In this paper, we present some initial results about vulnerability of control systems that can be used in Internet of Things (IoT) applications. Up to our best knowledge, this paper is the first study about vulnerability of applied control systems in general, and especially in the IoT environment. The purpose of this paper is to examine fundamentals of linear control systems and consider vulnerability of its main features and concepts used in Internet of Things applications under potential malicious attacks. We examine vulnerabilities of system stability, controllability, observability, design of feedback loops, and design and placement of sensors and controllers (actuators). The detailed study is limited to the most important vulnerability issues in time-invariant, unconstrained, deterministic, linear physical systems. Several interesting and motivating examples are provided. We have outlined also some basic vulnerability studies for time-invariant nonlinear unconstrained systems, and indicate that such a study is particularly needed for distributed parameter systems that are very prone to outside physical and cyber-attacks.

INTRODUCTION

Internet of Things (IoT) is a recent research trend in computer science, computer engineering, electrical engineering, and other engineering fields [1]-[2] with the goal of constructing “immersive and pervasive networks that enable easy accesses and interactions among ”things”,” [۳]. The things are represented by any type of devices that can be connected to the network and interact with the network such as personal computers, smart phones, machines, vehicles, appliances, and in general their sensors and actuators, benefiting for example home automation, manufacturing, public transportation, and so on. Consequently, networks can be home networks, device-todevice networks, unmanned aerial networks, body area networks, satellite networks, and similar, providing person-toperson, person to machine, machine-to-machine communication. The things connected are used in different operations such as sensing [7], computing [11], communication [1], actuation and control [1], [4]-[5], [9], [13]. Closely related to IoT, and practically representing its integral parts are cyberphysical systems [6], [12], wireless sensor networks [8], mobile computing [1], and pervasive computing [1]. Computing involves also cloud computing, and recently fog computing [14]-[15] for end devices of IoT. For example, paper [15] uses fog computing [14] to solve the problem of selection of workers to complete some tasks in specified locations (spatial crowdsourcing). The problem is formulated as an optimization problem with two utility functions, one for the fog platform and another one for the workers. Security and privacy also play very important roles in IoT, [6] – [7], [9]. Wireless sensor networks have been instrumental in the development of IoT. Simultaneously used in numerous industrial applications these networks require efficient spectrum sharing. This problem has been recently studied in [16], where autonomous channel switching has been proposed to achieve fairness of spectrum usage among many users (things) connected to IoT. A survey of network methodologies that are efficient for big data IoT with billions of things, and related future research challenges have been considered in [17]. A nice overview of the present state of IoT and its future prospects can be found in [18]. The paper also discusses importance of security for IoT using the Smart Grid example. Detailed consideration of security and privacy in IoT smart city applications can be found in [19].