2025

Goals

The goal of this project was to develop a safe, affordable, and fully sensorised elastic actuator suitable for prosthetic and orthotic devices. A key objective was to make compliant actuation accessible to students, researchers, and practitioners through an open-source hardware and software framework.

Scope

Research and development project conducted within an academic research group, resulting in a peer-reviewed HardwareX publication. The work covers mechanical design, electronics, control architecture, and experimental validation, with emphasis on reproducibility, modularity, and low-cost fabrication using 3D printing.

Implementation

The actuator is based on a series elastic actuator (SEA) architecture using a belt-driven DC motor, radial springs, dual encoders, load cells, and inertial sensing. Control software is implemented in embedded C++ on low-cost microcontrollers and supports both impedance and admittance control. The complete system—including CAD files, schematics, firmware, and documentation—is released under an open-source license.

What I Learned

This project provided extensive experience in designing compliant robotic actuators for safe physical interaction. I gained deep insight into elastic actuator modelling, impedance and admittance control, sensor fusion, and real-time embedded programming. The work also strengthened my skills in research-driven development, experimental validation, and producing reproducible open-source hardware.

Read the paper Open-source repository
2023-2025

Goals

The primary goal of this project was to design a simple and predictable application dock for tiling window managers such as bspwm and i3. A major personal objective was to deepen my understanding of Go by building a complete, real-world application with a clean internal structure.

Scope

Personal tooling project focused on usability, maintainability, and minimalism. The dock intentionally avoids dynamic layouts and complex behaviors, favoring static positioning and explicit configuration to integrate cleanly with window manager workflows.

Implementation

Icebar is implemented in Go and organized into clearly separated packages for configuration parsing, dock logic, GUI handling, event processing, and application launching. Configuration is provided through external TOML files, allowing margins, transparency, monitor selection, and styling to be controlled without recompilation. The application is distributed as a Docker container to ensure reproducible builds and straightforward deployment across Linux systems.

What I Learned

This project significantly improved my proficiency in Go, particularly in structuring medium-sized applications, managing packages, and writing clear, maintainable code. I also gained experience with Docker-based distribution for desktop applications, Linux window manager integration, and designing configuration-driven user interfaces.

View source on GitHub
Icebar dock running on bspwm
2024

Goals

The primary goal of this project was to design and evaluate a haptic teleoperation system that enables intuitive and safe control of an industrial robotic manipulator. A key objective was to convey meaningful force feedback to the operator, allowing them to adapt grip force and motion based on tactile cues rather than vision alone.

Scope

This project was developed as a research-oriented prototype with real-time performance constraints. The focus was on modularity, human–robot interaction, and experimental validation rather than production readiness. Emphasis was placed on low-cost hardware, 3D-printable mechanical components, and an open software architecture suitable for rapid iteration and future research.

Implementation

The system is implemented in C++ using ROS 2, with MicroROS-based microcontrollers handling distributed sensing and actuation. A custom-designed haptic glove tracks finger motion and provides force feedback using servo-actuated links and impedance control. The glove teleoperates an industrial robot arm equipped with a soft robotic gripper, enabling pick-and-place of objects with varying stiffness. The software architecture is fully modular, allowing vision, haptics, control, and robot interfaces to be developed and tested independently.

What I Learned

This project provided hands-on experience with designing real-time human–robot interaction systems, highlighting the challenges of stability, latency, and feedback fidelity in haptic control loops. I gained practical insight into ROS 2 system design, distributed control using microcontrollers, and impedance-based force feedback. The work also reinforced the importance of modular software architecture and mechanical simplicity when developing experimental robotic systems.

2023

Goals

The goal of this project was to develop a parametric tool capable of generating accurate involute gear profiles from fundamental design parameters such as module, number of teeth, pressure angle, and profile shift. Emphasis was placed on correctness, flexibility, and clear geometric interpretation.

Scope

Course project focused on analytical modeling and implementation of gear geometry rather than CAD automation or manufacturing optimization. The tool is intended for educational use, early-stage design exploration, and validation of mechanical concepts.

Implementation

The gear generator was implemented in MATLAB using closed-form expressions for involute geometry and gear tooth construction. The script computes base, pitch, and addendum circles, generates tooth profiles parametrically, and visualizes complete gears by rotational replication. The implementation allows rapid iteration and immediate visual feedback when adjusting design parameters.

What I Learned

This project deepened my understanding of involute gear theory and the relationship between abstract design parameters and physical geometry. I gained experience translating mechanical design equations into robust, reusable code and learned the value of visualization as a validation tool for analytical models.

2022

Goals

The goal of this project was to design a safe and autonomous physical arena where drones can be trained using reinforcement learning techniques outside of simulation. The system aims to protect both the drone and its surroundings while enabling long-duration, unattended training cycles.

Scope

Bachelor-level group project focused on conceptual design, mechanical construction, and system architecture. The work emphasizes safety, modularity, and portability, while laying the foundation for future integration of reinforcement learning algorithms and distributed control.

Implementation

The arena consists of a modular mechanical structure with a central actuator system that repositions and reorients the drone to a known starting pose after each training iteration. A depth camera and visual markers are used to estimate drone position and orientation, while a Raspberry Pi controls arena actuators and peripheral hardware. The system is designed to communicate with an external computation unit responsible for vision processing and learning.

What I Learned

This project provided practical experience in designing safety-critical robotic systems and translating abstract research requirements into mechanical and architectural solutions. I gained hands-on experience with system-level design, sensor integration, and the challenges of bringing reinforcement learning concepts from simulation into physical environments. The work also highlighted the importance of modularity and maintainability in long-running autonomous systems.

2022

Goals

The goal of the LightCar Race was to create an engaging, hands-on competition that introduces students to embedded programming, sensor-based control, and mechanical design. The challenge encourages iterative development, clean code, and empirical tuning while remaining accessible to beginners.

Scope

Educational competition developed for makerspace use. Participants design and build their own autonomous vehicle, including chassis, sensor placement, and control logic. The project balances simple hardware requirements with room for advanced control strategies and creative mechanical solutions.

Implementation

Each vehicle is built around an Arduino Uno controlling two DC motors via an L298 motor driver. Two photoresistors are used to track a lead car equipped with a rear-mounted light source. By combining the sensor signals, the vehicle follows the lead car through a predefined course with varying speed and curvature. Evaluation criteria include tracking performance, stopping accuracy, code quality, and overall design creativity.

What I Learned

This project strengthened my understanding of how to design educational engineering challenges that scale from simple rule-based control to more advanced feedback control concepts. I gained experience translating control theory into intuitive learning objectives, defining fair evaluation metrics, and designing constraints that promote both creativity and technical rigor.

View resources on GitHub
View project description on UiA Makerspace
2021

Goals

The goal of this project was to create an affordable and portable mechatronics laboratory that enables students to perform meaningful hands-on experiments outside traditional university lab facilities. The platform was designed to strengthen the connection between theory and practice in instrumentation, measurement, and control.

Scope

Bachelor-level collaborative project focused on educational system design. The work includes mechanical design, electronics integration, control experiments, and curriculum-aligned laboratory exercises. Particular emphasis was placed on accessibility, robustness, and reproducibility for large student cohorts.

Implementation

The home lab is built around a microcontroller-based platform with modular sensors and actuators, mounted on a custom 3D-printed base to ensure mechanical stability and low noise. A sequence of experiments was developed, ranging from first-order temperature control and load cell calibration to multi-axis drone thrust, pitch, roll, and altitude control using MATLAB/Simulink and hardware-in-the-loop techniques.

What I Learned

This project provided experience in designing robust experimental platforms for education, balancing technical complexity with usability. I gained insight into control education, system identification, sensor fusion, and hardware-in-the-loop simulation, as well as the practical challenges students face when transitioning from simulation to real hardware.

Read the paper