Let s Not Throw Away that Big and Bulky Manipulator Revitalize It!

Size: px

Start display at page:

Download "Let s Not Throw Away that Big and Bulky Manipulator Revitalize It!"

Amy Burke
5 years ago
Views:

1 Paper ID #23606 Let s Not Throw Away that Big and Bulky Manipulator Revitalize It! Dr. Trung H Duong, Colorado State University-Pueblo Dr. Trung Duong is currently a Research Professor at Engineering Department, Colorado State University- Pueblo. From 2014 to 2017, Dr. Duong worked as a Post-doctoral Research and then a Research Faculty at Center for Advanced Infrastructure and Transportation (CAIT), Rutgers the State University of New Jersey. He involved in research activities of the Long-Term Bridge Performance (LTBP) program funded by the Federal Highway Administration (FHWA) in U.S. Department of Transportation and the Bridge Resource Program (BRP) funded by the New Jersey Department of Transportation. Dr. Duong earned his M.S and Ph.D. degrees in Mechanical and Aerospace Engineering at the Oklahoma State University. His research interests are mechatronics, robotics, NDE technologies, image processing and computer vision, and artificial intelligence. He is a member of IEEE, IEEE Robotics and Automation Society, and ASEE. Dr. Nebojsa I. Jaksic, Colorado State University, Pueblo NEBOJSA I. JAKSIC earned the Dipl. Ing. degree in electrical engineering from Belgrade University (1984), the M.S. in electrical engineering (1988), the M.S. in industrial engineering (1992), and the Ph.D. in industrial engineering from the Ohio State University (2000). He is currently a Professor at Colorado State University-Pueblo teaching robotics and automation courses. Dr. Jaksic has over 70 publications and holds two patents. Dr. Jaksic s interests include robotics, automation, and nanotechnology engineering education and research. He is a licensed PE in the State of Colorado, a member of ASEE, a senior member of IEEE, and a senior member of SME. c American Society for Engineering Education, 2018

2 Let s not throw away that big and bulky manipulator - Revitalize it! Abstract In this work, we proposed a framework to interfacing the existed manipulators with a modern programming environment. A serial protocol (RS-232) communication between the robot and a computer was developed. Modern and state-of-the-art programming environments, such as MATLAB Robotics Toolbox, or free version of Robotics Toolbox from Peter Corke, etc., can be used to communicate and control the robot. Bypassing the provided and outdated software retired together with the robots by their manufacturers, students now have a greater flexibility to choose programming languages as their preference. Moreover, they can utilize many open source and upto-date libraries to create their own software to solve vast problems ranging from forward/inverse kinematics/dynamics, path planning, to robot vision and learning in a graphical user-friendly environment. This work was implemented at Colorado State University-Pueblo undergraduate engineering program. Students were exposed to both approaches, using provided software from robot providers and following the new proposed framework. The preliminary student survey was conducted and produced a promising feedback. Introduction With the current scarcity of funding, especially for undergraduate student teaching and research, not every department and research center can afford to replace decade-old industrial robots in their labs with brand new and at state-of-the-art ones. Thanks to the industrial-grade standard, the decade-old manipulator's mechanism most likely works fine. However, the electrical control board (hardware), firmware, and software are greatly outdated. Even in industry, most robot arms are supposed to serve for a long period of time because the cost related to upgrading hardware and software is high. In our lab, there are several decade old, industrial-grade robot arms providing undergraduate students with hands-on exercises for mechatronics engineering and industrial engineering curricula. Many PCs that used to communicate with the robots cannot be upgraded from the very old Intel 8088-based PCs, because the interface of the robot controller cards had been abandoned from modern PCs. Moreover, the original software provided with the robots on floppy disks is not only limited in capacity but also user-unfriendly (e.g. DOS-like environment with nothing close to the graphic user interface GUI used in most of current software). Students deserve learning and working with a better system rather than being forced to program these robots via an obsolete programming language. Therefore, in this work, we proposed a framework to interface the existing manipulators with a modern programming environment. The reminder of this paper is organized as follows. Literature review is presented in the next section followed by the curricular context. After that, we introduces hardware setup and then software development. The next section describes the evaluation of students survey followed by the conclusions section.

3 Literature Review Throughout engineering education curriculum in general as well as in our institution, the hands-on experiments and laboratory projects play an essential role in the success of the program. According to Kolb [1], students learn best if they are exposed to a four steps/axes learning cycle/spirals, namely, 1- experiencing (concrete experience), 2- watching (reflective observation), 3- thinking/modeling (abstract conceptualization), and 4- applying/doing (active experimentation). Various engineering education programs, such as mechanical, industrial, manufacturing, and civil engineering, adopted this learning cycle into their curriculum [2] [6]. Many educational institutions have implemented robots of some kind, e.g. industrial robot arms, mobile robots, educational robot kits, etc. to support their science and engineering program [2] [4]. Laboratory exercises and tutorials, educational robotics projects, and open-source software and toolboxes are available and are reported in literature. For example, Corke [7] provides extensive toolboxes for various robotics, vision, and control tasks, which are applicable for both mobile robots and industrial robot arms. The modeling and simulation techniques as well as a vision-based controller of some specific robot arms are introduced in [8] [10]. Curricular Context In our institution, with the two ABET accredited engineering programs, namely Bachelor of Science in Engineering program with Specialization in Mechatronics (BSE-Mechatronics) and Industrial Engineering program, senior students can utilize industrial robots in few classes, such as computer-integrated manufacturing (CIM), robotics, and senior design, as well as undergraduate student research. This work presents the use of an old Mitsubishi robot arm in the CIM class, which is three credit hours and one semester long. At the end of this course, students are expected to have the following knowledge, attitudes or skills. Letters in parentheses indicate how these goals support the program s Student Outcomes as listed in Criterion 3 (ABET): - Demonstrate an understand various concepts used in CIM (a, h, i) - Design and implement small automation projects using digital electronics devices, relays and PLCs (a, b, c, d, e, g, k) - Perform end-of-tool manipulation using robots (a, b, c, d, e, g, i, k) - Successfully program and use a CNC machine (a, b, c, d, e, g, k) - Successfully print a part using a rapid prototyping machine (a, b, c, d, e, g, k) - Develop criteria for the selection, justification, and implementation of selected CIM technologies (c, f, i, k) Among the objectives of this class, students are expected, at the end of the course, to have knowledge and skills to perform end-of-tool manipulation using robots. The lab portion of this class, one credit hour - two hours per week, covers logic controls, programmable logic controllers (PLCs), computer numerical controls (CNCs), robotics, and additive manufacturing (AM). There are five robotics lab sessions. Within robotics labs, student learnt to perform three tasks: manual robot control using a teach pendant, lead-through programming, and computer programming. The manual robot control and lead-through programming session deals with manipulating various small objects. The computer programming task (two week session) is to have the robot write a word (student s name) on an 8 ½ by 11 inch sheet of paper. Students evaluation survey, collected

with the lab reports at the end of lab sections, plays an important role of closing the loop in students experiential learning process. Figure 1: Laboratory setup for the RV-M2 robot.

The robot arm can be controlled manually by a teach pendant or programmatically by a Q-Basic program, which originally run on an Intel 8088 PC.

4 with the lab reports at the end of lab sections, plays an important role of closing the loop in students experiential learning process. Figure 1: Laboratory setup for the RV-M2 robot. Hardware Setup The robot system setup, shown in Fig.1, includes the RV-M2 robot arm, the teach pendant, the control module, and a computer. The robot arm can be controlled manually by a teach pendant or programmatically by a Q-Basic program, which originally run on an Intel 8088 PC. The PC was later on upgraded to Pentium-4 PC with a separated RS232 card which provided two additional serial ports. This PC communicates to the robot control module via a serial protocol, RS232C. The detailed settings of this serial protocol together with the complete command set and the dedicated Input-Output (I/O) timing diagrams in the I/O signal lines are presented in the manufacturer s manual [11]. Figure 2: Mitsubishi s Movemaster RV-M2 robot configuration and dimensions.

The Mitsubishi Movemaster RV-M2 robot has 5 degree of freedom (not including the gripper open-close movement), in which each joint rotation is driven by a DC servo motor.

This manipulator follows a standard vertical and articulated construction with waist, shoulder, elbow, wrist pitch, and wrist roll rotation.

5 The Mitsubishi Movemaster RV-M2 robot has 5 degree of freedom (not including the gripper open-close movement), in which each joint rotation is driven by a DC servo motor. The gripper or handset can be configured as step motor operated or pneumatically operated. This manipulator follows a standard vertical and articulated construction with waist, shoulder, elbow, wrist pitch, and wrist roll rotation. The robot arm is 28 kg in weight and can lift a 2 kg object with ±0.1mm position repeatability at the wrist tool plate. The operation ranges of the joints are: ±150 o waist rotation, -30 o to +100 o shoulder rotation, ±120 o elbow rotation, ±110 o wrist pitch rotation, and ±180 o waist roll rotation. The upper arm is 250mm in length with the offset of 120mm from the base axis. The robot configuration and some key dimensions are shown in Fig. 2. (b) (a) (c) Figure 3: (a) Old software, Q-Basic in DOS-like environment; (b) and (c) New software program using Robotics Toolbox for Matlab. New Software Development A serial protocol (RS-232) communication between the robot and a computer is developed. Modern and state-of-the-art programming environments, such as open source Robot Operation System (ROS) Industrial [12], [13], MATLAB Robotics Toolbox, or free version of Robotics Toolbox from Peter Corke [7], etc., can be used to communicate and control the robot. Notice that the Corke s library is used only for simulation purpose. This library does not has capability to communicate with the particular version of robot (hardware) that we have. Our main contribution is creating the connection between modern programming languages and the old hardware through basic serial port (RS232) protocol (it has been done by our instructors, not students). And

Bypassing the provided and outdated software tired with the robots by their manufacturers, students now have a greater flexibility to choose programming languages as their preference.

6 MATLAB is not the only option, for example, program in ROS and/or C/C++ can be used to communicate with the robot. However, our students only have experience with MATLAB, so they wrote their own programs in this language to control the robot. Bypassing the provided and outdated software tired with the robots by their manufacturers, students now have a greater flexibility to choose programming languages as their preference. Moreover, they can utilize many open source and up-to-date libraries to create their own software to solve vast problems ranging from forward/inverse kinematics/dynamics, path planning, to robot vision and learning in a graphical user-friendly environment. The comparison between the old and the new software development environments is shown in Fig.3. It is anticipated that the students prefer the new software environment with rich GUI tools, easy to plot and visualize, and much more flexibly to create and test the program. With the new approach, it is possible to model the robot according to the standard DH table (Denavit and Hartenberg [14], [15]) representation and then setup and solve the forward/inverse kinematics and dynamics problems. Fig.3.b shows a code snip that print out the DH parameters of the robot into the Matlab s workspace. These parameters follow the standard notation in the Robotics Toolbox for MATLAB from Peter Corke [7]. The units of link offset d and link length a are in mm; the link twist angle alpha is in radians. The visualization of the robot modelled by this toolbox is presented in Fig.3.c. Some examples of student projects that develop computer programs to have the robot write a word (student s name) on an 8 ½ by 11 inch sheet of paper are shown in Fig.4. Evaluation of Student Responses Figure 4: Examples of student projects. The new approach of programming the robot was incorporated into our CIM course syllabus. In the future, this approach is meant to provide resources for students in other classes within our two engineering program curriculums, such as robotics, senior design, as well as undergraduate student research. For the preliminary assessment, a questionnaire were given to our CIM class of 17 students in Fall semester, As discussed earlier, students had five robotics lab sessions, included a two week session of computer programming. They learnt to perform three tasks: manual robot control using a teach pendant, lead-through programming, and computer programming. The tasks deal with manipulating various small objects and control the robot write a word (student s name) on an 8 ½ by 11 inch sheet of paper. For the computer programming tasks, all students used

7 both the approaches (with DOS environment and MATLAB) to control the robot. At the end of the lab, students write lab reports where they include their working programs, their names written by the robot, sections on task challenges and solutions, and sections on self-reflections. The Selfreflections Section is a crucial part of experiential learning. The Likert scale of these questions is from 1 to 5, where 1 means not at all and 5 means very. Figure 5: CIM Lab Students Attitudes and Perceptions Survey. The whole survey questions is presented in Fig.5 and results with standard deviations of the scores are shown in Table 1; while the detailed answers to each question are presented in Fig. 6. Table 1: Survey questionnaire and results Num. Question Average St. dev. 1 How relevant were the CIM labs to your interests as an engineer? How much did you like performing different robotic tasks with the 2 robots in the lab? Computer Technology in the Computer-Integrated Manufacturing Lab: Students Attitudes and Perceptions Survey In the Computer-Integrated Manufacturing (CIM) lab, you experienced a variety of computer technologies including the Disk Operating System (DOS) running the Serpent robot on an computer; the DosBox programming environment running a DOS-based PLC software for Allen Bradley SLC 150 programmable logic controller (PLC) under Windows 7 64-bit operating system, and DosBox with Q-Basic as well as MATLAB to control the lab s Movemaster RV-M2 industrial robot. During this process you learned that the DOS-based computer was easy to boot and did not require a procedure to stop it one could simply turn off the computer by flipping a switch. Then, you learned that one can run Windows 7 non-compatible programs by going through an intermediary programming environment such as DosBox. However, the control of the serial port through DosBox was available only if the program was running with the administrator privileges (which you as individual student users didn t have). Finally, you were able to use MATLAB from your individual university accounts to operate the Movemaster RV-M2 robot. You could develop your design files for the robot on one of many computers having MATLAB installed. Also, the files saved in your account would follow you to other computers as well as to the computer controlling the robot. I am an industrial/mechatronics engineering student. On the scale 1 5 (where 1 means not at all and 5 means very ) please rate the following three questions. 1. How relevant were the CIM labs to your interests as an engineer? How much did you like performing different robotic tasks with the robots in the lab? 3. How much did the different computer technologies distract/impede you from the programming tasks? Please comment on the various computer technologies used in the lab (DOS, DosBox, Win 7 with MATLAB, etc.). What is it that you liked/didn t like, appreciated, etc.? 5. What other technologies would you like to explore/experience in CIM labs? 6. Please provide any other comments on computer technologies in CIM labs. How much did the different computer technologies distract/impede you from the programming tasks?

8 The students responses clearly show the interest in the subject in general and in programming the robot in particular. Likert scale: 1 not at all, 5 very Figure 6: Student responses to the survey. The open-ended Question 4 was designed to obtain students feedback on the familiarity and access of the lab s computer technologies. Students understand the simplicity of DOS, but they preferred the familiarity of Windows and MATLAB. One group of student indicated the major drawback of the old technology that related to the accessibility. We have only one working industrial robot for the class. Most importantly, there is only one computer in CIM lab that support DosBox with administrative privileges (required to control the robot). Hence, students had to arrange the lab time to take turn to access the robot. Some typical comments include: It would be nice if it was easier to work with the computer technologies outside of lab time. I was just frustrated by the fact that there is only one computer that can communicate to the hardware to complete the lab. It made the overall wait much longer than it really should have been. Question 5 asked students what they would like to experience in the lab. Students answers included statements like spend more time on programming robots in the lab, buy more robots, provide more complicated robots like dexterous robots, have image processing robotic related lab, and teach us how to use your humanoid robots. It was surprising that a couple of students wanted more DosBox instructions. For Question 6, most of the students reiterated that they loved the lab and even wanted to spend more time programming more complicated robotic tasks. Some of them wanted to include the robot in their senior projects, and one of them provided an advice to make sure that industrial engineering and mechatronics students partner in all labs. In this class, the sections on students self-reflections were written individually, even though their lab sessions and lab reports were done in pairs. The self-reflections focused on the question: what did I like about this lab. Some students indicated that:

9 I loved that I was able to use Matlab code to make the robot do what I wanted. I really enjoyed this lab because I like learning how to program the robot to be able to do anything I want it to. It was also really cool to watch the robot spell my initials successfully! This lab was very inspiring as I was able to work with different robot arms in a very hands-on environment. One problem I had was that it took way too long to plot all 119 points in the cursive form of Mike. It wasn t super challenging; it was just fun. It was definitely a welcome break from some of the other rigors of this program. It was really entertaining to use the robot arm to stack blocks in this lab. It felt like a game rather than like school. This lab was one of the most fun labs that I have experienced here at Its simplicity and ease of MatLab made it an enjoyable lab, and also easy to fix. As the testimonials shown, students enjoyed the old robot in both manually control (by teach pendant) and computer programming task (by MATLAB). The age of the robot did not pose any problem. Student actually would like to have more programming (and more complex) tasks. As the result, students showed interest in the new software programming paradigm mainly because of their familiarity of Windows and MATLAB. In addition, it increases the accessibility of the system, for example students can create programs ahead of time on available computers outside the lab instead of waiting in line as before. Many students loved the lab and did stay after hours to program more complicate tasks than required, such as shown in Fig. 4. Some of them plan to have their senior design projects be robotics-related. Conclusion In this work, a framework for interfacing the existing manipulators with a modern programming environment via serial protocols (RS-232) was developed. State-of-the-art programming environments, such as MATLAB Robotics Toolbox or free version of Robotics Toolbox from Peter Corke, etc., can be used to communicate and control the robot. By replacing the obsolete software, i.e. DOS-like environment, as well as upgrading the hardware, from an old Intel 8088 PC to a Pentium-based PC running MATLAB on Window 7 64 bit system, we provide students with a better environment to study and perform research. After the first time implementation of this new approach in a CIM course for one semester, the students surveys showed the strong interest in the subject in general especially the programming robot tasks. They were concerned less about the old robot hardware, but enjoyed programming the robot through MATLAB. We believed that this old robot, provided with our new software compatibility, is still capable of providing students a decent source to improve their robot programming skill.

10 References [1] D. A. Kolb, Experiential learning : experience as the source of learning and development. Englewood Cliffs: Prentice-Hall, [2] J. L. Newcomer, An Industrial Robotics Course for Manufacturing Engineers, 2016 ASEE Annu. Conf. Expo. Proc., Jun [3] S. Das, S. A. Yost, and M. Krishnan, Development Initiative : Relevance, Content, A 10-Year Mechatronics Curriculum Development Initiative : Relevance, Content, and Results Part I, IEEE Trans. Educ., vol. 53, no. 2, pp , [4] N. I. Jaksic and B. Li, Humans vs. Robots Workout Challenge, in 124th ASEE Annual Conference and Exposition, 2017, vol June. [5] M. Muscat and P. Mollicone, Using Kolb s Learning Cycle to Enhance the Teaching and Learning of Mechanics of Materials, Int. J. Mech. Eng. Educ., vol. 40, no. 1, pp , Jan [6] F. A. Candelas, S. T. Puente, F. Torres, F. G. Ortiz, P. Gil, and J. Pomares, A virtual laboratory for teaching robotics, Complexity, vol. 1, p. 11, [7] P. Corke, Robotics, Vision & Control, vol. 73. Springer, [8] H. Zhao, Z. Lu, C. Liu, and H. Wang, Model and Simulation of the Mitsubishi RV-M1 Robot using MATLAB, in 2016 IEEE International Conference on Signal and Image Processing (ICSIP), 2016, pp [9] R. Čermák, Web Camera-based Control of a Mitsubishi MELFA Robotic Arm with MATLAB Tools, [10] J. Swider, K. Foit, G. Wszołek, and D. Mastrowski, The off-line programming and simulation software for the Mitsubishi Movemaster RV-M1 robot, J. Achiev. Mater. Manuf. Eng., vol. 20, no. July, pp , [11] Industrial Micro-Robot System Model RV-M2 Instruction Manual. Mitsubishi Electric Corporation. [12] L. Joseph, Mastering ROS for Robotics Programming. Packt Publishing, [13] M. R. Product, M. Manager, and H. Kong, Interfacing MATLAB and ROS, [14] R. S. Hartenberg and J. Denavit, Kinematic synthesis of linkages. McGraw-Hill, [15] Denavit and R. Hartenberg, A kinematic notation for lower-pair mechanisms based on matrices., Trans. ASME. J. Appl. Mech., vol. 22, 1955.

Humans vs. Robots Workout Challenge

Paper ID #18986 Humans vs. Robots Workout Challenge Prof. Nebojsa I. Jaksic, Colorado State University, Pueblo NEBOJSA I. JAKSIC earned the Dipl. Ing. degree in electrical engineering from Belgrade University