Specification Mining of Time-sensitive Systems for Safety and Security Applications

点击翻译按钮获取中文摘要

Embedded safety-critical systems are essential for Canada and Canadians. Should safety-critical systems fail to provide their required functionality, they have the potential to cause harm to people, loss of capital infrastructure, or significant damage to the environment. Canadians rely on safety-critical systems every day as they commute to work, travel in airplanes, and undergo medical treatment in hospitals. The lives of many Canadians are dependent on these systems; people rely on pacemakers, implanted insulin pumps, and neonate incubators. Businesses rely on these systems to steer supply chains, control factories, and manage finances. Finally, Canada's safety and independence relies on these systems for modern surveillance and defense technology; all of these systems are controlled by computers and software in embedded safety-critical applications.***Safety-critical systems are becoming increasingly complex; the more complex, the greater the risk of safety hazards for the public. Today's control software in vehicles and aircraft counts millions of lines of code. Increase in software complexity strongly correlates with an increase in system defects. Over the last 20 years, for instance, automotive software has become the prime reason for vehicle recalls --- accounting now for over 30% of all recalls. Defects in safety-critical systems heighten the risk of failure, which can lead to accidents severely impacting human life, property, and the environment.***This proposal outlines a research programme that will research and develop technology to (1) assist in understanding complex systems and (2) build a monitoring system to detect failures before they become hazards. The programme concentrates on specification mining and runtime monitoring as the core building blocks. Specification mining uses algorithms to analyze artifacts such as system traces with the goal to extract system behaviour. Extracted system behaviour can provide dynamic views on specific aspects that a developer might be interested in, and can be used as an effective tool to comprehend complex systems. Runtime monitoring provides mechanisms to observe a system and prevent it from entering undesirable, or even erroneous states. Runtime monitoring typically uses checker code synthesized from high-level specifications to check current system behaviour, and alert or steer the system in case of violations of high-level specifications. For example, an automotive transmission shall shall remain for a minimum time on a gear before switching to avoid rapid (malicious) engine wear.***By automating tools for safety and security monitoring, the programme will fundamentally advance the state-of-the-art of system development for complex embedded systems. The programme will also train HQP necessary for industry to translate the results into products to build safe and secure next-generation embedded systems.