A Blockchain-Oriented Task Scheduling and Allocation System for ROS Enabled Mobile Robots

In multi-mobile robot applications, the operational processes such as the management of robots, task assignment, monitoring of assigned tasks, communication and coordination between robots, and data storage are executed through a centralized server system. Therefore, most critical decisions are made on this centralized server rather than by the robots themselves. However, working in a centralized system has numerous disadvantages such as the obligation to maintain a server, routing all communication through a central unit, susceptibility to connectivity issues that render the system inoperable, and increased bandwidth requirements as the number of robots increases. Moreover, any communication issues between the server computers and any of the robots in centralized systems affect the entire system’s operation. To address these limitations, a blockchain-powered distributed communication system for inter-robot communication has been developed in this study. A task allocation application between robots has been implemented on this developed distributed communication system. In the application, Hyperledger Fabric (HLF) has been utilized as the blockchain platform due to its advantages. Each robot is a peer in the blockchain network in the proposed system. A cost function which is computed in all robots has been introduced to reduce the communication load in the blockchain network during task distribution and to enable optimal task allocation among robots. With the proposed system, robots compute and choose the most suitable tasks using the cost function, hence transactions on the blockchain network are kept at optimal level. After reaching consensus of peers on task allocations via HLF, task data are transmitted to robots by Robot Operating System (ROS) integration. With the proposed system, a dynamic and distributed architecture has been introduced and implemented where mobile robots can communicate with each other over a blockchain network without the need for a centralized server. In experimental studies conducted on real robots, the proposed system demonstrated optimal task allocation across multiple phases, effectively adapting to various task requirements in different scenarios. For instance, in one scenario, the system effectively allocated a total of 9 tasks, distributed across two phases: 3 tasks in the first phase and 6 tasks in the second phase. This study presents an innovative contribution to the literature on communication of robots and task allocation. Also, this study has a high potential to be adapted to industrial applications including robotic instruments.