A Framework for Developing Cyber Physical Systems

Abstract

Abstract-Cyber physical systems (CPSs) are pervasive in our daily life from mobile phones to auto driving cars. CPSs are inherently complex due to their sophisticated behaviors and thus difficult to build. In this paper, we propose a framework to develop CPSs based on a model driven approach with quality assurance throughout the development process. An agent-oriented approach is used to model individual physical and computation processes using high level Petri nets, and an aspect-oriented approach is used to integrate individual models. The Petri net models are systematically mapped to classes and threads in Java, which are enhanced and extended with domain specific functionalities. Complementary quality assurance techniques are applied throughout system development and deployment, including simulation and model checking of design models, model checking of Java code, and run-time verification of Java executable. We demonstrate our framework using a car parking system.

A Systematic Approach for Developing Cyber Physical Systems

“…To effectively model and analyze the complex behaviors of CPSs, many modeling techniques have been proposed and adapted in recent years including formal methods such as hybrid automata [6] and special graphical modeling languages such as actor-oriented MoC [7]. High level Petri nets [8] are well suited to model the complex behaviors of CPSs, especially combined with well-established software engineering approaches such as agent-oriented approach and aspectoriented approach [3]. However most existing techniques only provide very general guidelines and lack fine grained rules.…”

“…With regard to behavioral programming, we can define many generic safety and liveness property patterns using linear time temporal logic, and then instantiate the patterns using concrete behaviors and check them using SPIN model checker. Some generic property patterns include (where B, B1, B2 denote place names representing different behaviors, x and y denote symbolic tokens): (1) <>B(x) (G1) This liveness property states that a behavior B will eventually active; (2) [](B(x) → <>!B(x)) (G2) This liveness property states that an active behavior B will eventually terminate; (3) []! (B1(x) ∧ B2(y)) (G3) This safety property states that two behaviors B1 and B2 cannot be active at the same time due to the sequential nature of behavioral programming; (4) [](B1(x) → <> B2(y)) (G4) This liveness property states that a behavior B1 leads to behavior B2.…”

“…However since there are multiple UML notations such as class diagram, state machine diagram, and sequence diagram for representing different aspects of a system, it is not easy to obtain a coherent set of code. In [3], we presented a model driven approach to realize our high level Petri net models, which provided a systematic way of writing Java programs and establishes the traceability between the models and resulting programs. Our model driven approach consists of the general code structure and domain specific refinement.…”

“…However the domain specific refinement requires manual process in identifying and defining additional features of the system, especially with regard to the physical devices. Different from [3] where each agent net was mapped to a thread in Java, this paper maps each behavior net into a behavior program [11] that includes 3 template methods: boolean takeControl(),void action(), and void suppress().…”

“…In [3], we provided a concrete framework to realize the ideas in [2], where a model driven approach from high level Petri nets to Java programs was presented. Essential CPS design issues including concurrency and timing are modeled using high level Petri nets and analyzed through model checking and simulation.…”