Swarms of robots neutralize dangerous Byzantine robots using a blockchain-based token economy

In a new study, we demonstrated the potential of blockchain technology, known from cryptocurrencies such as Bitcoin and Ethereum, to secure the coordination of swarms of robots. In experiments conducted with real and simulated robots, we demonstrated how blockchain technology allows robot swarms to neutralize harmful robots without human intervention, thereby enabling the autonomous and safe deployment of robot swarms.

A robot swarm is a multi-robot system consisting of many robots collaborating to perform a task. They do not require a central control unit but the collective behavior of the mob is more a result of local interactions among the robots. Thanks to this decentralization, robot swarms can work independently of external infrastructure, such as the Internet. This makes them highly suitable for applications in a variety of different environments such as underground, underwater, at sea and in outer space.

Although swarm robotics applications are currently exclusively demonstrated in research environments, experts anticipate that in the not-distant future, swarms of robots will support us in our daily lives. Swarms of robots can carry out environmental monitoring, underwater exploration, inspection of infrastructure and waste management—and thereby make a significant contribution to the transition to a fossil-free future with low pollution and a high quality of life. In some of these activities, swarms of robots will even outperform humans, producing higher-quality results while ensuring our safety.

However, once a robot swarm is deployed in the real world, it is likely that some of the robots in the swarm will be damaged (for example, due to bad weather conditions) or possibly even hacked. Such robots would not behave as expected and were called “Byzantine” robots. Recent studies have shown that the actions of a very small minority of such Byzantine robots in swarms—similar to a virus—spread in swarms and thereby destroy the entire system. Although security issues are critical to real-world deployment of robot swarms, security research in swarm robotics is lacking.

On the Internet, Byzantine users, such as hackers, have been successfully prevented from manipulating information using blockchain technology. Blockchain technology is the technology behind Bitcoin: it allows users to agree on ‘who owns what’ without the need for a trusted third party such as a bank. Initially, blockchain technology was only meant to exchange units of digital currency, such as Bitcoin. However, a few years after the release of Bitcoin, blockchain-based smart contracts were introduced by the Ethereum framework: these smart contracts are programming code that runs on the blockchain network. Since no one can manipulate or stop this code, smart contracts enable “code is law”: contracts are automatically executed and do not require a trusted third party, such as a court, to enforce.

So far, it is not clear whether large robot hordes can be controlled using blockchain and smart contracts. To answer this open question, we present a comprehensive study with real and simulated robots in collective sensing scenarios: the aim of robot swarms is to provide approximations of environmental features. To do so, the robot needs to sample the environment and then agree on the feature values. In our experiment, each robot is a member of a blockchain network managed by the robot itself. The robots send their environmental feature estimates to a smart contract which is shared by all the robots in the swarm. These estimates are combined with a smart contract that uses them to generate an estimate of the requested environmental feature. In this smart contract, we implement an economic mechanism that ensures that good (non-Byzantine) robots are rewarded for sending useful information, while harmful Byzantine robots are punished. The resulting robot economy prevents Byzantine robots from participating in herd activities and influencing herd behavior.

Adding blockchain to robot swarms increases the computational requirements of robots, such as CPU usage, RAM and disk space. In fact, it is an open question whether running blockchain software on swarms of real robots is even possible. Our experiments have shown that this is indeed possible because the computational requirements are manageable: the use of additional CPU, RAM, and disk space has little impact on the robot’s performance. The successful integration of this blockchain technology into the robot pool paved the way for a wide range of secure robotic applications. To support this future development, we have released our software framework as open source.