Work History:

Robotics Software Engineer - Japan

Currently I am working at SEQSENSE / シークセンス writing software for SQ2 series of security robot. Primarily the robot acts as an autonomous security camera that carries out scheduled, pre-defined patrols each day in large office buildings. The types of buildings our robots operate in all have dedicated security centres, so there robots are used to complement the jobs of the already-present security personnel.

SQ2 at Narita Airport
SQ2 performing a patrol at Narita Airport (Tokyo)
Test robots
Test robots being serviced in the laboratory

SQ2 uses a cluster of three 2D laser scanners which are rotated on a turntable to perceive the surrounding environment and navigate. Additionally it uses four wide-angled cameras placed around the body for image capture. Data is continuously streamed to a web-based interface which allows security guards to review patrols and see what the robot is doing at any time. The robot is capable of self-docking with it's battery charger when required and can also use specific elevators to enable multi-floor patrols.

My work is mainly focused on object detection and tracking for a safety feature which is currently still under development. Previously I worked on a mapping system used to generate the 3D maps used for navigation. Working here has taught me about software containerisation, continuous integration, and how different cultures think about development.

Robotics Post-Doctoral Researcher - New Zealand

Between 2015-2018 I was a post-doctoral researcher on a kiwifruit orchard robotics project. The project was a collaboration between two universities, a government research institution and a private robotics company and was government funded. As a team of about 12 people, we developed robotics for autonomously pollinating and harvesting kiwifruit as well as a multi-purpose platform to which we could attach the pollinating and harvesting robots.

The multi-purpose platform in a kiwifruit orchard

Autonomous platform

The unloaded platform driving around an orchard

My work primarily focused on the development of the multi-purpose platform. The platform was a hybrid petrol-electric vehicle capable of autonomous row-following in kiwifruit orchards by way of lidar based sensing of the orchard's structure. It could drive fully electrically by way of six electric motor/gearbox units while providing AC and DC power outputs any mounted robotic systems. Also, it had a carrying capacity of 1 tonne on its payload space, plus extra capacity to carry kiwifruit in the space between the rear wheels.

Kiwifruit harvesting

The project's final season of kiwifruit harvesting

The kiwifruit harvesting and pollinating units are quite spectacular to watch. The harvester could identify kiwifruit in 3D using stereo cameras and artificial neural network, decide in which order to harvest the kiwifruit (to minimise damage to neighboring fruit), then instruct the custom robotic harvesting arms to take the fruit. You can see it working in the video below.

Kiwifruit pollination

The pollinator spraying a targeted cluster of kiwifruit flowers

The pollinator could use the same cameras and neural-networks to identify flowers in 3D and then shoot pollen solution at flowers as the vehicle drove underneath. The time-of-flight of the pollen solution was calculated for each shot so the pollen would hit the flower while the vehicle was in motion. Unfortunately there is no publicly available video of this robot to show, but this animation shows the general process.

Overall, my role was quite varied and involved mechanical design and fabrication, electrical design and fitment, as well as a variety software development. Software included custom computer-vision algorithms to detect kiwifruit, lidar processing for autonomous navigation, custom firmware to control motor drivers, as well as low-level software for safety. Looking back at this work, it was an amazing experience and opened my eyes to the possibilities of robotics in the agriculture industry.


Curriculum Vitae:

For more details on my work or research experience, my CV can be downloaded by clicking the image below. If you're interested in adapting this document for yourself, download the source code.