CoCPN-ng – Cooperative Cyber-Physical Networking: Next Generation
Automation, Mechatronics, Control Systems, Intelligent Technical Systems, Robotics
Final Report Abstract
The recent advancements in digitalization and automation enable a more flexible, intelligent, and autonomous operation of production and manufacturing lines, and water and power distribution systems. This development is additionally fostered by the ubiquity of generalpurpose networks, which provides the opportunity to re-use already existing infrastructure. However, the quality of service provided by general-purpose networks is fluctuating, resulting in varying latencies, packet loss rates, and achievable data rates for all end systems that share the communication resources. Control applications typically have stringent requirements regarding the timely transmission of control commands and sensor data, rendering them particularly susceptible to variations of the network quality of service. Consequently, they must be enabled to operate under variable communication conditions without performance degradation. Traditional approaches either rely on customized networks, e.g. fieldbuses, to satisfy the tight timing constraints of control applications, or they co-design and optimize control applications and communication protocols such that the demands and expectations of control applications are met. These approaches are expensive, inflexible, and of high intrinsic complexity, i.a. because they introduce a tight coupling between control applications and communication system or because they require significant efforts to meet timing requirements. CoCPN implements a different concept that enables control applications to cooperatively share the capacity of a general-purpose network in a lightweight and distributed manner. It combines elastic network-aware control applications with a control-aware communication system and introduces the CoCPN translator as a method to loosely couple these two parts. Elastic control applications can tolerate quality of service fluctuations to a certain extent and support an adjustable trade-off between control performance and sending rate. The control-aware congestion control CoCC leverages this elasticity to prevent possibly severe quality of service degradations due to congestion. It also distributes the available link capacity within the network such that fairness w.r.t. control performance is achieved. The CoCPN translator facilitates this cooperation, it provides methods for a bidirectional exchange of information between control applications and the communication system. The cooperative approach of CoCPN shows promising results: Harmful congestion has been avoided by CoCPN in simulative evaluations, even when network links where highly utilized. Control applications and non-control traffic were able to share the network without any observable undesired interferences.
Publications
- Sequence-Based Stochastic Receding Horizon Control Using IMM Filtering and Value Function Approximation, Proceedings of the 58th IEEE Conference on Decision and Control, 2019, pp. 6424–6430
Florian Rosenthal and Uwe D. Hanebeck
(See online at https://doi.org/10.1109/CDC40024.2019.9029717) - Stability Analysis of Polytopic Markov Jump Linear Systems with Applications to Sequence-Based Control over Networks, IFAC-PapersOnLine 53 (2020), no. 2, 3104–3111, 21st IFAC World Congr
Florian Rosenthal and Uwe D. Hanebeck
(See online at https://doi.org/10.1016/j.ifacol.2020.12.1030) - Cooperative Congestion Control for Cyber-Physical Systems, Proceedings of the 2021 10th Mediterranean Conference on Embedded Computing, 2021, pp. 125–128
Markus Jung and Martina Zitterbart
(See online at https://doi.org/10.1109/MECO52532.2021.9460239) - Hop-By-Hop: Advancing Cooperative Congestion Control for Cyber- Physical Systems, Proceedings of the 2021 IEEE 46th Conference on Local Computer Networks, 2021, pp. 511–518
Markus Jung and Martina Zitterbart
(See online at https://doi.org/10.1109/LCN52139.2021.9524887)
