Mobile manipulation

Leading an in-store order-picking robot

Led development of a mobile manipulator for in-store order picking, from system architecture to simulation, containers and end-to-end testing.

Team Lead / Robotics Software Engineer · AIRLab Delft · 2021–2024

A mobile manipulator that navigates a store and picks products brings perception, navigation, manipulation and user-facing software into one system. I led its development and designed how those components interacted.

System architecture for the in-store order-picking robot

System design

I designed the system architecture and the interfaces between its components. This made responsibilities and data flow explicit across a stack that combined the mobile base, manipulator, perception and application services.

Development environment

To make work across the team repeatable, I created:

  • A Gazebo simulation containing the store and products
  • Docker containers with GPU acceleration and graphical display access
  • GitHub Actions workflows for automated end-to-end tests
  • A Markdown documentation site linked to the repositories' READMEs

What shaped my work

Leading the project reinforced that robotics productivity is a systems problem. Simulation, containers, tests and documentation are not secondary tasks—they determine how confidently a team can change and ship the robot software.