Abstract

Abstract
In an embedded system, functions often operate under different requirements. In the extreme, a failing safety critical function may cause collateral damage (and hence consider to be a system failure) while non critical functions affect only the quality of service. Approaches by partitioning the system's functions into sandboxes require virtualization mechanisms by the underlying platform and thus prohibit deployment to the bulk of microcontroller based systems. In this paper we discuss an alternative approach based on static semantic analysis performed directly on the system specification expressed in the form of an object oriented (00) model in the experimental language RTFM-lang. This would allow to (at compile time) to discriminate in between critical and non-critical functions, and assign these (by means of statically checkable typing rules) appropriate access rights. In particular, one can imagine dynamic memory allocations to be allowed only in non-critical functions, while on the other hand, direct interaction with the environment may be restricted to the critical parts. With respect to scheduling, a static task and resource configuration allows e.g. Stack Resource Policy (SRP) based approaches to be deployed. In this paper we discuss how this can be achieved in a mixed critical setting.

Abstract
The analysis of the spatial structure of animal communities requires spatial data to determine the distribution of individuals and their limiting factors. New technologies like very precise GPS as well as satellite imagery and aerial photographs of very high spatial resolution are now available. Data from airborne LiDAR (Light Detection and Ranging) sensors can provide digital models of ground and vegetation surfaces with pixel sizes of less than 1m. We present the first study in terrestrial herpetology using LiDAR data. We aim to identify the spatial patterns of a community of four species of lizards (Lacerta schreiberi, Timon lepidus, Podarcis bocagei, and P. hispanica), and to determine how the habitat is influencing the distribution of the species spatially. The study area is located in Northern Portugal. The position of each lizard was recorded during 16 surveys of 1 h with a very precise GPS (error<1 m). LiDAR data provided digital models of surface, terrain, and normalised height. From these data, we derived slope, ruggedness, orientation, and hill-shading variables. We applied spatial statistics to determine the spatial structure of the community. We computed Maxent ecological niche models to determine the importance of environmental variables. The community and its species presented a clustered distribution. We identified 14 clusters, composed of 1-3 species. Species records showed two distribution patterns, with clusters associated with steep and flat areas. Cluster outliers had the same patterns. Juveniles and subadults were associated with areas of low quality, while sexes used space in similar ways. Maxent models identified suitable habitats across the study area for two species and in the flat areas for the other two species. LiDAR allowed us to understand the local distributions of a lizard community. Remotely sensed data and LiDAR are giving new insights into the study of species ecology. Images of higher spatial resolutions are necessary to map important factors such as refuges.

Abstract
Hard real- time multiprocessor scheduling has seen, in recent years, the flourishing of semi-partitioned scheduling algorithms. This category of scheduling schemes combines elements of partitioned and global scheduling for the purposes of achieving efficient utilization of the system's processing resources with strong schedulability guarantees and with low dispatching overheads. The sub-class of slot-based "task-splitting" scheduling algorithms, in particular, offers very good trade-offs between schedulability guarantees (in the form of high utilization bounds) and the number of preemptions/migrations involved. However, so far there did not exist unified scheduling theory for such algorithms; each one was formulated in its own accompanying analysis. This article changes this fragmented landscape by formulating a more unified schedulability theory covering the two state-of-the-art slot-based semi-partitioned algorithms, S-EKG and NPS-F (both fixed job-priority based). This new theory is based on exact schedulability tests, thus also overcoming many sources of pessimism in existing analysis. In turn, since schedulability testing guides the task assignment under the schemes in consideration, we also formulate an improved task assignment procedure. As the other main contribution of this article, and as a response to the fact that many unrealistic assumptions, present in the original theory, tend to undermine the theoretical potential of such scheduling schemes, we identified and modelled into the new analysis all overheads incurred by the algorithms in consideration. The outcome is a new overhead-aware schedulability analysis that permits increased efficiency and reliability. The merits of this new theory are evaluated by an extensive set of experiments.

Abstract
Several methodologies have been proposed in the last decades to improve the quality of Safety-Critical Embedded Systems (SCES) and, at the same time, keep costs and schedule compatible with project plans. In particular, approaches such as Product Line Engineering (PLE) and Model-Driven Engineering (MDE) offer an interesting solution to reduce development complexity and time to market due to their synergy and common goals. However, the current state of how MDE and PLE can be combined to enhance productivity in the domain of SCES is not clear yet. This paper presents a systematic literature review, with the purpose of obtaining the state of the art of the aproaches, methods and methodologies whose goal is the combination of PLE and MDE for the development of SCES, and to verify the existence of empirical studies that demonstrate the application of these techniques in this type of development. We drew the following conclusions from the review results: (1) The number of studies using PLE with MDE to build SCES is relatively small, but has increased gradually in recent years. (2) The approaches diverge about what is needed to build Model-driven Product Lines. (3) Most of the approaches do not consider to differentiate between hardware and software variabilities. (4) Most of the studies propose the use of UML and feature diagrams. (5) The studies present case studies implemented in different tools and most of them are free. (6) The approaches do not cover the entire development lifecycle.

Abstract
The development of Critical Embedded Systems (CES) like Unmanned Aerial Vehicles (UAV) is complex because it needs to ensure a high degree of quality, with affordable cost and delivery time. It is also necessary to ensure security since failures in this type of system can lead to catastrophic results. In this sense, a Model-Driven Development (MDD) approach presents itself as a good alternative to the traditional development because coding complexity will be reduced by the use of high level models. In addition, it avoids the introduction of coding errors by human programmers, since the critical code will be built automatically through models transformation. From another perspective, Embedded Systems Development can benefit from Software Engineering techniques like Product Lines to reduce costs and time-to-market. While other works propose the use of Product Line techniques to improve Embedded Software development, we propose a Product Line approach to the whole Critical Embedded System development life cycle, including hardware variability management. Therefore, this paper proposes a Critical Embedded System Product Line Model Based approach, which aims to reduce the above mentioned challenges. The development approach proposes a Domain Engineering and Application Engineering focused on the system, with both software and hardware. To illustrate the proposed approach we include some artifacts from a case study in the UAV domain.