Modelling and Analysing Collaborating Heavy Machines

Cyber Physical Systems (CPS) consist of embedded computers and networks for controlling and monitoring the physical processes associated with it. In practice, these systems are complex in nature. Also most of the CPS systems handle concurrent and real time operations and they are mostly distributed in architecture. Suitable analysis techniques need to be used in order to design and develop a CPS system. Formal verification is a branch of software engineering which is used to verify the correctness of a design. Model checking is one technique used in formal verification, where a formal model of the real system is developed. This model is used for verify the correctness of the system design against its specified requirements. For model checking, a formal model and a formal specification language is required. There are several formalism available to represent concurrent distributed systems such as Petrinets, Actor models etc. Actor models are suitable for modeling the functional behaviour of distributed and asynchronous systems. If formal verification support could be incorporated with an actor model, it shall be beneficial in analyzing cyber physical systems. In this thesis, we use an actor based modelling language which is supported by formal semantics to analyze a complex cyber physical system. This study aims to analyze the distributed nature and concurrent behaviour of the system and not timing constraints. The system selected for this study is a safety critical system which consists of several autonomous machines which are operating in a fleet manner carrying out concurrent operations at a site. We also analyze the distributed design architecture of autonomous machine individually. The design implementation of the system is developed in a Robot Operating System and the actor based modelling language used for developing the models is ’Reactive Objects Language’ (Rebeca). Thus major tasks of the study include understanding the ROS based system, transformation of ROS elements to Rebeca models and verification of system specific properties. The study outcomes include identification of few mapping patterns between ROS and Rebeca. During this process a reusable algorithm describing the procedure of transforming any ROS code to Rebeca model was developed. Our study results hence proves the potential of Rebeca in verifying real robotic applications.