Team Kryha’s previous achievements
Or: the road to a paper in a journal
Writing an academic article is not a simple task. Getting it reviewed and published in a journal is the next step and not necessarily achieved. The winners of the machine-to-machine track on the Blockchaingers hackathon in April 2018, Odyssey’s former name, were Team Kryha with a system to put ‘blockchain on drones’, an ecosystem for machines which operate with other machines. No humans needed.
Team Kryha developed a way to let drones communicate with each other using the BigchainDB 2.0 blockchain. At the hackathon itself, they fitted a drone with a Raspberry Pi, so the drone was not only that, a quadcopter, but also a node in a network of wirelessly communicating drones. This system could be used for rescue missions or exploring remote planets.
The team deployed a network of nine nodes running BigchainDB, MongoDB and Tendermint for consensus. This way they could create a fully distributed network of running full nodes. They simulated the whole system and showed the nodes actually did do their job.
After their effort and win, they wrote a scientific paper on the experiment, which was conducted at the hackathon in Groningen. The paper, titled “Grex: A Decentralized Hive Mind”, researched Swarm Robotics, or SR, and the challenges ahead for widespread adoption for real-world applications. Some of these challenges could be overcome by using distributed ledger technologies (DLT).
If you wonder why they chose the name ‘Grex’, well that’s a no brainer for Latinists: it means ‘swarm’ in Latin.
As an experiment Kryha build a search and rescue mission, simulating drones which are coupled by single board computers, like the Raspberry Pi’s they used in their experiment in Groningen, and several simulated agents. The communications between the agents were facilitated through a distributed ledger in a completely decentralized network. In their paper they show they can easily extract information from the system. Also, when one agent would go rogue, it doesn’t matter as the other drones also have the same information due to the nature of distributed ledger technologies. If a drone crashes, still no problem.
Eventually all of this might come to future work where focus lies on more complex tasks through means of federated learning or inter-swarm communications.
If you wonder what ‘federated learning’ is, this is a system of collaborative machine learning without centralized training data. Initially it is focussed on mobile phones to collaboratively learn a shared prediction model while keeping the training data on the phone itself. But that’s for future research.
