Enabling reliable control with communication constraints
From connected vehicles to smart factories, networked autonomous systems will enable future technological advancements. In many such systems, functional requirements like mobility will require that communication between autonomous agents be wireless. To ensure system reliability, and, most importantly, safety, the communication networks that control these autonomous systems will face strict latency and reliability requirements. Wireless resources are fundamentally scarce, and this scarcity is further compounded by the need for high reliability and low latency. This motivates research along two related threads; namely the development of wireless technologies to support the low latency and high reliability requirements of control and the development of communication cost-aware control systems. In this dissertation, we pursue reliable control over wireless via this two-pronged approach. On the channel side, we develop a novel signal processing framework for noncoherent communication using ideas from quantum error correction. This leads to the development of a family of full-diversity space-time block codes for multiantenna wireless communications. The codes are specialized to the domain of ultra-reliable low latency communications (URLLC). On the source side, we derive lower bounds on the minimum feedback bitrate required to sustain a given control performance in a Linear-Quadratic-Gaussian (LQG) control system. We then demonstrate that these bounds are nearly achievable using a time-invariant data compression architecture. We use the proposed architecture to develop algorithms for minimum rate LQG control that are suitable to a real-time implementation.