High stability Windows programming for real time control

The use of the Microsoft windows operating system as a basis for integrated security systems provides a pathway to lower development and training costs while increasing commercial hardware compatibility. However, the historical view of Windows as a high-overhead, low-stability operating system has tended to limit the number of Windows-based mission-critical automation systems relative to the more conventional embedded micro-controller or Linux-based approaches. This paper documents the use of the Windows operating system for mission critical control of an active denial security technology. The advantages offered by Windows are weighed alongside the risks. We examine specific technical risks to long-term application stability including memory leaks, page faults and stack pointer corruptions. The risks and benefits of using third-party software libraries to accelerate the development and implementation of control systems are evaluated. The trend towards the use of managed versus native code is examined along with its associated impact on application stability and testing methods. Specific counter-measures aimed at reducing the incidence of operating system instability and application fault are discussed including the use of memory analyzers, inter-locking application threads, state-based persistence recovery objects and watchdog timers. We detail some techniques for effectively employing these countermeasures toward the production of highly stable Windows-based control systems. We conclude by presenting the stability performance of a government-developed remotely operated weapon system during recent testing and qualification.