BraceAssertion: Runtime Verification of Cyber-Physical Systems

Abstract

Abstract-Cyber-Physical Systems (CPS) have gained wide popularity, however, developing and debugging CPS remain significant challenges. Many bugs are detectable only at runtime under deployment conditions that may be unpredictable or at least unexpected at development time. The current state of the practice of debugging CPS is generally ad hoc, involving trial and error in a real deployment. For increased rigor, it is appealing to bring formal methods to CPS verification. However developers often eschew formal approaches due to complexity and lack of efficiency. This paper presents BraceAssertion, a specification framework based on natural language queries that are automatically converted to a determinitic class of timed automata used for runtime monitoring. To reduce runtime overhead and support properties that reference predicate logic, we use a second monitor automaton to create filtered traces on which to run the analysis using the specification monitor. We evaluate the BraceAssertion framework using a real CPS case study and show that the framework is able to minimize runtime overhead with an increasing number of monitors. I. INTRODUCTIONCyber-Physical Systems (CPS) are found in applications in structural monitoring, autonomous vehicles, and many other fields. Compared with the growth of the domain, verification and validation of CPS lags far behind [32]. The state of the practice in debugging CPS generally entails a combination of simulation and in situ debugging. In a 2007 DARPA Urban Challenge Vehicle, a bug undetected by more than 300 miles of test-driving resulted in a near collision. An analysis of the incident found that, to protect the steering system, the interface to the physical hardware limited the steering rate to low speeds [23]. When the path planner produced a sharp turn at higher speeds, the vehicle physically could not follow, and this unanticipated situation caused the bug. The analysis concluded that, although simulation-centric tools are indispensable for rapid prototyping, design, and debugging, they are limited in providing correctness guarantees. State of the art formal methods tools, including static analysis, theorem proving, and model checking, are insufficient in tackling the challenges in CPS verification and validation [32]. Other verification techniques, including model-based testing [11] and simulation [16] have high learning curves, impractical development costs, and scalability issues. Domain specific tools (e.g., passive distributed assertions [29] and symbolic execution [30]), though more scalable, fail to formally verify either qualitative constraints (e.g., the ordering of events), quantitative ones (e.g., timing), or both. Ad hoc debugging has become the de facto

Real-Time Simulation Support for Real-Time Systems

“…On the other hand, runtime verification where properties are formally specified and checked at runtime, receives a lot of attention to verify real-time systems. In Zheng et al (2015a), runtime monitors can check both qualitative (e.g., safety, liveness) and quantitative (e.g., bounded safety and liveness, responsiveness) properties. However, to detect the insidious real-time systems bugs that are manifest only in a specific deployment environment, runtime verification techniques require repetitive deployments that are either too expensive (e.g., in labor, time, and/or money), dangerous (e.g., involving autonomous vehicles), or infeasible.…”

“…A lot of research efforts has been made to improve and ensure the reliability of such systems. For instance, formal methods (Zheng et al 2015bCassez et al 2017;Bouyer et al 2011;Zheng 2014) and testing have been widely used to improve the quality in real-time systems. However, the tighter coupling between physical processes and software components in the modern real-time systems along with the varying spatial and temporal runtime contexts made such system exhibit diverse behaviors across runs (Zheng et al 2017a).…”

“…The simulators allow the designer to conduct a variety of test earlier at the planning and design stage and in a repetitive and safe fashion. At the same time, with the rapid development of computing technologies, the cost of running simulations is steadily reduced and the performance is increasingly improved, making the simulators available to more researchers and engineers for a wider variety of real-time systems (Zheng 2015). The adoption of real-time simulation allows testing real-time systems under faulty and extreme conditions without damaging equipments under test while maintaining sufficient flexibility in choosing test parameters and components.…”

An Overview of Attacks and Defences on Intelligent Connected Vehicles

“…2) Real-time Simulation and Formal Verification: On a separate regard, the industry is adopting real-time simulation and formal verification as part of security compliance check for intelligent vehicles [78], [202], [203]. In real-time simulation, computer models are used to accurately re-create repetitive and flexible test environment for vehicular systems [160] while formal verification provides security guarantee [204], [205]. As a promising future direction, deep learning based models can be combined with real-time simulation and formal verification to provide more rigid yet accurate security assurance.…”

A Preliminary Study On Emerging Cloud Computing Security Challenges

“…Cloud computing is a broad topic with equally broad security issues and surrounding threats and challenges. In order to protect the cloud computing service providers and cloud consumer, effective runtime verification [23,24,20,6,25] along with anomaly and signature based intrusion detections system are strongly recommended to handle the emerging and evolving security threats in cloud computing [9,12]. Anomaly based intrusion detection system analyzes all the network traffics and detects the abnormal behaviors of network packets and applications.…”