Koguma — Space Robotics & Swarm Navigation

This research project was related to the Google Lunar X Prize initiative at Yoshida's Space Robotics Lab in Japan. The purpose was to distribute a swarm of minimalist collaborative robots for planetary exploration and mapping of the moon.

Autonomous Swarm Algorithms : Created autonomous algorithms for edge systems with strictly limited resources. These could be used for a swarm of mini-robots to follow each other in formation, thereby saving energy.

Communication Protocol : Developed a modular communication protocol to manually control, stream video, and take photos with each individual robot over LTE.

The client needed autonomous, energy-efficient algorithms that can execute locally "on the edge" with low latency and limited memory/CPU, as well as a modular communication protocol for remote control, real-time video streaming, and image capture over LTE.

As Solution Architect Robin was responsible for the overall system design in an edge environment. He designed the pipeline for real-time video streaming and image capture over LTE from autonomous robots, with clear latency and bandwidth budgets, and modular separation between sensor input, encoding/compression, transmission, and decoding. He designed object-tracking algorithms based on classical state-of-the-art methods in image and video analysis, adapted for limited CPU and memory resources.

As Full Stack Developer Robin implemented the full solution: from camera reading and efficient frame handling to encoding/compression and network transmission, as well as client logic for control commands, testing, and media handling. He optimized algorithms to meet real-time requirements on limited hardware (profiling, memory footprint, threading), built test suites for both offline and integration tests in field-like scenarios.