Project 26: Digital Twin of Medical Devices
About this course
- Lecturers: Olaf van der Sluis (Coordinator), Simon Eugster, Victoria Dolean-Maini, Anton Wijs, Kayleigh Williams.
- Academic level:
- Number of students: 100
- Number of ECTS: 5
- Disciplines: Applied Physics and Science Education, Applied Physics and Science Education
- Status:
- Topic:
Set-up
Digital twins are digital models that are connected to their physical counterpart by means of data. In this project, digital twins of devices and systems (e.g., medical devices such as catheters, stents, thrombectomy devices, robotics, and manufacturing equipment) will be developed. These digital twins can have different purposes, such as device and system design, training, studying ‘what if’ scenarios, process planning, logistics, (predictive) maintenance, or recycling and sustainability (e.g., product lifecycle management). To enable the real-time simulation capability, game engines will be applied for realistic visualization. In addition, to describe the physical behavior of these devices and systems, so-called physics engines are required. Examples are SOFA, LapGym, Mujoco, DART, PyBullet, Gazebo, OpenSIM, and PhysX. Multiple aspects will be covered in this project, such as verification and validation of software results, understanding of (nonlinear) physics-based simulations and solvers, assessment of the accuracy, speed, robustness, maintainability, and usability of different types of software, accelerating simulations (e.g., by graphic cards), assessing differences and limitations of open source, free, and commercial software tools, and how to break up complex problems into smaller but still meaningful problems. Teams receive a budget of up to €100 for materials to build a hardware prototype, and the completed prototype remains the property of the university
Intended Learning Outcomes (ILO’s)
- Testing and verifying the accuracy of simulation software.
- Learning how to simulate complex physical behaviours, such as nonlinear deformations.
- Comparing different software tools regarding speed, accuracy, ease of use, and maintainability.
- Using graphic cards (GPUs) to accelerate simulations.
- Understanding the strengths and limits of open-source, free, and commercial software.
- Learning how to break down complex problems into smaller, manageable parts.
Assessment
This course used the following methods of assessment:
- Individual assessment (50%): The individual grade reflects personal engagement, collaboration, the quality of self-study assignments, unique contributions, and peer feedback.
- Group Assessment (50%)
- Intermediate Presentation (20%)
- Final Presentation with a Functional Demonstrator (30%)
Attendance at all on-campus meetings is mandatory, and unexcused absences can negatively affect the final individual grade through peer evaluations
Learning Activities
The course employs a Challenge-Based Learning approach based on the Design Thinking methodology. Students work in multidisciplinary teams and meet twice a week to discuss progress and plan next steps. Between these sessions, students complete individual Self-Study Assignments to advance the project. Practical activities include building hardware prototypes, running real-time simulations, and connecting digital models to real physical data using sensors. The process is structured into phases of empathising and defining (Weeks 2 to 3), ideating (Weeks 4 to 5), prototyping (Weeks 6 to 7), and testing (Weeks 8 to 9)
Organisation of the course
- Course Guidance: The project is coordinated by Olaf van der Sluis. The teaching staff also includes Simon Eugster, Victorita Dolean-Maini, Anton Wijs, and Kayleigh Williams. Tutors are present at all structured group meetings to monitor progress and provide guidance.
- AI Policy: The course maintains an AI policy where generative tools can be used for brainstorming, spelling checks, and generating example code, provided their use is clearly declared. However, submitting AI-generated content directly as original work is forbidden
Contacts
