<\/a><\/p>\nSmart factory (i.e., Industry 4.0)<\/strong> is a revolutionary trend of manufacturing industries, which promises to deliver a more responsive, adaptive, and connected manufacturing line. New market requirements and emerging autonomously technologies such as IoT are shifting the manufacturing companies\u2019 environment toward smart factories. Thanks to IoT physical objects are seamlessly integrated into the information network where they can become active participants in business processes, communicate information about their status, surrounding environment, production processes, maintenance schedule and even more. In addition, rising energy prices, increasing ecological awareness, and changing consumer behaviors toward greener products are driving decision makers to put green manufacturing and energy efficient production processes at the top of their priorities. Furthermore, unlike the way in which existing factories provide products to customers, situations are changing in which factories produce and provide customized products to meet the needs of customers of late years. As a representative example, Jun Ji-hyun’s lipstick, which appeared in the drama “My Love from the Star,” has gained huge popularity in Korea, Japan and China. However, the company has never experienced inventory problems because the company analyzed customers’ needs and increased production in advance by conducting a preliminary survey of SNS such as Instagram and Facebook. As such, Smart Factory should enable efficient production of products according to customer needs.<\/p>\nThere are four main components in this smart factory scenario. First is robots which are considered as autonomous constituent systems (CSs) of this scenario. In this scenario, there are two organizations, Manufacturing and Distribution. In manufacturing organization, we assume that there are three constituent systems, component management robots, process robots, packaging robots. In distribution organization, we assume that there are four CSs, stock management robots, cleaning robots, transferring robots (e.g., KIVA system), delivery robots (e.g., Amazon Prime Air). Second one is policies generated by analyzing customer behaviors. In this smart factory scenario, this smart factory can generate its policies by analyzing customers\u2019 behaviors in social networks. The third component is energy suppliers for robots. Each robot automatically charge itself before discharged. The last part is cloud computing platform that enables communication between robots.<\/p>\n
<\/p>\n
(2) Real-world Example: SoS for Autonomous Collaborative Robots<\/p>\n
<\/p>\n
We consider a scenario that autonomous robots collect trash with cooperation. This robot scenario simplifies the main features of the smart factory SoS scenario. First, there are robots that collect trash at the sea as autonomous constituent systems (CSs). Each robots communicate to other robots for sharing the existence of trash and its positions. Second, the policies of the SoS affect the cooperative behavior of this robot system. In the smart factory scenario, the policies of SoS are automatically changed by the analysis of customer behaviors in SNS. However, in this robot scenario we assume that SoS managers set the policies of SoS. Third, this robot scenario also considers the energy issue same as smart factory. There are energy supplies that robots should revisit before they are discharged and SoS managers can set a policy of robot behaviors for minimizing the energy consumption. Lastly, we assume that each robot communicate by using internet connection at the cloud computing platform.<\/p>\n
<\/p>\n
(3)Why is the autonomous collaborative robot scenario an SoS?<\/p>\n
\n\n\nDimension<\/td>\n | Observations in the scenario<\/td>\n<\/tr>\n |
\nAutonomy of constituents<\/td>\n | Constituent systems in SIMVA-SoS can have range of actions. These actions need not be known at the SoS level, except the actions that are used to perform a role in SIMVA-SoS.\u00a0For example, the robots in this scenario could have range of actions including traversing a specific area, collecting garbage, recharging, sending messages to nearby robots, etc. SIMVA-SoS considers only those actions which are used to fulfill the common goal, i.e. the SoS level goal, and only those the constituent systems publically use to perform their role.<\/td>\n<\/tr>\n |
\nIndependence<\/td>\n | Constituent systems in SIMVA-SoS perform their action to fulfill specific role identified\/decomposed from the SoS level goal.\u00a0In doing so, regardless of how they do it, the only requirement is communicating state of their action to related constituent systems. For example, how a robot perform the searching and collecting operation with or without other robots is encapsulated in its action. Independently they perform their action, and only communicate action status to the respective constituent systems.<\/td>\n<\/tr>\n |
\nDistribution<\/td>\n | Constituent systems in SIMVA-SoS exchange only information using the infrastructure provided by SIMVA-SoS. For example, a robot communicate with other robot by using the Cloud infrastructure. Thus, SIMVA-SoS provide\u00a0distribution\u00a0<\/em>by enabling constituent systems communicate concurrently on logical time.\u00a0Each constituent systems have a shared logical time tick to perform their atomic action at once.<\/td>\n<\/tr>\n\nEvolution<\/td>\n | The SIMVA-SoS simulation engine always keep track of constituent systems activity logs at each tick, and this enables to create consistent evolution steps and preservation of specified properties.\u00a0In the collaborative robot scenario, at each tick SoS state update, for example additional trash founded, is communicated to all constituent systems (i.e., robots). The logs and update information helps to maintain and monitor evolution steps preserved.<\/td>\n<\/tr>\n | \nDynamic Behavior<\/td>\n | SoS structure and composition changes dynamically as constituent systems make autonomous decisions. SIMVA-SoS provide a mechanism to manage the changes and direct them to the intended SoS level goal. In the collaborative robot scenario, if a robot is broken, the simulation engine can easily recognize it by counting the missed self-status notification. The simulation engine, thus, free SoS level resources which were assigned to that specific constituent system (this can include restructuring and rescheduling the remaining constituent systems, or admitting if there are constituent systems that can play the role).<\/td>\n<\/tr>\n | \nEmergence of Behavior<\/td>\n | SIMVA-SoS identifies emergent behavior based on SoS level goal progresses. In the collaborative robot scenario, certain robots performing searching <\/em>action, and some other robots performing collecting<\/em> action contribute to the progress of SoS level goal only when the actions are synergized.\u00a0 The simulation engine compute and determines the cumulative contribution of each constituent systems actions to the progress of SoS level goal by analyzing log data.<\/td>\n<\/tr>\n\nInterdependence<\/td>\n | By analyzing log data, SIMVA-SoS provide explicit identification of interdependence, the tracing of mutual dependencies, and the ability to use these links to assess the impact of constituent system changes. When some robots search garbage and sends signal to other robots, activity details are added to the log database. Similarly, when robots given messages collect and transfer garbage, activity details are added to the log database. Since these two activities are related and contribute to the SoS level goal progress, SIMVA-SoS can determine interdependences, and can assess the impact of either of the constituent system performance in fulfilling their role.<\/td>\n<\/tr>\n | \nInteroperability<\/td>\n | SIMVA-SoS assumes heterogeneous constituent systems in terms of capability and role, but having identical model and semantics.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n <\/p>\n 2. Goals & Requirements<\/strong><\/h5>\n(1) SoS level goal<\/p>\n | | |