Skip to content

Projects


2019 - 2024 - Embedded Software - Robotics engineer at Bitcraze

ROS 2 Navigation with a nano-quadcopter

mapping crazyflie

During this project, I've connected a nano quadcopter with the ROS 2 framework through the Crazyswarm2 project. This allowed me to connect it to simulation, navigation packages like NAV2, or simpler mapping and navigation package strategies.

Related Links

  • Blogposts:
  • Videos:

  • ROSCon 2022 Presentation

  • Robotics Developer day 2024
Python Embedded C Gazebo Webots ROS 2 Nav2

Robotic simulators of nano-quadcopters

mapping crazyflie

I've built multiple models for the nano-quadcopter, the Crazyflie, creating a low-poly Collada model suitable for multi-quadcopter simulation. These have been implemented with collision models, propeller physics, and control in both Webots and the new Gazebo.

Related Links

  • Blogposts
  • ICRA 2023 Workshop on Aerial Robotics Simulation:
  • IEEE Robotics & Automation journal paper:
Blender Python Embedded C C++17 Gazebo Webots

2015 - 2019 - PhD Researcher at TU Delft

Swarm Gradient Bug Algorithm

mapping crazyflie

I've researched how a swarm of tiny quadcopters can autonomously fly indoors without external positioning help. This posed a significant challenge, requiring individual quadcopters to fly independently and the swarm to communicate for coordination and avoidance. To address this, I developed SGBA, the Swarm Gradient Bug Algorithm.

Related Links

  • Blogpost:
  • Science Robotics Publication
  • Video Press release
Embedded C FreeRTOS ESP State Machines ROS 1 Gazebo Classic ARGoS Matlab

Autonomous pocket drone for Swarm Exploration

mapping crazyflie

Back in 2015, there was no standard platform for autonomous swarming for indoor exploration. Until the folks at Bitcraze released the right expansion decks, I had to develop my own platform for this by combining different modules together, developed internally by the engineers from MAVlab, TU Delft.

Related Links

  • Final Experiment Video:
  • IEEE Robotics and Automation Letters publication:
Embedded C GPIO Bluetooth ESP Optical Flow Paparazzi Matlab FreeCAD 3D printing

2014 Intern at Computational Photography - NEC Japan

Coded Infrared-Cut Filter

mapping crazyflie

As an intern at the NEC Corporation in Tokyo, Japan, I worked in the Information and Media Processing Labs. I focused on validating a novel high-sensitivity sensing method for computational photography. With a coded infrared cut filter, the concept could be used on regular security cameras to capture both images in the visual spectrum and infrared, which can be useful in low-light conditions.

Related Links

  • Conference Paper at Springer CAIP conference:
Matlab Diffraction Simulation OpenCV Experimental setup

2011 - 2014 Study Final Projects - TU Delft

MSc Final Project - Binocular Vision Stabilization

mapping crazyflie

I've developed a neural network model for the stabilization of the cameras of a binocular robotic head, based on the human cerebellum. This is based on the theory that the cerebellum processes a massive amount of sensor information and filters out the most important data for the task. This enabled the robot to more smoothly track the marker despite external disturbances, inspired by mechanisms in our own brain

Related Links

  • MSc Final Thesis:
C++11 Matlab Solidworks ROS1 Gazebo Classic OpenCV Optical Flow Neural Networks

BSc Final Project - Wearable timelapse camera

mapping crazyflie

Before switching to Mechanical Engineering, I completed my final project in Industrial Design Engineering. John's Phone is a company that made extremely simple cellphones as an anti-movement to smart iPhones. As the next product in their line, I conceptualized the 'John's Camera,' a wearable timelapse camera that can be clipped on and record the user's day.

Solid Works Industrial Design Photoshop Indesign Assembly planning 3D printing