Control under limited information: Special issue (Part I)

摘要

In recent years, new applications have changed the way that systems interact and communicate: users working across the Internet, people communicating with cell phones, entire manufacturing facilities that share real-time data for product tracking, and wireless systems to network computers. New technologies have brought extraordinary advantages to the developers of control systems including the simplification in wiring and maintenance as well as cost reduction for teleoperation, automobile applications, multi-robot systems, etc. The introduction of these technologies into the control systems, however, has also brought about new challenges. Control theorists have usually assumed perfect flow of information among the different devices (sensors to controllers, controllers to actuators). In the new control systems, however, dedicated and unconstrained communication channels have been replaced by shared communication systems or very constrained communication links. Perfect communication links are now the shared networks that suffer from traffic congestion and limited communication bandwidth. Such communication media feel the quantization effects, the potential loss of information, and random delays. These different aspects in the feedback control loop are what we refer to ‘control under limited information’. Control engineers are thus forced to expand their application domain by incorporating the communication infrastructure into their designs, and by considering the impact of link capacity, latency, and packet loss on the performance of feedback control systems. The need for new paradigms for control design is particularly evident in large-scale interconnected multi-agent systems. For such systems, signals need to flow quickly and efficiently, but interconnected components may not be able to store and manipulate the complete state of the system. Although complexity barriers render the design of controllers for high-dimensional systems impractical, the ability to reason about global network properties based on locally available information enables the design of decentralized control laws. These topics are receiving a lot attention and researchers are bringing important contributions to the area 1-4. The objective of this Special Issue (and its companion) is to present the latest developments in the field of control systems with limited information and to stimulate further research within the community. The first article in this first special issue is titled ‘Kalman Filtering over Unreliable Communication Networks with Bounded Markovian Packet Dropouts’, by Xie, Xiao, and Fu. This work addresses the peak covariance stability of a time-varying Kalman filter. The main contribution is that it assumes the presence of packet losses in transmitting measurement outputs to the filter when a packet-based network is used for the transmissions. The main assumption is that the packet losses are bounded and driven by a finite-state Markov process. This work uses the observability index of the discrete-time linear time-invariant (LTI) system under investigation as well as the system dynamics and the probability transition matrix of the Markov chain to establish conditions for the peak covariance stability of the Kalman filter. The validity of the results is demonstrated by numerical simulations. The second article is titled ‘Average Consensus On Networks with Quantized Communication’ by Frasca, Carli, Fagnani, and Zampieri. This article presents a contribution to the solution of the average agreement problem on a network when the links are quantized. Starting from the well-known linear diffusion algorithm, the authors propose an adaptation that is able to preserve the average of states and to drive the system near to the consensus value. This is achieved in the presence of uniform quantized communication between agents. The article investigates the properties of this algorithm using worst-case analysis as well as probabilistic analysis. Similar to