Personalised products like medical devices or safety shoes offer numerous benefits, but there are challenges when it comes to designing them. Current methods often include manual or semi-automated steps, making it tedious and time consuming. For his PhD, Farzam Tajdari developed new methods of pattern recognition, created an optimised 4D scanning device, and produced a novel 4D data set in the process, showing there is a better way to do personalised design.
Maths as a bridge
In mass production, large numbers of identical or similar products are produced, but there are some issues in this process. It can result in things like waste of materials from production as well as sizing issues that cause discomfort or non-use of a product.
Although advances in technology have improved the potential for personalised mass production, there are challenges in the design process when it comes to personalised fit. Understanding the geometry of both the user and the product is the starting point to address these challenges.
With a background in automation and mechanical engineering, Farzam Tajdari at first felt a bit out of his element doing a PhD with a design focus. But his experience in the field of non-linear and control systems helped form a bridge to the design of ultra-personalised products. “When I started my PhD, I was working with mathematical parts of pattern recognition, but as it went on, I moved more towards topics like human centred design and how people feel,” he said. The human aspect, though new to him, was fundamental to his work. Through his multi-disciplinary research, Tajdari aimed to develop a systematic framework including methods and tools for 3D and 4D (3D in motion) scanning to help designers create personalised products with more comfort and safer performance for the users.
Non-rigid mesh registration and murmuration
Mesh registration in 3D scanning is the process of aligning a point cloud in a fine source mesh to the output point cloud of the scanner, called target mesh. During the process, the source mesh may deform for the “best” fit to the target mesh known as non-rigid mesh registration. For a simpler explanation, Tajdari likened it to a murmuration of birds, where they fly together but the shape is always changing. “Non-rigid mesh registration is when a group of points are moving while the identity of each point and the neighbouring points are remaining the same, but they can freely move and the shape of the group can freely deform,” he said. His research focussed on finding ways of automatically extracting 3D and 4D features from raw scanned data using non-rigid registration that accounts for addressing design challenges, particularly for human body shapes.
Novel results
There were several important and novel contributions that resulted from Tajdari’s research. One was the development of software programmes that enable an automatic method of pattern recognition in 3D and 4D scanning. Next, he designed and manufactured new hardware, specifically a 4D scanner capable of achieving high-speed 4D scanning of a moving object at 15 frames per second. This scanner was then used to collect and create a unique dataset Tajdari called 4D Feet by scanning the actively walking feet of 58 people.
All the data and code that resulted from his research is accessible via the 4TU.ResearchData website, as well as DINED, IDE’s anthropometric database. The methods, tools, and data have a variety of applications. Tajdari said, for example, it could be used to design a safety shoe that is optimally comfortable and safe or a personalised garment. He also noted that the 4D scanner could be adapted for any other object, including the dimension of a full human body during different activities.
Double PhD, multiple perspectives
A unique aspect of Tajdari’s research journey is that he was simultaneously pursuing two PhDs. About a year before coming to TU Delft, he started a PhD at Aalto University in Finland on automated and connected vehicles. Having made good progress, he decided to pursue the work at TU Delft while completing his first PhD remotely. They were two independent PhDs with different topics, methods, and publications. Tajdari said it was unconventional and required a lot of work, but there was an advantage to taking this route.
Doing two PhDs in two different fields gave me more creativity and made my mind more dynamic. It helped me to think at the same time on different things from a broader point of view.
―
Farzam Tajdari
“Doing two PhDs in two different fields gave me more creativity and made my mind more dynamic. It helped me to think at the same time on different things from a broader point of view,” he said. “That's why I always say bridging is a key word in my PhD. I could bridge because I could think at the same time on multiple topics, which also gave me a vision about the methods I could develop for TU Delft,” he said.
Tajdari recently finalised a postdoctoral project at TU Eindhoven related to his first PhD. He is now starting a new two-year postdoctoral project in the Faculty of Mechanical Engineering at TU Delft, where he will study human movement with the aim of reducing motion sickness.
