Jean-Philippe Babau

A DSML for reversible transformations

In this paper, we investigate a way to promote the reuse of legacy tools (or transformations) in specific contexts (defined by specific metamodels). More precisely we suggest a model transformation approach to achieve this purpose. We first introduce a language based on a metamodel called Modif in order to specify the differences between two semantically close metamodels. We can generate automatically data migration components from a Modif specification. They enable to put data complying with the specific context under the scope of the legacy tool. But more importantly in the case of a rewriting tool, they enable to put the tools outcome back into the original specific context. Then we propose a process and a set of helpers based on Modif to automate the reuse of legacy tools for domain-specific contexts. To illustrate this approach, we apply it to the case of simple finite state machines.

Optimizing the Deployment of Distributed Real-Time Embedded Applications

The synthesis of a valid and optimized deployment model from functional and platform models is a crucial issue in the development of distributed real-time systems. The synthesis consists in the allocation of functions/signals to execution nodes/communication buses, the mapping of functions/signals into tasks/messages and the priority assignment to tasks/messages. Current approaches provide partial solutions for the synthesis of a deployment model on distributed platforms, as either the allocation or the mapping is fixed a-priori. In order to tackle this problem we propose an optimization technique, based on two different mathematical programming formulations, to handle optimization of both allocation and mapping. The optimization is multi-objective and considers extensibility maximization, latency minimization and minimization of the number of tasks. The obtained solutions satisfy timing and platform resources requirements. An automotive case study shows the effectiveness of our approach.

Real-time design models to RTOS-specific models refinement verification

One key point of Real-Time Embedded Systems development is to ensure that functional and non-functional properties (NFPs) are satisfied by the implementation. For early detection of errors, the verification of NFPs is realized at the design level. Then the design model is implemented on a Real-Time Operating System (RTOS). However, the design model could be not implementable on the target RTOS. In this paper, we propose to integrate between the design and the implementation phases, a feasibility tests step to verify whether the design model is implementable on the target RTOS and a mapping step to generate the appropriate RTOS-specific model. This two-steps approach is based on an explicit description of the platform used for verification and the RTOS which is the implementation platform. Moreover an additional verification step is needed to ensure the conformity of the implementation model to the design model with regard to NFPs.

ORQA: modeling energy and quality of service within AUTOSAR models

Electric vehicles embed a low amount of energy, so their devices need to be managed efficiently to optimize the vehicle autonomy. A vehicle management is achieved by the embedded systems, modeled following the AUTOSAR standard. AUTOSAR covers most of the automotive concerns, but it lacks energy consumption and user-oriented Quality of Service models. This paper presents ORQA, a framework to model and manage the electric vehicle devices through energy consumption and user-oriented Quality of Service. At design time, the architects choose and tune the actual vehicle device models through their power requirements and, if appropriate, quality levels. The generated implementation is then embedded in the existing AUTOSAR models. Thus, at run-time, the vehicles system is able to evaluate the global consumption of a trip and to propose the user a specific driving strategy. The optional devices are managed throughout the trip, based on the driver preferences. ORQA is illustrated with a classic use-case: a work to home trip.

SRMP: a software pattern for deadlocks prevention inreal-time concurrency models

Model-based approaches for the development of software intensive real-time embedded systems allow early verification of timing properties at the design phase. At this phase, the Real-Time Operating System (RTOS) may not be chosen, hence some assumptions on the software platform are made to achieve timing verifications such as schedulability analysis of tasks describing the application. Among these assumptions, the synchronization protocol which is used to manage the concurrent access to resources that are shared between tasks. A classical solution is to consider the Priority Ceiling Protocol (PCP) synchronization protocol to avoid deadlocks. However, when this protocol is not provided by the target RTOS on which the application will be deployed, the concurrency model becomes not implementable and a new synchronization protocol must be considered. In this paper, we propose the Shared Resource Merge Pattern (SRMP) which aims to prevent deadlocks when the use of PCP protocol is not allowed by the target RTOS. The application of this pattern on the concurrency model must guarantee that the timing properties of the real-time application are still met.

VBR Video Abstraction for Home-Network Reservation

Home network reservation is classically based on token bucket policy. Token bucket parameter setting is a trade-off between maximizing the quality of service and optimizing the resource usage. In this paper, we propose dynamic parameter setting, based on a bitrate hull, following bitrate evolution. A shortest path algorithm is used to compute an optimal hull, according to implementation constraints, reservation period and number of reservations.

ImocaGen: A model-based code generator for embedded systems tuning

IMOCA is a model-based architecture model dedicated to embedded process control systems in disturbed environment. These systems depend on various parameters which are difficult to set because they are bound to environment changes. In this paper we propose to extend IMOCA with the meta-model ImocaGen for managing the aspects of the code generation. ImocaGen allows to target multiple platforms and different programming languages, generates both embedded code as well as tuning and reconfiguration tools, takes into account different communication protocols and offers a mechanism for integrating handwritten code. This approach is tested on a basic control application for a NXT brick for which three generations are performed: the first one for a PC with an USB connection, the second one for an Android tablet with a Bluetooth connection and the last one for a simulator in Java.

Modeling legacy code with BIP: how to reduce the gap between formal description and real-time implementation

To reduce the gap between high-level functional descriptions and real-time multitasking implementation, this paper proposes a set of modeling and code generation principles. Modeling principles are based on integration of a specific BIP concurrent component. This component follows a specific behavioral pattern based on periodic activation of data consumption, data processing and data production. It acts as a periodic task at execution stage. The pattern proposes two variants for eventtriggered and time-triggered platforms. The approach has been tested on three case studies, showing the interest of formalization for behavioral verification. The proposed pattern allows real-time validation and offers classical advantages of high-level modeling.

IMOCA: A Model-Based Code Generator for the Development of Multi-platform Marine Embedded Systems

Process control systems embedded in disturbed environments are usually developed case by case for specific deployment platforms and their behaviours closely depend on the characteristics of the environment. The obtained code is not portable and not reconfigurable. In order to help the software development of such applications, IMOCA offers architectural modelisation tools. The associated code generator allows to product adaptative and reconfigurable code for a simulator as well as embedded code for various platforms. This approach has been tested on NXT bricks, Arduino boards and Armadeus boards.

Improving reuse by means of asymmetrical model migrations: An application to the Orcc case study

The legacy code of a tool handling domain specific data gathers valuable expertise. However in many cases, it must be rewritten to make it apply to structurally incompatible data. We investigate a co-evolution approach to avoid this update by making the call context meet the a legacy tool definition domain. The data conforming to the call context co-evolve into data conforming to the definition domain. Once processed by the tool, they can be put back into their original context thanks to a specific reverse transformation which enables the recovery of elements that had been initially removed. This approach is applied to Orcc, a compiler for dataflow applications. Orcc requires many common functions that are expected to be adapted to its own context. Our approach is an effective way to reuse them instead of rewriting them.