Short-term availability and performability analysis for a large-scale multi-state system based on robotic sensors

Abstract

Abstract This paper presents dynamic (short-term) availability and performability analysis for highly responsive large-scale distributed robotic systems. Such systems are intended for the exploration and measurement of different parameters in aggressive environments by using robotic sensors. The typical system, which is considered in this paper, consists of N platforms and m robotic sensors, which initially are installed on each platform. Platforms are distributed on a safe distance from the aggressive (dangerous) environment to be explored. At some time, sensors should be shot from the platform to the environment. Then each platform begins to process all data gathered by its sensors. Each sensor, when it is installed on a platform, and each platform are individually represented as a three-state component – completely available state, minor repairable state, and replacement state. Thus, the entire system and its components are interpreted as a multi-state system. In general, numbers N and m may be enough big. In such cases, the system is large-scale. The problem is to find the short-term availability and the performance for an entire large-scale system which strongly depends on the initial conditions of the system. It has been shown, that there is a great difference between steady-state and short-term availability and performance indices which significantly depend on the initial states of all robotic sensors and platforms. For a highly responsive system, this difference is very important. The main obstacle in finding a solution to the problem is its large multi-dimensionality. To overcome this difficulty, the Lz-transform method was used. This paper suggests a multi-state model for the system and measures for system availability and performance assessment. The paper has also developed an algorithm for an Lz-transform application to evaluate these measures and presents recommendations for the prevention of availability and performance decrease in the initial short-time period. A numerical example is presented to illustrate the proposed method.