标题： 基于通信约束的随机网络化系统的安全控制器综合 显示英文标题

标题： Safety Controller Synthesis for Stochastic Networked Systems under Communication Constraints

摘要： 本文提出了一种框架，用于合成在通信不完善条件下运行的离散时间随机线性控制系统（dt-SLS）的安全控制器。控制单元是远程的，并通过一个不完善的无线网络与传感器和执行器进行通信。我们考虑了从传感器到控制器信道（上行链路）中的恒定延迟，以及在传感器到控制器和控制器到执行器（下行链路）信道中的数据丢失。在我们提出的方案中，每个信道中的数据丢失被建模为独立的伯努利分布随机过程。为了系统地处理上行链路延迟，我们首先通过连接所有足够表示原始dt-SLS在延迟和数据包丢失约束下状态-输入演化的状态和控制输入，来引入一个增广的离散时间随机线性系统（dt-ASLS）。然后，我们利用dt-ASLS的控制屏障证书（CBCs）来合成一个控制器，该控制器在随机意义上保证dt-SLS的安全性，确保dt-SLS的所有轨迹在具有量化概率边界的安全区域内。我们的方法将安全约束转化为矩阵不等式，导致一个优化问题，最终量化了在存在通信不完善的情况下满足安全规范的概率。我们在一个受到恒定延迟和概率数据丢失影响的RLC电路中验证了我们的结果。 显示英文摘要

摘要： This paper develops a framework for synthesizing safety controllers for discrete-time stochastic linear control systems (dt-SLS) operating under communication imperfections. The control unit is remote and communicates with the sensor and actuator through an imperfect wireless network. We consider a constant delay in the sensor-to-controller channel (uplink), and data loss in both sensor-to-controller and controller-to-actuator (downlink) channels. In our proposed scheme, data loss in each channel is modeled as an independent Bernoulli-distributed random process. To systematically handle the uplink delay, we first introduce an augmented discrete-time stochastic linear system (dt-ASLS) by concatenating all states and control inputs that sufficiently represent the state-input evolution of the original dt-SLS under the delay and packet loss constraints. We then leverage control barrier certificates (CBCs) for dt-ASLS to synthesize a controller that guarantees dt-SLS safety in a stochastic sense, ensuring that all trajectories of dt-SLS remain within safe regions with a quantified probabilistic bound. Our approach translates safety constraints into matrix inequalities, leading to an optimization problem that eventually quantifies the probability of satisfying the safety specification in the presence of communication imperfections. We validate our results on an RLC circuit subject to both constant delay and probabilistic data loss.