Adaptive Control of Networked Distributed Systems with Unknown Interconnections Georgios Lymperopoulos, EE Controls PI: Petros Ioannou Introduction Known Interconnections Networked Distributed Control Systems (NDCS) consist of a plethora of agents/subsystems, each one of which is controlled by a local controller. Information on the states and control signals is exchanged between the agents through a control/communication network. Common problems with NDCS include delays in the communication network or uncertainties in the interconnections, with examples in economic systems, power systems or robotic systems. Local controllers can weaken interconnections to avoid instability and at the same time avoid problems due to dimensionality.   Proposed Controller:   Problem Statement Multi-zone HVAC systems can be modeled as interconnected systems with uncertainties and delays in communication. Unknown Interconnections   Proposed Controller:         The controller gains of the distributed controllers are estimated by the proposed adaptive law. The computational burden for each adaptive controller depends on the number of neighbors that the respective subsystem has. Discussion & Future Work In NDCS, the proposed distributed control scheme that weakens the effect of the interconnections via feedback can stabilize the local dynamic, in the case of known interconnections. In the case of unknown interconnections the proposed adaptive control scheme guarantees that the states converge close to zero in a set that is small in the mean square sense provided the interconnections can be weakened and the time delays are relatively small. Future work includes calculating accurate bounds for the delay that can be handled and finding a way to make a trade-off between delay and interconnections strength.   Georgios Lymperopoulos, lymperop@usc.edu