 |
| Title of Course |
Intelligent Robotics |
| Credits |
2 |
| Target Students |
graduate student |
| Course start date |
1Q |
| Day and Time |
Tuesday 3rd period, Friday 3rd period |
| Teachers in Charge |
Wang Shuoyu |
| Contacts |
Teacher's Room：A480， Extension No.: 2306 |
| Keywords |
Robots, robotic systems, intelligent robots, knowledge, learning, reasoning, AI, specifically neural networks, reinforcement learning, genetic algorithms, human reasoning, welfare care robots, life support robots |
| Objectives of the class |
Specifically, ① basic robot mechanisms, ② robot system configuration, and ③ neural networks and reinforcement learning and genetic algorithm learning methods and from the standpoint of knowledge acquisition,④ From the standpoint of knowledge use, you can learn about human reasoning methods based on fuzzy sets, which provide a quantitative representation of ambiguous concepts from reasoning in binary logic, and ⑤ how to realize intelligent robotic systems using these methods. |
| Lesson Objectives |
Unlike industrial robots, which can only perform simple, predetermined tasks, intelligent robots can autonomously reason about and execute actions in response to different environments.
Reasoning about appropriate actions requires knowledge. Knowledge is acquired through learning. In other words, to construct an intelligent robot, one must equip it with reasoning and learning functions in addition to mechanical mechanisms.
In intelligent robotics, students first learn the basic mechanisms and system configuration of robots, as well as typical intelligent methods, such as learning and inference algorithms.
|
| Method of teaching |
In this lecture, in order to enhance understanding, emphasis will be placed on the meaning of computation, i.e., physical concepts and the background of algorithm development, while valuing theoretical computational algorithms. In order to deepen understanding, the students will experience inference and learning algorithms using freeware to understand how intelligent robots work. In addition, state-of-the-art intelligent robots will be introduced through the use of numerous photographs and videos.
|
| Achievement Goals |
①You will be able to understand the basic mechanisms of robot systems, system configuration, and concepts of robot intelligence.
②This course provides basic concepts for knowledge representation using neural networks, backpropagation learning algorithms, and robots.
③As an unsupervised learning method, this provides the learning mechanism of reinforcement learning and the basis for making robots more intelligent.
④By understanding the learning mechanism of genetic algorithms, which acquire knowledge by searching for optimal states, we can gain a foundation for making robots more intelligent.
⑤Basic concepts of inference, inference methods in binary theory, flexibility of knowledge representation using production rules, expert systems.
⑥You will be able to understand how to quantify vague concepts using fuzzy sets and perform basic operations on fuzzy sets and relationships.
⑦You will be able to understand the basic concepts of fuzzy reasoning and express three types of fuzzy reasoning algorithms mathematically.
⑧This enables students to understand the relationship between knowledge acquisition, learning, and inference, and to create language rules for obstacle avoidance in intelligent robots.
|
| Lesson plans |
1. Explain the basic mechanism of a robot and the system configuration of a robot.
2. We will provide an overview of various machine learning methods, explain the difference between supervised and unsupervised learning, and the basic concepts when applying them to intelligent robots.
3. Explain the advantages and disadvantages of knowledge representation using neural networks.
4. Explain how to give a neural network a learning function by introducing the backpropagation learning method.
5. Explain the basic concepts of reinforcement learning and Q-learning, a representative method.
6. Explain the learning algorithm using genetic algorithms.
7. The advantages and disadvantages of each learning method are theoretically explained and demonstrated using examples of intelligent robots.
8. Summary of the first half, exercises, and proficiency check
9. Basic concepts of inference, inference methods in binary theory, flexibility of knowledge representation using production rules.
10. Fuzzy theory and intelligent control
Introduction of the concept of fuzziness and its importance, examples of intelligent control systems that utilize the theory, and intelligent robots.
11. Fuzzy sets and quantification of vague concepts
Sets, crisp sets and defining functions, fuzzy sets and their properties, basic operations.
12. Fuzzy numbers and their operations
Definition of fuzzy numbers, extension principle, decomposition theorem, expression of intelligence in language, action planning of intelligent robots based on fuzzy inference
13. Fuzzy Inference I
Mamdani's reasoning method, application example to autonomous control of robots Second half of the exam 60 points (basic content 50 points, application problem 10 points)
14. Fuzzy Inference II
Concept of distance, calculation of distance between sets, properties of distance-based fuzzy inference method
15. Action planning for intelligent robots based on fuzzy reasoning
Relationship between knowledge acquisition, learning and inference, trajectory and path planning for intelligent robots, obstacle avoidance based on inference methods with learning capabilities.
|
| Texts |
『Introduction to Machine Learning』，Tomohiro Kodaka（Ohmsha）ISBN：9784274068461
『Fuzzy Control』，Michio Kanno（The Nikkan Kogyo Shimbun）ISBN：4526023485
|
| Reference books |
『Robot Intelligence: Evolutionary Computation and Reinforcement Learning』，Kazuyuki Ito（Ohmsha）ISBN：9784274203985
『Advanced Fuzzy Control』，Kazuo Tanaka（Kyoritsu Publishing Co., Ltd.）ISBN：978-4-320-08530
『Knowledge Representation and Fast Reasoning』，Mitsuru Ishizuka（Maruzen）ISBN：4621042068 |
| Grading methods and standards |
Mid-term exam achievement level: 20 points (Basic content 40 points, advanced questions 10 points）
Midterm Test: 50 points (Basic content 40 points, advanced questions 10 points)
AA: 90 points or more
A: 80 points or more and 89 points or less
B: 70 points or more and 79 points or less
C: 60 points or more and 69 points or less
F: 59 points or less
|
Study outside of class hours (preparation, review, etc.)
| It is recommended that you prepare in advance using the distributed textbooks and reference books. |
Relationship with other subjects
| We recommend taking Robotics I and Robotics II. |
|
 |
|
|
| Title of Course |
Robotics I |
| Credits |
2 |
| Target Students |
Second-year Students |
| Course start date |
4Q |
| Day and Time |
Tuesday 2rd period, Friday 2rd period |
| Teacher in Charge |
Wang Shuoyu |
| Contacts |
Teacher's Room: A480， Extension No.: 2306 |
| Keywords |
Robots, wheeled robots, bipedal robots, quadrupedal robots, industrial robots, medical and health promotion robots,Welfare and nursing robots, human-friendly robots, sensors, robot arms and hands, manipulators, position, posture, joints, kinematics, dynamics, control, learning control
|
| Objectives of the class |
Students can learn basic knowledge about mechanisms, dynamics, measurement, and intelligent control for wheeled, walking, and manipulators. |
| Lesson Objectives |
Students will understand basic concepts of control qualitatively using video such as active suspension control for automobiles and animation of intelligent robot control, and then Laplace Transform method is explained, and based on them, learn about transient response analysis and stability criterion of control systems.As practical applications, you can also understand cleaning robots, rescue robots, soccer robots, educational robots, guide robots, mental care robots, welfare care robots, rehabilitation and health promotion robots, etc.The word "robot" is so well known that there is probably no one who doesn't know it. Robots are made from machines, so they have a certain shape. If they have a shape, they are visible and easy to understand.
Robots are also fun to watch because they can move intelligently. Robotics is what supports these robots behind the scenes.
Robotics I covers wheeled, manipulator, and walking robots, and introduces everything from basic knowledge to practical applications. |
| Method of teaching |
The content of the text will be supplemented as necessary as the course progresses, with numerous examples, video, and videos being used to focus on explaining the physical meaning of complex calculations.
Each time, we will start by reviewing the content of the previous session and end by summarizing the main points of today's session. |
| Achievement Goals |
① Define a robot in your own words.
② You can organize the types of robots.
③ Able to explain the mechanisms of mobile robots and manipulators
④ Understand the kinematics and dynamics of mobile robots and manipulators.
⑤ You will be able to understand the steps to make robots intelligent.
|
| Lesson plans |
1. What is a Robot?
History of robots, origin of the word robot, three laws of robotics, classification by mechanism, classification by application field
2. to 4. Wheeled robots
Classification of wheeled robots, Structure and arrangement of wheels, steering and casters, Special mechanisms, Omnidirectional vehicles, Static dynamics and running of wheeled robots, Dynamics of wheeled robots, Speed ??control, Practical applications
3. Progress of Control Engineering
History of Control Engineering, Classification of Automatic Control, Controls in Future, Future Control-Intelligent Robot
4. Open-loop Control and Closed-loop Control
Advantages and Disadvantages of Open-loop Control, Advantages and Disadvantages of Closed-loop Control, Components of Feedback Control Systems, Negative Feedback and Positive Feedback
5.-6. Mathematical Basis of Control Theory Ⅰ
Necessity of Mathematic Models, Dynamic System and Static System, Concept of Transient Response, Step Response and Impulse Response, Definition and Properties of Laplace Transform, Calculation of Laplace Transform of Exponential Function and Unit Step Function
7.Exercises
Exercises for 1. to 6.
8.Midterm Test
Range: 1. to 7.
9.-10. Mathematical Basis of Control TheoryⅡ
Usage of Laplace Transform, Input/Output, Transfer Function, Block Diagram, Block Diagram Transformation
10.Modeling of Control Objects
Transfer functions of control system components (Proportional, Derivative and Integral, First-order Lag, Second-order Lag, Dead-time) through examples.
11.Response of Control System
Necessity of Dynamic Response. The way to obtain a transient response when a unit step function is input to a system in which a transfer function is a derivative element, a first-order lag element, and a second-order lag element.
12.-13. Stability of Control System
Concepts of Stability and Instability, Poles of Characteristic Equation and Transfer Function, Criterion of Stability by Positions of Poles, Stabilization by Feedback.
14.Exercises
Exercises for 9. to 13.
15.Regular test
Range: 9. to 14. |
| Texts |
『Basic Robotics』，Koichi Ogawa/Ryozo Kato（Tokyo Denki University Press）ISBN：4501414103
|
| Reference books |
『Basic Robotics』，Sadao Kawamura（Ohmsha）ISBN：4274130355
『Robotics of Atom the Iron Plan』，Toshio Fukuda（collectors' editorship）ISBN-13:978-408
|
| Grading methods and standards |
A score of 60 or more overall is required to pass the examination. The distribution of points is as follows.
Mid-term exam achievement level: 20 points (Basic content 40 points, advanced questions 10 points）
Midterm Test: 50 points (Basic content 40 points, advanced questions 10 points)
AA: 90 points or more
A: 80 points or more and 89 points or less
B: 70 points or more and 79 points or less
C: 60 points or more and 69 points or less
F: 59 points or less
|
Study outside of class hours (preparation, review, etc.)
| It is recommended that you prepare in advance using the distributed textbooks and reference books. |
| Relationship with other subjects |
A basic knowledge of mechanics and control will help you to understand it more deeply. |
|
| |
|
|
| Title of Course |
Robotics Ⅱ |
| Credits |
2 |
| Target Students |
Third-year Students |
| Course start date |
2Q |
| Day and Time |
concentration |
| Teacher in Charge |
Wang Shuoyu |
| Contacts |
Teacher's Room：A480， Extension No.：2306 |
| Keywords |
Autonomous robots, probability and statistics, self-location estimation, environmental mapping |
| Overview of the class |
This class will be given as an intensive lecture by Prof. Katsuhiro Hori (Tomakomai National College of Technology), a part-time lecturer.
It has been nearly 20 years since cleaning robots based on mine-clearing algorithms were introduced to the world.
In recent years, AI-powered cleaning robots have become common. One of the reasons for this is that robots are now able to take information from sensors as probabilistic, process it statistically, and choose which action to take or not to take based on an integrated approach.
The field of “probabilistic robotics” addresses these issues. This is an unavoidable situation for studying robotics in the future.
In this lecture, we aim to understand “Stochastic Robotics” and learn how to control an autonomous robot in a real environment.
The content of the lecture is based on the textbook: “Stochastic Robotics” authored by Dr. Chiba Takumi Ohta (recommended book for the JSME Educational Award in 2020). The contents are roughly divided into the following four categories.
◎How can we use knowledge of probability and statistics in the control of autonomous robots?
◎How to model an autonomous robot for simulation in a real environment
◎How can self-location estimation and environmental mapping be achieved?
◎How can we realize action decisions of autonomous robots in real environments?
|
| Objectives of the class |
In Robotics 1, students have learned the basics of the mechanisms and movements of robots that are used in various situations. Through these experiences, we believe that we have gained a bird's-eye view of the system structure of robots and the technologies required for them. In Robotics II, we will focus on self-position estimation and environmental mapping, which are a kind of environment-resistance technology commonly used in cleaning robots, drones, automatic driving, etc., to understand “probabilistic robotics,” which is how to make autonomous robots operate in a rapidly changing surrounding environment. |
| Method of teaching |
Lectures will be given using PowerPoint. Lecture materials and exercises will be presented on the web. The Jupyter Notebook environment is used for the implementation of the algorithms in the exercises.
|
| Achievement Goals |
①Understand the basics of probabilistic robotics and probability and statistics and be able to explain the concepts.
②Understand autonomous robot modeling and be able to explain and implement concepts and algorithms.
③Understand self-location estimation and environmental mapping, and be able to explain and implement concepts and algorithms.
④Understand the behavioral decisions of autonomous robots and be able to explain and implement concepts and algorithms.
|
| Lesson plans |
1st Introduction
To learn about the concept and history of probabilistic robotics and to get an overview of this lecture.
2nd-4th Basics of Probability and Statistics
Probability and statistics, which are the basis of data processing for robots, are studied: probability models of data, Bayes' theorem to estimate the cause from the result, and multidimensional Gaussian distribution.
5th-6th Modeling of Autonomous Robots
Students learn modeling of an autonomous robot equipped with sensors. In addition, students learn how to take into account the uncertainty.
7th-9th Self-location estimation
Students will learn self-position estimation by particle filter and extended Kalman filter. Also, understand various problems of self-position estimation.
10th-12th SLAM
Understand the concept of SLAM, which simultaneously performs self-location estimation and environmental map construction, and learn about sequential SLAM and graph-based SLAM.
13th-15th Action Decision
Students will learn about dynamic programming methods that are effective for Markov decision processes in robot action decision making. The students will also learn about the fundamentals of reinforcement learning.
16th Credit Approval Examination
|
| Texts |
『Detailed Explanation of Probabilistic Robotics: Implementation of Basic Algorithms in Python』，Ryuichi Ueda（Kodansha）ISBN：978-4-06-517006
|
| Reference books |
『Probabilistic ROBOTICS(Intelligent Robotics and Autonomous Agents series 』，Sebastian Thrun, Wolfram Burgard and Dieter Fox(The MIT Press) ISBN：9780262201629
『Probability Robotics (Premium Books Edition)』，Sebastian Thrun, Wolfram Burgard and Dieter Fox （Mynavi Coming to work）ISBN：978-4-8399-5298
|
| Grading methods and standards |
Grading will be based on a comprehensive evaluation of the students' performance in the credit examinations and exercise assignments.
The ratio of evaluation is 60% for the credit examination and 40% for the exercises, and the passing score is 60 points or higher.
AA: 90 points or more
A: 80 points or more and 89 points or less
B: 70 points or more and 79 points or less
C: 60 points or more and 69 points or less
F: Less than 60 or grading materials not submitted
|
Study outside of class hours (preparation, review, etc.)
| Lecture materials and exercises will be posted on the web, and students are expected to work on them sequentially. |
| Relationship with other subjects |
The knowledge learned in Robotics 1 will be the foundation of the course. In addition, since Python is used to implement the algorithms, it is desirable that students have already acquired basic knowledge of Python programming.
|
|
| |
|
|
| Title of Course |
Modern Control |
| Credits |
2 |
| Target Students |
Second-year Students |
| Course start date |
3Q |
| Day and Time |
Monday 2nd period,Thursday 2nd period |
| Teacher in Charge |
Wang shuoyu |
| Contacts |
Teacher's Room：A480， Extension No.：2306 |
| KeyWords |
Control, state variables, equation of state, controllability, observability, equilibrium point, stable and asymptotically stable, pole assignment, Riccati equation, optimal control |
| Overview of the class |
In Modern Control II, students will be able to use the equation of state to quantitatively understand controllability, observability, stability discrimination methods, pole assignment methods, and optimal control methods. |
| Objectives of the class |
In the category of mechanical engineering, control means to put a machine, device, etc. in the desired operating condition.
In a broader sense, it means to subdue others and control them to do what you want. Humans have a desire to have everything the way they want it or the way they think it should be.
For example, television channels, room temperatures from heating and cooling systems, driving cars, elevator rises and falls, physical and mental health, human organization, economic systems, environmental systems, and so on.
In real life, however, common sense tells us that everything cannot be controlled as we wish. This hinges on the controllability of the system.
Modern Control provides an understanding of analytical and design methods for linear systems, using equations of state. |
| Method of teaching |
The course will focus on the contents of the textbook, but will utilize real-life examples, videos and animations to explain the physical meaning of abstract calculations with emphasis on the physical meaning of the calculations.
Each lecture will begin with a review of the previous lecture and end with a summary of the main points learned. |
| Achievement Goals |
①Explain basic concepts such as stability, controllability, and observability in your own words.
② Can determine the stability, controllability, and observability of a control system.
③ The equation of state can be established from a physical model represented by a second-order differential equation.
④ Understand the properties of Laplace transforms and be able to use them.
⑤ The poles can be freely positioned using constant feedback.
⑥ Basic calculations of matrices and determinants can be performed.
|
| Lesson plans |
1. concept of modern control theory
To explain the history of development of control engineering, and to explain the concept, characteristics, and practical applications of modern control.
2. Mathematics Fundamentals I
Matrices and vectors, addition and multiplication of matrices, determinants, inverses.
3. Basic Mathematics II
Eigenvalues, eigenvectors, matrix rank, positive definiteness, matrix exponential function.
4.-5. Equation of state
Discriminant Methods for Controllability and Observability.
6.-7. Controllability and observability
Discriminant method for controllability, Discriminant method for observability.
8. exercises and mid-term test
A mid-term exam will be given after the exercises in Sections 1 through 7.
9.-10. Stability
Equilibrium points, stability and asymptotic stability, stability of linear systems, Hurvitz's stability discriminant, Lyapunov's method, application of Lyapunov's method to linear systems
11.-12.Design method of control system by pole arrangement
Concept of regulator, poles (eigenvalues) of regulator, necessary and sufficient conditions for pole placement,
method of constructing control system by pole placement method
13.-14. Design method of control systems by optimal control
Basic concept of optimal, evaluation function and optimal control, stability of optimal control system, solution of Riccati equation
15. exercises
Exercises on 9. to 14.
16. final examinationRegular Test
An overall summary and final exam will be given.
|
| Texts |
『Modern Control Engineering - From Basics to Applications』，adashi Egami, Takeshi Tsuchiya（Sangyo Shobo）ISBN: 9784782855584
|
| Reference books |
『Basic Robotics』，Sadao Kawamura（Ohmsha）ISBN：4274130355
『Robotics of Atom the Iron Plan』，Toshio Fukuda（collectors' editorship）ISBN-13:978-408
|
| Grading methods and standards |
A score of 60 or more overall will be considered passing. The distribution of points is as follows
Mid-term exam 20 points for quiz ([2 points / 1 time] x 10 times = 20 points)
Mid-term exam: 30 points (breakdown: 20 points for basic questions, 10 points for advanced questions)
Routine examination: 50 points (breakdown: 40 points for basic questions, 10 points for advanced questions)
AA: 90 points or more
A: 80 points or more and 89 points or less
B: 70 points or more and 79 points or less
C: 60 points or more and 69 points or less
F: 59 points or less
|
Study outside of class hours (preparation, review, etc.)
| Preparation and preparation are strongly recommended, using the textbooks distributed in advance as well as reference books. |
| Relationship with other subjects |
Mastery of linear algebra and control fundamentals is required for deeper understanding. |
|
| |
|
|
| Title of Course |
Systems Engineering Experiment |
| Credits |
2 |
| Target Students |
Second-year Students |
| Course start date |
2学期 |
| Day and Time |
Thursday 4 ～ 5nd period |
| Teacher in Charge |
Wang Shuoyu, Orange Masayoshi, Yukio Miyama, Noda Satoshi, Shibata Kyoko, Ouchi Masahiro, Nakamura Takashi, Makino Hisao, Kusaka Kusaka, Nanokawa Mitsumi, Shiba Tatsuya |
| Contacts |
(Overall coordination): Masahiro Ouchi |
| Keywords |
|
| Objectives of the class |
Experiments that form the basis of the major area of study are provided in the mechanical, electronics, and construction fields, respectively.
The entire school year is divided into three groups, and through rotation, experiments in all systems are conducted four times each.
*This course is primarily a practical education course, including work experience under the guidance of companies and organizations with a wealth of practical experience. |
| Lesson Objectives |
Attendance check attendance points and submission of reportsIn studying each major area of the Systems Science Group, experiencing and understanding phenomena through experiments will help students gain a deeper understanding of theory.
In addition, in order to be active in research and development work in the future, students must have the skills to conduct highly accurate and reproducible experiments, to organize and analyze the results of experiments in an easy-to-understand manner, and to report the contents and results of experiments in an easy-to-understand manner.
In the Systems Engineering Laboratory, students will acquire the basic experimental skills and the ability to compile easy-to-understand reports in their respective fields of specialization. |
| Method of teaching |
【Confirmation of attendance and submission of attendance score and reports】
Attendance points will be 25 points for attending all experiments and submitting all required assignments, and no partial points will be given.
If a report is not submitted for an experiment in each system, the score for that system will be zero.
Since each laboratory does not have an IC card reader, as a rule, the teacher in charge or TA will call roll at the beginning of class. Those who are not present at the time of attendance check will be considered tardy, and those who are substantially tardy will be considered absent. If a student is absent without notice, he/she is considered to have abandoned the course and will fail the class.
If you are absent for unavoidable reasons, submit a notice of absence to the Educational Affairs, consult with your instructor to schedule the experiment on a different date, and submit a report by the designated date.
【Report Submission】
Students are required to prepare reports according to the instructions of the instructor in charge of each theme and submit them by the due date.
【Text】
Distribute as needed in each system. |
| Achievement Goals |
① Acquire the background, basic principles, and fundamental laws of a given experimental task.
② To be able to cooperate with others in a group by assigning roles and responsibilities, and to be able to carry out tasks to the best of one's ability.
③ Compose a clear and concise report of what was actually done.
④ The experiments involve various hazards, so safety should be taken into consideration.
⑤ Find a specialty that matches your interests. |
| Lesson plans |
Orientation to the first lecture
Term 1 (Lecture 2,3,4,5,6)
Term 2 (Lectures 7,8,9,10,11)
Term 3 (Lectures 12,13,14,15,16)
The students are divided into three groups and take mechanical, electronic, and construction experiments in each term by rotation.
|
| Texts |
|
| Reference books |
|
| Grading methods and standards |
【Grading】
Attendance points (25 points (including submission of reports) + 25 points for the content of each system x 3 = 100 points maximum. The content points for each system will be evaluated based on the student's attitude, attitude toward the experiment, preparation, and the content of the report. If the total number of points is less than
Less than 50 points: F
51-69 points: C
70-79 points: B
80 points or more: A
AA may be given to those with a grade of “A” in cases of very exceptional excellence.
A student will receive a failing grade (F) if he or she is absent for more than one unexcused absence or if no assignments have been submitted.
|
Study outside of class hours (preparation, review, etc.)
| 【Preliminary study】
In order to work efficiently on the day of the experiment and to improve mastery of the content, sufficient preparation should be made by means of a bookmark of the experiment, etc.
【Review】
Follow the instructions and prepare the report carefully. This is important in order to acquire skills that will be needed in the future, such as report formatting, significant figures, and graph writing. |
|
| |
|
|
