Senior Robotics Software Engineer
at Blue Ocean Robotics
Jan 2020 — Jan 2021
Developed the robot's full navigation, cognition, and perception stacks — localization, mapping, autonomous driving, mission planning, sensor fusion, and decision making. Built the safety-critical systems and led the TUV Rheinland certification from the functional safety side (IEC 61508, ISO 13849).
Led remote and on-site commissioning at client facilities including Clarion Hotel (Choice Hotels International) and Novo Nordisk, across Europe, Asia, and the Americas. Typical commissioning spanned two full days per site — mapping the facility, configuring disinfection areas, programming the full mission, and walking through every operation the customer would perform independently.
Key achievements
- Researched and implemented a state-of-the-art trajectory planner that enabled the robot to navigate smoothly in narrow, dynamic, and cluttered hospital environments.
- Raised manufacturing and product development quality from a new defect every day to robots deploying at client sites without incident.
Gallery
Behind the Scenes
Under the hood
Implemented and benchmarked trajectory planning for constrained hospital navigation, focusing on safety envelopes, optimizer stability, and real-time performance. Upstreamed the key improvements to teb_local_planner (teb_local_planner#210, teb_local_planner#209, teb_local_planner#259, teb_local_planner#208, teb_local_planner#234). Also designed a lightweight exploratory coverage mode for smart disinfection in confined spaces without map/localization overhead.
Led localization benchmarking in field and simulation, then contributed the improvements to the upstream localization library. Contributed production hardening and ROS-integration updates across iris_lama (iris_lama#25, iris_lama#26, iris_lama#27, iris_lama#28) and iris_lama_ros ( iris_lama_ros#21, iris_lama_ros#22, iris_lama_ros#23, iris_lama_ros#24, iris_lama_ros#26, iris_lama_ros#27, iris_lama_ros#28)
Built the mission planning application in React and TypeScript, with a behavior tree architecture driving mission execution on the robot. Contributed to hardware quality beyond software: electrical harness review, gear ratio dimensioning, and suspension system tuning — including behavior under motor brake engagement during safety stops.
Tools & Technologies
Related publications
Optimal control allocation of quadrotor UAVs subject to actuator constraints
American Control Conference (ACC), 2016
The control allocation theory from this paper — distributing high-level commands across actuators under saturation — directly informed the saturation strategy contributed to teb_local_planner (PR #209), translating academic research into production robot navigation.
Control Allocation for Wheeled Mobile Robots Subject to Input Saturation
IFAC-PapersOnLine, 2020
Extension of the control allocation framework to wheeled mobile robots, co-authored during this period and published in 2020.