SCEL Modeling and jRESP Programming. In this section, we outline how SCEL [DLPT13, DFLP11b] can be used to model the robotic scenarios. In the considered model, each robot is described via a SCEL component. To model an autonomic system in SCEL, the first step is to identify the basic attributes that charac- terize the relevant aspects of Autonomic Components (ACs) involved. ACs are entities with dedicated knowledge units and resources. ACs can cooperate while playing different roles. Awareness is enabled by providing ACs with information about their own state and behavior; such information is stored in each AC’s knowledge repository. These repositories also allow ACs to store and retrieve information about their working environment. Ultimately, this information is used to redirect and adapt their be- havior. Each AC is equipped with an interface, consisting of a collection of attributes, such as provided functionalities, spatial coordinates, group memberships, trust level, response time, etc. Attributes are used by the ACs to dynamically organize themselves into autonomic-component ensembles (ACEs). These are sets of ACs featuring goal-oriented execution. ACEs are highly dynamic and flexible, and not curbed by rigid structures. These attributes are typically identified during the requirements analysis and are related to the set of goals and utilities that are part of SOTA specification. In the robot scenarios, one of the main attributes is the task that a robot is performing. For instance, in the hard exploration variant, when individual mapping is not possible, possible values for this attribute, that we call task, are: exploring, networking, searching and rescuing. The value of attribute task is determined by what the robot sensors perceives from the enclosing environment. The schema of robot behaviour is the following: a group of robots starts in task explor- ing and performs a random walk across the area. A robot switches to task networking when either it perceives a victim, or it detects a given number of robot invoked in task networking. The following code segment is the SCEL process identifying the exploring task: Pexp(id) ¾ //While executing this behaviour a robot moves following a random walk. get(“victim sensed”)@self. //When a victim is perceived, the robot stops moving put(“stop”)@self. //The fact that a victim has been directly found is published in the local knowledge put(“victim”, 0, id)@self. nil + qry(“victim”, ?d, ?r)@true. //A robot also stops the presence of a victim is ...
