Real-Time Networks for Robotics and Industrial applications: Research Challenges and Novel Solutions

The increasing adoption of smart sensors and actuators enables industrial applications to process data in a decentralized way. In this context, communication networks play a key role, as industrial automation applications require network architectures and communication protocols, wireless and wired, able to support the interaction between multiple devices not only in a reliable way, but also guaranteeing the meeting of real-time constrains of the supported applications. The distributed processing among coordinated automation devices includes the cooperation of robot teams. However, cooperative robot networks impose additional communication constraints to those required in the automation context. This thesis originates from the collaboration with STMicroelectronics and targets mechanisms, algorithms and protocols that meet the communication requirements of cooperative robot applications in the industrial automation scenario. With the aim to avoid the definition of "yet another protocol", the work mainly focuses on the existing communication technologies adopted in the industrial context. In particular, innovative mechanisms, algorithm and protocols built upon standard communication technologies are investigated, with the aim to meet both the specific requirements imposed by cooperative robot applications (e.g., mobility and low latency) and those that are typical of industrial automation networks. In the thesis, innovative solutions for several communication technologies, both wired and wireless, are presented and described. Among the network technologies addressed, EtherCAT, Bluetooth Low Energy, Sub-GHz Communications, the IEEE 802.15.4e and IEEE 802.11 standards. Evaluations through simulations, analysis and experiments on proof-of-concept implementations are reported, which prove the effectiveness and the suitability of the proposed solutions.