Building Bots, Solving Problems: A New Approach to Robotics Education

by

Author: Denis Avetisyan

A project-based learning framework combining robotics and agile methodologies empowers students to tackle real-world challenges like automated disassembly and contribute to a circular economy.

The project’s progression unfolds across defined phases, each representing a stage in its inevitable evolution from inception to completion-a natural arc mirroring the decay inherent in all constructed systems.
The project’s progression unfolds across defined phases, each representing a stage in its inevitable evolution from inception to completion-a natural arc mirroring the decay inherent in all constructed systems.

This paper details an educational framework leveraging Scrum and ROS to provide practical experience in automation and robotics for sustainable manufacturing.

Traditional robotics education often prioritizes theoretical knowledge over practical application, leaving students unprepared for industry challenges. This paper, ‘Bots and Blocks: Presenting a project-based approach for robotics education’, details an innovative framework employing agile project management and a semester-long robotics project to bridge this gap. The core of the approach centers on students developing a disassembly software ecosystem for hardware robots, utilizing the Robot Operating System (ROS) and contributing to concepts like automated disassembly for a circular economy. Could this immersive, project-based model represent a scalable solution for fostering the next generation of robotics engineers and addressing real-world automation needs?

The Inevitable Shift: From Extraction to Circulation

The prevailing linear economic model – often described as “take, make, dispose” – presents a fundamental sustainability challenge by relentlessly extracting resources, converting them into products, and ultimately discarding those products as waste. This system not only depletes finite natural resources at an accelerating rate but also generates substantial environmental pollution through manufacturing processes and landfill accumulation. Consequently, critical materials are lost, valuable energy is wasted, and ecosystems are increasingly stressed. The sheer volume of waste produced globally underscores the unsustainability of this approach, prompting a growing recognition of the urgent need for systemic change and alternative economic frameworks focused on resource conservation and circularity.

Shifting towards a Circular Economy fundamentally requires reimagining how products are designed, used, and ultimately, dealt with at the end of their service life. Traditional ‘take-make-dispose’ models create substantial waste streams and rely heavily on virgin resource extraction; a circular approach prioritizes keeping materials in use for as long as possible. This necessitates innovative strategies extending beyond simple recycling, including design for disassembly, component remanufacturing, and materials recovery at a scale previously unseen. Effective end-of-life processes are no longer an afterthought, but a core element of product lifecycle management, demanding collaboration across industries and the development of new business models centered around resource stewardship and closed-loop systems. Ultimately, a successful transition hinges on viewing discarded products not as waste, but as valuable sources of secondary materials, driving economic opportunities while minimizing environmental impact.

Automated disassembly represents a pivotal technological advancement in the pursuit of resource efficiency and a circular economy. Traditional manual disassembly is often labor-intensive, costly, and inconsistent in material recovery; however, robotic systems, coupled with artificial intelligence, offer a pathway to systematically deconstruct products at scale. These systems utilize computer vision and machine learning algorithms to identify components, assess their condition, and precisely separate materials – even in highly complex assemblies. This precision not only maximizes the recovery of valuable resources like rare earth metals and plastics, but also reduces contamination, improving the quality of recycled materials and lowering the environmental impact associated with waste streams. The development of increasingly sophisticated automated disassembly lines is therefore crucial for closing the loop on material lifecycles and fostering a more sustainable, resource-conscious future.

Effective transition to a circular economy hinges on the creation of sophisticated automated disassembly systems. These aren’t simply robotic arms taking things apart; they demand integrated technologies capable of ‘seeing’ and ‘understanding’ the composition of complex products. Current research focuses on combining computer vision, machine learning, and advanced sensor data – including tactile and acoustic feedback – to identify materials, recognize component geometries, and map internal connections. The goal is to create systems that can dynamically adapt to variations in product design, safely separate hazardous materials, and optimize disassembly sequences for maximum material recovery. Such robust systems will be crucial for unlocking the valuable resources currently locked within end-of-life products, minimizing waste, and fostering a truly circular flow of materials.

Perceiving the Components: The Eyes of Automated Systems

Computer Vision equips automated disassembly systems with the ability to process and interpret visual data, functioning as the primary sensory input for understanding an object’s structure. This is achieved through algorithms that analyze images or video feeds to identify features, shapes, and spatial relationships between components. Unlike traditional robotic systems reliant on precise pre-programming and fixed coordinates, Computer Vision enables adaptability to variations in object pose, lighting conditions, and even minor design changes. The system constructs a digital representation of the physical assembly, allowing it to ‘perceive’ the object’s composition and prepare for targeted disassembly actions. This perception is crucial for navigating the complexities of real-world objects that deviate from ideal conditions.

Object Detection, a fundamental technique within the field of Computer Vision, involves algorithms that not only identify objects present in an image or video stream but also precisely locate their positions. This is achieved by generating bounding boxes around each detected object, providing coordinates that define the object’s spatial boundaries. The process relies on training machine learning models – typically convolutional neural networks – on large datasets of labeled images to recognize patterns and features associated with specific objects. In the context of disassembly, Object Detection enables a system to distinguish individual parts within a complex assembly, moving beyond simple image recognition to provide spatial awareness critical for robotic manipulation and selective deconstruction.

The YOLO (You Only Look Once) algorithm is a convolutional neural network (CNN) architecture specifically designed for real-time object detection. Unlike traditional methods that require multiple passes to identify objects, YOLO processes an entire image in a single instance, significantly increasing speed. This is achieved by dividing the image into a grid and simultaneously predicting bounding boxes and class probabilities for each grid cell. Current iterations, such as YOLOv8, prioritize both high accuracy – measured by metrics like mean Average Precision (mAP) – and frames per second (FPS) performance, making it suitable for dynamic environments where disassembly robots must react to changing conditions and identify components quickly and reliably. Its efficiency stems from its ability to minimize localization errors and confidently classify objects even with variations in scale, pose, and lighting.

Object detection enables disassembly systems to distinguish between different brick components – such as 1×1, 1×2, and 2×4 bricks – within a larger assembly. This differentiation is achieved by training algorithms on labeled datasets of brick imagery, allowing the system to identify each component’s class and precise location in 2D or 3D space. Successful component identification is a prerequisite for selective disassembly, where the system can target specific bricks for removal based on their type, position, or desired disassembly path, rather than performing a completely randomized or brute-force approach.

Cultivating Future Systems: A Project-Driven Approach

The Digital Technologies Study Program centers around Project-Based Learning (PBL) as a core pedagogical approach. This methodology prioritizes practical application by tasking students with completing projects that mirror challenges encountered in professional settings. PBL allows students to develop technical skills alongside crucial soft skills such as problem-solving, collaboration, and critical thinking. By actively engaging in the design, development, and implementation phases of projects, students gain hands-on experience and a deeper understanding of the subject matter compared to traditional lecture-based learning. The program’s structure is designed to ensure students are consistently applying theoretical knowledge to real-world scenarios.

The Digital Technologies Study Program employs Agile Project Management methodologies to facilitate flexible and iterative project development. Specifically, the program utilizes the Scrum Framework, adapted for educational purposes as eduScrum. This approach centers on short development cycles – sprints – allowing for frequent review and adaptation based on ongoing results. Key components include defined roles, daily stand-up meetings – referred to as “Daily’s” – and intermediate evaluations – “Inbetweens” – to ensure continuous feedback and improvement throughout the project lifecycle. This iterative process contrasts with traditional, linear project management approaches, enabling students to respond effectively to changing requirements and challenges.

Educational Robotics provides a practical context for students to apply Project-Based Learning and Agile methodologies through the design and implementation of robotic disassembly systems. These systems require students to integrate principles of mechanical engineering, electrical engineering, and computer science to create robots capable of deconstructing objects. The complexity of this task necessitates iterative development, mirroring the sprint-based structure of the eduScrum framework, and demanding continuous refinement of design and implementation based on testing and feedback. This hands-on approach allows students to directly translate theoretical knowledge into functional robotic solutions, reinforcing learned concepts and fostering problem-solving skills.

The Digital Technologies Study Program allocates 60 of its 180 total European Credit Transfer and Accumulation System (ECTS) points to project-based coursework. This represents a significant 33.3% weighting towards practical application within the bachelor’s program. The substantial ECTS point allocation signifies a deliberate curricular emphasis on hands-on learning, exceeding typical theoretical-to-practical ratios found in many undergraduate engineering and technology curricula. This commitment ensures students gain considerable experience applying learned concepts through sustained, extended projects.

Project workflows within the Digital Technologies Study Program are organized into seven iterative sprints, each with a fixed duration of two weeks. These sprints incorporate two key meeting types: Daily’s, brief daily check-ins to assess progress and identify impediments, and Inbetweens, intermediate review sessions conducted to evaluate deliverables and adjust project direction. This sprint-based structure, emphasizing short development cycles and frequent feedback, is designed to facilitate rapid iteration, continuous improvement, and adaptability throughout the project lifecycle, allowing students to respond effectively to challenges and refine their solutions based on ongoing evaluation.

Expanding the Horizon: Dissemination and Impact

The Digital Technologies Study Program is demonstrably advancing the field of Educational Robotics through consistent research initiatives. This contribution isn’t merely theoretical; investigations delve into the practical applications of robotic systems to enhance learning experiences, focusing on areas like student engagement, personalized instruction, and the development of crucial STEM skills. Studies frequently explore novel hardware and software integrations, alongside innovative pedagogical approaches that leverage robotics to teach complex concepts in accessible ways. The program’s research actively shapes the discourse surrounding effective robotic integration in educational settings, providing empirical evidence to support-and sometimes challenge-existing methodologies and ultimately fostering a more robust understanding of how technology can best serve learners.

The Web of Science (WoS) database functions as a central hub for monitoring the output and evaluating the significance of research originating from the Digital Technologies Study Program. This comprehensive citation database allows for detailed analysis of publication trends, identifying highly influential articles, and mapping the reach of innovative concepts within the field of Educational Robotics. By tracking citations, WoS provides quantifiable metrics – such as impact factor and h-index – that demonstrate the scholarly influence of the program’s contributions. Furthermore, the database facilitates the discovery of related research, fostering collaboration and preventing redundant efforts, ultimately ensuring the program’s work builds upon and advances the existing body of knowledge in a meaningful and verifiable way.

The Digital Technologies Study Program prioritizes the widespread adoption of effective educational robotics techniques through rigorous peer-reviewed publication. This commitment extends beyond simply documenting research; it actively contributes to a continuously evolving body of knowledge accessible to educators and researchers globally. By subjecting findings to external scrutiny and validation, the program ensures the reliability and replicability of its best practices. This dissemination strategy fosters collaboration, accelerates innovation in the field, and ultimately enhances the quality of educational experiences centered around robotics for students worldwide. The program views sharing insights not as a culminating step, but as an integral component of its ongoing commitment to advancing the field and improving learning outcomes.

The Digital Technologies Study Program fosters a continuous loop of advancement by intertwining rigorous research with practical application and broad knowledge-sharing. This cyclical process doesn’t merely generate new techniques in educational robotics; it actively embeds them within the program’s core activities, ensuring ongoing refinement and improvement. Crucially, disseminating findings through peer-reviewed channels guarantees that these innovations extend beyond the program, contributing to the wider field and promoting sustainable development by providing educators with validated, effective tools and methodologies. This commitment to both internal growth and external impact solidifies the program’s position as a driver of progress and a valuable resource for the educational community.

The presented framework inherently acknowledges that any system, even a carefully designed educational one, is subject to entropy. It’s not enough to simply build a robotics curriculum; the structure must allow for adaptation and improvement, mirroring the circular economy principles it aims to instill. As Andrey Kolmogorov observed, “The most important things are not those that are easy to define, but those that are difficult.” This sentiment resonates deeply with the challenge of preparing students for a future demanding iterative design and problem-solving. The project-based approach, coupled with Scrum, isn’t about achieving a perfect, static solution; it’s about embracing change and building resilience through continuous refinement – a slow, deliberate process of improvement that honors the inevitable weight of the past while striving for graceful decay.

What’s Next?

The presented framework, while offering a structured approach to robotics education, merely postpones the inevitable entropy of any pedagogical system. To suggest a curriculum can ‘solve’ the circular economy is optimistic, perhaps even naive; it merely introduces a new, more complex layer of intervention into an already decaying system. The real challenge isn’t building automated disassembly lines in classrooms, but accepting that all assemblies, ultimately, disassemble.

Future iterations must address the inherent limitations of project-based learning-the tendency for enthusiasm to wane as complexity increases, the difficulty in scaling individualized learning, and the constant need for recalibration as technology inevitably shifts. The current emphasis on ROS and Scrum, while valuable, represents a snapshot in time. These tools, too, will yield to newer paradigms. A more robust framework would acknowledge this transience, prioritizing adaptable skillsets over specific software proficiencies.

The pursuit of ‘graceful decay’ within educational infrastructure demands a shift in focus. Less emphasis should be placed on achieving perfect uptime-a rare phase of temporal harmony-and more on developing resilient systems capable of absorbing inevitable failures. The true metric of success won’t be the number of robots built, but the capacity to dismantle and rebuild, again and again, in the face of constant change.

Original article: https://arxiv.org/pdf/2603.14529.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/

See also:

2026-03-17 19:59