Compliant mechanisms, characterized by monolithic structures utilizing motion derived from the deflection of flexible parts, known as compliant joints, represent a paradigm shift in mechanism synthesis. This shift introduces numerous challenges, especially for learning institutions, since its early adoption may benefit students' competences towards novel technologies in addition to strength problem solving abilities. A promising path for teaching compliant mechanisms synthesis is through experimentation, this poses numerous challenges, foremost among them being the need for accurate methods for validation that incorporate both rigid-body and flexible-body kinematics. The validation of such mechanisms poses a key challenge, particularly in the context of non-planar mechanisms. This work addresses this challenge by proposing an algorithm grounded in visual object tracking and stereo vision for the kinematic validation of a novel 3D compliant gripper mechanism. The experimental setup employs two cost-effective commercial USB cameras for stereo vision, three 3D-printed gripper prototypes with varying material properties, and an actuation system. The motion data obtained is validated against finite element simulations. The comparison with the experimental validation has an error rate of less than 10% which indicates the efficiency of the tracking system created through this paper. This implementation holds promise as an inexpensive tool for teaching state-of-the-art kinematic models in a hands-on fashion.

Detecting the pose of an student in an educational workspace that has heavy machinery or dangerous hazards is a safety precaution that in automated safeguard protocols is not yet applied. To address this gap, we propose a novel solution that combines the Intel® RealSense™ Depth Camera D435 with the YOLOv8 algorithm for object detection. The utilization of YOLOv8, is a technology which allows for rapid detection of objects of interest, making it ideal for maintaining low response times. Additionally, the integration of stereo camera technology enables the precise identification of human positions within an XY divided space. The convolutional neural network architecture formulated by the algorithm enables the ability to identify and classify human beings on an image with a confidence level superior to 50 %. Once this is done, the infrared sensors in the camera are utilized to deliver an estimate of the distance in meters from the camera to the median location of the person in the picture. With the use of triangulation performed by the stereoscopic vision with a reference marker, one can proceed to generate an approximation of the individual's relative coordinates in the closed space where the task is monitored. This technology empowers students to proactively assess their proximity to restricted or hazardous areas in educationallabora-tories, enhancing their safety awareness. By avoiding potentially dangerous spaces, students increase their adherence to safety protocols, leading to improved work efficiency and productivity within real workshop settings.

In the age of swift technological advancements, precise object classification and arrangement-termed “bin picking” -is essential. The challenge revolves around accurately identifying and positioning diverse objects in terms of color, size, and orientation within simulated environments. This work delves into a comparative analysis of two distinct methodologies: Seman-tic Segmentation through YOLO integrated with Convolutional Neural Networks (CNN), and a stereo camera-centric approach amalgamating segmentation with pose estimation. These method-ologies find their application within Python frameworks in a virtual environment. To ascertain the efficiency of each method, evaluation metrics like the precision recall curve, mean average precision 50–95 and euclidean distance are employed. The main objective of this paper is to quantitatively measure the gain of specific competencies and skills in postgraduate students resulting from the development of two bin-picking algorithms in the field of computer vision. These competencies and skills, such as advanced programming proficiency, problem-solving awareness, and in-depth knowledge of computer vision, will be assessed through pre- and post-development evaluations that consist of self-assessment surveys. Additionally, we aim to highlight the educational impact of acquiring these competencies and skills, demonstrating how they can enhance student's abilities and understanding, ultimately benefiting their academic and professional pursuits. The study focuses on segmentation and pose estimation using synthetic scenarios. Results with YOLO show 100% precision and recall rates due to a well defined testing set, with negligible translation errors for most shapes. Stereo Vision assessment compares estimated centroids with ground truth values. Emphasis is placed on prioritizing pose estimation in the XY plane for bin picking purposes, indicating potential for improvement despite notable distances achieved between centroids. Participant competency development reveals significant growth in relevant skills. Overall, the study highlights the efficacy of synthetic data in deep learning, suggests further research for pose estimation algorithms, and provides valuable insights into stereo vision development. In educational contexts, bin picking extends beyond technical knowledge, providing students with a distinctive perspective on how machines perceive and engage with the world. This research, rooted in the fundamental principles of both algorithms, not only enriches knowledge dissemination but also enhances learner's understanding of automation and its practical applications.

Teaching Cyber-Physical Systems (CPS) can be challenging due to the involvement of diverse, complex, and expensive devices. During the past year, Tecnologico de Monterrey together with Manchester Robotics Ltd. successfully implemented a teaching strategy for CPS under its model TEC21 on the way to graduate high-level students, and ready to join the workforce. Theoretical lectures combined with hands-on practices allowed students to solve a real problem using off-the-shelf mobile robots and ROS as the software platform. After implementing this teaching-learning strategy, significant improvements have been observed in students' learning outcomes. They have had the opportunity to work collaboratively in teams, develop their research skills, enhance their programming abilities, and gain a deeper understanding of mobile and manipulator arm robotics through practical implementation in the laboratory. The planning, methodology, and solutions performed by students are used to showcase the results of the teaching strategy and its outcomes.

Reinforcement Learning (RL) is an area of Machine Learning (ML) that takes care of what decisions are better to make with no prior information; it creates its own datasets via reward shaping. It has taken importance in the last years in the industrial field since it is a significant tool to make decisions under uncertainty. However, there is a lack of research in RL projects that contribute to develop innovative education in the professional education context. Therefore, in this work, a possible contribution of the AWS DeepRacer models to the autonomous driving for handicapped people is briefly presented in a theoretical way. Besides, through RL, three time-trial models were developed using Amazon Web Services (AWS) (AWS DeepRacer and AWS Console). These three models were developed for the first stage of the study, where there are no obstacles for the car. The resultant performance of each model is presented and discussed. Lastly, for future work three extra stages of the work are proposed: field tests of the presented models, static obstacles reward function design, and dynamic obstacles reward function development.

This work aims to present a roadmap towards the Cyber-Physical implementation at Tecnologico de Monterrey. This document describes the approach to conceptualize, design, and implement a mixed virtual environment that fulfills the Paradigm of Industry 4.0. This implementation has the following benefits: 1) The Cyber-Learning-Factory that emulates a real, controlled, Industry 4.0 workplace, 2) It allows the students to design a solution to manufacture and assembly a product from a realistic product specification 3) It allows the students to play similar roles to those found in industry positions. This concept was validated on a project to build a product that integrates autonomous, industrial and collaborative robots, Product and Process Digital Twins, flexible CNC machinery and Additive Manufacturing, Simulation stations and CAD/CAM/CAE applications, Horizontal and Vertical Integration (MES and ERP), Augmented-Virtual and Mixed Reality aids and all the IT infrastructure and applications supporting IoT, Cyber Security, Cloud Computing and Big Data & Analytics.

The work presented in this article shows the results of the implementation of the lean manufacturing methodology using the 5s and 7 wastes tools to improve the learning process for engineering undergraduates in the cyber-physical systems course for higher education students. The cyber-physical systems course is taught in the Smart Factory, a space that recreates a genuine factory in a controlled environment where students play the roles of real-life industry. We have observed that the lack of order in the workspace affects the overall performance of the students. This educational innovation project implements the lean manufacturing methodology in the learning environment providing students tools that will, not only positively impact their education but, prepare them with means to better perform in industry.

Industry 4.0 has revolutionized how factories operate, integrating new advanced technologies. This progress must be matched with innovation in engineering education, as these new technologies demand engineers with multidisciplinary skills. Microservices architecture offers a chance to implement project-based learning courses that cater to an individual's skills while exposing them to other disciplines' concepts and methods. This paper explores the implementation of such a project while implementing a security camera system for a Smart Factory, serving as a showcase of the strengths that project-based learning offers. Three microservices were successfully developed, with another three microservices pending further development. Project-based learning offered multidisciplinary learning opportunities during the construction of the services.

This paper proposes an approach to attract students to the study of Vision Systems to detect transparent materials, such as glass or plastic, through the comparison of videogame technologies, e.g. the Kinect Sensor, and more complex systems, such as Stereo Vision. Although the field of computer vision has advanced enormously in recent years, there are still problems when it comes to viewing transparent materials. Technologies such as the Kinect Sensor have been widely implemented in research before, but they lack the capabilities to detect transparent materials as well. A good alternative is the use of Stereo Cameras that can triangulate points from pictures with two perspectives. However, given that the Kinect is used in videogames and for hobbies, it represents a great area of opportunity to attract its intended public into research activities, as it could act as a catalyst to foster curiosity, mainly undergraduate students, into how technology intended for videogames can be applied for something as important as Computer Vision.

Teaching robotics demands the use of hands-on tools that enable and empower the skills and competencies acquaintance needed to successfully implement robot-based solutions. Real robots are expensive and prone to damage in the event of miss usage, a powerful solution is the use of simulators, ROS (Robot Operating System) which is an open-source software that is being viewed as a potential solution for tackling new and more complicated production challenges. This paper presents various projects that are aimed to give students high-quality courses in robotics and industrial control thanks to recent technical advancements, so they can work in simulated environments, and once that has been processed has finished, it can be rapidly implemented in real manufacturing processes. The robots may vary depending on the use they will have in the production area; this research shows the use of Autonomous Mobile Robots (AMRs) and Articulated Robotic Arms with numerous joints and articulated robotic arms that are used to move and lift objects in our facility. We used the UFACTORY xArm 6, which offers a good performance of 6 axis robot arms, and the DASHGO-B1 robot which is capable of indoor mapping, localization, and autonomous navigation. The suggested approach connects this arm robot and the mobile robot to PLCs (Programmable Logic Controller), which are industrial computer control systems that continually monitor input devices. This research’s possible use might strengthen educational approaches for teaching robotics and industrial control, as well as save costs and improve factory performance. This research shows that it is feasible to deploy numerous robots with industrial equipment to improve performance.

Using Deepracer, through experimentation and simulation, theoretical concepts can be applied to a practical application of reinforcement learning (RL) in a real-life problem. It was considered as a highly useful tool to develop many direct and transversal competences that students need to work within the field of artificial intelligence (AI). Nowadays the combination of Hardware, Computer Vision methods and Machine Learning (ML) algorithms for the development of controllers for vehicle driving automation have facilitated the development of solutions for this problem. The intention of this work is to show a Roadmap that was formulated to learn AI, ML and RL competencies required to prepare undergraduate students for the industry of this area, following a structured mostly practical learning plan using Deepracer AWS platform and local alternatives for training; and a physical vehicle as primary tools that have made an incredibly compact setup process and reduced complexity in the educational researching field related to learning autonomous vehicles (AV) software development process. The AWS DeepRacer framework includes all needed hardware based on a two front-camera vehicle for stereo vision and a LiDAR sensor, besides it provides a powerful computation computer for high performance in scale autonomous vehicles. The roadmap was executed testing and comparing different RL models over the modalities that Deepracer provides (by comparing both local and AWS console training) documentation of the execution of the generated Roadmap is shown to ensure that should be considered as a stable learning system that could be followed by college programs. Finally, models were tested on a physical track built, and limitations, considerations and improvements for this Roadmap are explained as a contribution for future work.

The combination of games, competitions around the world, and mobile robots cause motivation in students for learning while acquiring practical skills. However, the global pandemic has forced social and physical distancing measures, while the need for distance learning has made the teaching of robotics difficult. The intention of this work is the adaptation of the Very Small Size Soccer (VSSS) competition to be used as a distance learning strategy of robotics, following most of the rules of the original competition. The proposal consists in using free-available software and libraries for 1) modeling a non-holonomic wheeled mobile robot, 2) designing the graphical appearance, and 3) implementing different intelligent algorithms for computer vision, path planning, and decision-making. The main contribution of this work is the general methodology and development of a complete virtual system based on the VSSS, using only open and free software tools easily reachable by students. The fundamentals of mobile robotics (locomotion, perception, cognition, and navigation) have been successfully covered during the development of the system, while significant learning has been achieved by students.

This paper presents online simulation along with physical environments employing digital twins to develop robotic and computer vision applications, such as bin picking, by providing a virtual testing environment with real conditions for robotic systems and their surroundings. Particularly for situations such as the current COVID-19 pandemic, this approach permits the integration of knowledge and practice remotely without having the need of requiring a physical robot, increasing equity and inclusivity in the assessment, by reducing the gap between students from different institutions due to the lack of resources to buy the required robotic devices and even enabling remote connection to the classrooms. The proposed task can be performed as long as the educational center counts with basic computer equipment and internet connection. This paper presents a bin picker application based on mono and stereo vision to make the classification of cylinders of different colors and sizes and their arrangement in a base by using image processing and depth estimation algorithms. The performance of the classification was measured in terms of the error in millimeters from the difference of the desired and the obtained position. The images used for classification were taken in real life considering the environment conditions, and a digital twin of the bin picker and objects was made. The image classification and the overall instructions that the robot performed were generated on MATLAB, where the data from the real environment captured by cameras was sent. From here, the instructions were sent to URsim, which serves as the robot controller, and the final virtual simulation with the models of the base and cylinders was performed on Kuka-Sim Pro. This approach can then be utilized to implement different assessments based on personal results from simulation activities. As this assessment requires specialized knowledge regarding robotics, computer vision and classification algorithms, this methodology is intended to be applied for university students, at the level that is adequate according to the educational programs of each university.

The aim of this work is to present an open-source training system based on Virtual Reality and the Collaborative Universal Robot UR5e. This system uses the virtual controller URSim (Universal Robots Simulator) in conjunction with an open-source platform programed to overcome the lack of access to the robots during lockdowns or similar situations. The Virtual Reality platform was programed on Unity, it contains a model of the UR5e robot, all links were joined using the forward Kinematics obtained in house and then compared to that given by the manufacturer. For the inverse kinematics, the manufacturer kinematics calculations and path planning algorithms were used. As a result, a Virtual Reality environment for the kinematic chain was developed, the model can be fully joint jogged joint and linearly jogged using the manufacturer URSim virtual controller.