Hello! My name is Angelos Dimakos, and I am from Greece. I am currently studying for an MSc in Robotics at Cranfield University. I am also the Course Representative for this course. During this blog, post I’d like to take you through my group project experience on the Robotics MSc.
The group design project element makes up 20% of the MSc course total and provides an invaluable experience of delivering a project within an industry-structured team. You get to develop a range of skills including learning how to establish team member roles and responsibilities, project management, delivering technical presentations, and gaining experience of working in teams that include members with a variety of expertise.
This year the group project addressed the implementation of drones for part delivery in smart factories. The large workspace, in combination with the small size, makes them ideal for the delivery of small components, both to human operators, as well as existing reconfigurable production lines. However, the low payload, and limited on-board processing power, present limitations. As such, a novel approach for the localisation, and dynamic path planning and following had to be created. In the developed system, two laptops and two drones were used. The leading drone utilised software for the localisation and robot programming, through the Robot Operating System. For the localisation, ArUco markers were used, as feature extraction is more robust and efficient. For the following drone, an object following algorithm was implemented, which utilised thresholding to determine the moment of an object attached to the leading drone. In addition, an object was attached as a proof of concept for part delivery. Then, 3 proportional-integral-derivative controllers were utilised to determine the movement of the drone based on the offset of the moment and the centre of the camera.
The experience has been very unconventional. Restrictions due to government guidelines have shaped the final project. Access to labs, support, and equipment have also been affected. However, this situation also provided unique opportunities. The group had to be very efficient, both during the meetings and the lab sessions. We had to accurately determine potential issues and solve them in a timely manner. A systematic experimental and research approach was crucial, due to time constraints. Templates regarding literature review, validation documents, and video footage had to be taken in a consistent manner, to streamline the system development. Finally, the natural curiosity and rapid implementation and testing of different approaches to an open-ended problem developed our skills as researchers and system integrators. This has been an insightful experience regarding the responsibilities of professional robotics engineers.
The project was a life changing experience for me. Initially, I had to learn to be more flexible in my approach on how to solve problems. Complex problems do not have clear solutions. The ability to change research direction, based on academic papers and experimentation, was necessary for the successful completion. Furthermore, I learned and appreciated the value of brainstorming and horizontal, agile project management. Some options would not produce results, so I had to discard and modify a lot of ideas and concepts. Additionally, my research and integration were affected by a variety of external factors, such as the situation, and challenges during the implementation.
In terms of knowledge, I learned a lot about distributed control, swarm robotics, and machine vision. As each drone had individual capabilities, we had to be very creative during the conceptualisation and testing of the developed system. The use of perception for cohesion maintenance was something that we devised through extensive personal and group work. Furthermore, I learned how to approach hardware that I had no experience with and developed a methodology for the testing of the software, while considering safety.
The challenges were very varied during the undertaking of this project. An initial challenge was the very tight deadline. We had to submit an assignment during March, while the group project was ongoing. Furthermore, we had approximately one and a half months for the development of the system. This was further exacerbated by the very limited access to labs, which was around four hours per week. In addition, familiarisation with the equipment was a very slow process, due to the dynamic nature of drones. An approach that was very targeted, incremental, and safe had to be adopted. Furthermore, the limited research in our specific application (intralogistics with swarm of unmanned aerial vehicles), required us to gain a lot of insight through experimentation. Another important consideration was the implementation of software. A very slow and methodical, validation-based approach had to be adopted. As an error in programming could lead to equipment damage, we had to be very careful when testing different technologies. This was highlighted during the development of the overall system, which required a safe distance between the drones, as well as the team.
I am very proud of my team. The adversities and challenges we encountered during the project required strength of character and bravery in some cases. We had to be very multifaceted and had to be in harmony. All of us were software developers, testers, integrators, designers, and researchers. Furthermore, there was no clear hierarchy, as we all had relatively limited experience, and had to brainstorm and decide on crucial options for the development of the system. Our ingenuity and flexibility also shone through, as we explored many different options, with varying levels of success, due to time constraints. This project would not have been possible without my brilliant team, our supportive supervisor, and the skilled technician responsible for the lab.
The modules were invaluable during the undertaking of this project. An understanding of simultaneous localisation and mapping, as well as the control of mobile agents, was very important during the initial literature review, to determine gaps in existing research and to determine a sufficient speed for the leading drone. Knowledge regarding proportional-integral-derivative controllers was also crucial for the tuning of the following drone. Furthermore, sensor fusion was crucial during the experiments, as we had to ensure that the environment would not create issues with the vision system. Additionally, machine vision was important, as an object of appropriate size and colour had to be chosen, to ensure accurate following. It should also be noted that the extensive validation and programming skills during the assignment were directly transferred to the development of the system. Finally, the robot programming applied to the leading drone was a direct result of the experience that was developed during our first module.
The skills I have learned will be very important for my future career. Quality control, software development, and integration have been very beneficial. Programming which had a close effect on hardware also enabled me to understand that a slow, methodical approach is necessary, due to the inherent challenges when interfacing with equipment. As such, I feel much better equipped and comfortable when dealing with new challenges, which I now perceive as opportunities. Furthermore, trade-off analysis, communication, and flexibility will be defining for future projects, where I will have different roles and responsibilities. Flexibility and willingness to assist and take on new problems will enable me to push myself and develop new skills. Extensive and accurate documentation will also help in accurate prioritisation and completion of tasks.
The challenges would be what I will fondly remember. I developed skills, knowledge, and friendships that would be very difficult to develop. However, all of them were very tightly connected to the inherent difficulties. The road to personal and professional growth is not easy. Obstacles and setbacks will naturally occur. The ability to tackle these problems with a group of like-minded people is an opportunity that is very hard to come by. My team has and will achieve great things. I was very honoured to meet each one of them under these exceptional circumstances.
In the group project, I would say the single most important benefit is the ability to make mistakes in a controlled environment. Communication can be very challenging, especially in a team of people with a diverse set of skills. The communication must be both honest and consistent. Time management can also often be an issue, especially when exploring different options. However, the resilience to bring a project to completion, and flexibility in terms of approaches, are necessary. Furthermore, patience and willingness to change the plan are crucial, as projects are very fluid. Finally, the ability to apply knowledge in real hardware is of course very important. During integration, a lot of issues can occur. The ability to deal with each of them is crucial, especially in the case of future robotics engineers.
I am very much looking to my thesis project! I will be exploring the mechatronic and process control in wire additive manufacturing. As I have been certified in welding engineering, the combination of industrial standards and academic research is a combination of two very distinct approaches. Standards provide rigorous specifications, to ensure productivity, safety, and high quality. Research provides options for rigorous analysis and novel solutions to challenging problems. However, the group project gave me a much better perspective. Automation and intelligence can be very difficult to integrate. As such, experiments, research, and the determination of achievable goals are a very important consideration when creating knowledge.
After I finish my course, I hope to pursue a career in academia. The skills I learned, from a technical, personal and managerial point of view will be invaluable, to produce impactful and meaningful research in the field of intelligent welding, and additive manufacturing. Additive manufacturing, welding and automation industries need solutions that can potentially address existing problems. An open problem is the creation of closed-loop control to increase productivity and quality while ensuring low material wastage and good metallurgical properties. Ihope to address this issue through the implementation of machine learning, sensor fusion, and mechatronic control.
My next blog post will explain my individual research project experience, please check back soon!
Read my previous blog: My Cranfield MSc: Robotics MSc – my experience so far