Improving Formal Verification and Testing Techniques for Internet of Things and Smart Cities

Abstract

We are interested in formal verification and model-based testing for Internet of Things and Smart Cities. In general these two techniques suffer from state explosion problem. To remedy this situation we propose a set of techniques which aim to reduce the cost, duration and complexity of the considered problems. On the first hand the techniques realted to formal verification are as follows. First, Abstraction consists in modelling a part of the system accurately and the other parts at high level. Second, Modularization and Compositionality consist in splitting the whole system into smaller subsystems. Third, Symmetry Detection exploits symmetries that take place during the system execution. Fourth, Data Independence consists in detecting that the behaviour of the considered system does not depend on some data inputs. Fifth, Eliminating Functional Dependencies consists in removing dependency among state variables. Sixth, Exploiting Reversible Rules consists in collapsing subgraphs of the graph of states into abstract states. On the second hand the techniques related to model-based testing are as follows. First, Refinement Techniques extract test scenarios directly from the untimed specification. Second, the Reduction of the Size of Digital-Clock Tests Technique provides a heuristic to reduce the size of the generated tests. Third, the Timed Automata Testers Generation Technique allows to produce testers in the form of deterministic timed automata. Fourth, the Test Cases Updating Technique after System Evolution makes it possible to reduce the number of tests to be generated after each adaptation. Fifth, the Resource Aware Test Component Placement Technique allows to produce a placement plan of the different testers. Sixth, Coverage Technique generates a reasonable-size set of tests. A case study is proposed in order to illustrate the use of these techniques.