieter Katoen (MIT press, ISBN: 9780262026499), and “Planning Algorithms” by Steven M. LaValle (Cambridge University Press, ISBN-13: 9780521862059).
ME 450 – Special Topics: Robot Motion Planning and Control, MW 1:35-2:50, Professor Subhrajit Bhattacharya, FALL 2020, Website
This course will start with an introduction to the configuration spaces and kinematics of different robotic systems, including holonomic & non-holonomic mobile robots, spatial robots, and robotic manipulators. Following that basic motion planning algorithms, including potential & navigation function-based motion planning and graph search based motion planning, will be introduced. Sensor-based motion planning and motion planning under uncertainties using probabilistic representations will be introduced. Students will learn about estimation and filtering (Kalman filter, Markov filter, particle filters) and probabilistic robot action models (Markov chains, Markov decision processes, POMDP). Students will get hands-on experience in implementing the algorithms on MATLAB/C++. Application to multi-robot coordination problems, multi-robot coverage problems, pursuit-evasion problems, task allocation problems and exploration problems will be discussed. If time permits, students will be briefly introduced to topological motion planning, motion planning on manifolds and motion planning on flow fields. The evaluation will be based on two term projects and a final presentation.
ME 433 – Linear Systems and Control, MW 10:45-12:00, Professor Nader Motee, FALL 2020
This course covers the following topics in linear systems and control theory: review of fundamental concepts in linear algebra, state-space representation of linear systems, linearization, time-variance and linearity properties of systems, impulse response, transfer functions and their state-space representations, solution to LTI and LTV state equations, Jordan form, Lyapunov stability, input-output stability, controllability, stabilizability, observability, detectability, Canonical forms, minimal realizations, introduction to optimal control theory, Linear Quadratic Regulator (LQR), Algebraic Riccati Equation (ARE), frequency domain properties of LQR controllers.
ME 450 – Special Topics: Introduction to Nonlinear Dynamics and Chaos, TT 3:00-4:15, Professor Nader Motee, FALL 2020