Cooperative Consensus-based Control of Multi-agent Systems over Wireless Channels
PIs: Jörg Raisch (TU Berlin), Slawomir Stanczak (TU Berlin)
The central objective of this proposal is to investigate and exploit trade-offs and synergies involving control and communication in the context of cooperative control of multiagent systems over wireless channels. In multiagent systems, a potentially large number of entities (“agents”) operate in a shared environment and need to jointly solve tasks while observing restrictions regarding communication (and other) resources. There are numerous application scenarios, ranging from groups of terrestrial vehicles to swarms of unmanned aerial vehicles employed for, e.g., monitoring air or water quality, detecting forest fires, etc. Distributed control of such systems has been recognized as an important topic, asmonolithic control approaches exhibit a number of practically important disadvantages. Such distributed control schemes typically feature both local control and a cooperative control level, often based on consensus mechanisms. Whereas in the former, local measurement information is fed back to local actuators, the latter involves information exchange between agents, typically over wireless networks. In this scenario, there are a number of mutually related trade-offs: (i) control versus communication on the cooperative level, (ii) local versus cooperative control, and (iii) trade-off involving communication aspects, as, e.g., accuracy and delay. These trade-offs clearly impact on the communication demand imposed by control. Strong emphasis will be on matching the communication supply to this demand. E.g., as agents typically do not need to know individual specific information from other agents but rather to evaluate a function of this information, superposition in transmission channels may in fact become a property that allows drastic savings in communication.
Involved PhD candidates
- F. Molinari, S. Stanćzak, J. Raisch: Exploiting the Superposition Property of Wireless Communication for Average Consensus Problems in Multi-Agent Systems, European Control Conference (ECC), 2018.
- F. Molinari, J. Raisch: Automation of Road Intersections Using Consensus-Based Auction Algorithms, American Control Conference (ACC), 2018.
- S. Gallenmüller, S. Günther, M. Leclaire, S. Zoppi, F. Molinari, R. Schöffauer, W. Kellerer, G. Carle: Benchmarking Networked Control Systems, 1st Workshop on Benchmarking Cyber-Physical Networks and Systems.