Godkänd; 1997; 20070418 (ysko)

As real-time systems become more complex, the need for more sophisticated runtime kernel features arises. One such feature that substantially lessens the burden of the programmer is automatic memory management, or garbage collection. However, incorporating garbage collection in a real-time kernel is not an easy task. One needs to guarantee, not only that sufficient memory will be reclaimed in order to avoid out of memory errors, but also that the timing properties of the systems real-time tasks are unaffected. The first step towards such a garbage collector is to define the algorithm in a manageable way. It has to be made incremental in such way that induced pause times are small and bounded (preferably constant). The algorithm should not only be correct, but also provably useful. That is, in order to guarantee that sufficient memory is reclaimed each time the garbage collector is invoked, one need to define some measure of usefulness. Furthermore, the garbage collector must also be guaranteed to be schedulable in the system. That is, even though the collector is correct and proved useful, it still has to be able to do its work within the system. In this thesis, we present a model of an incremental copying garbage collector based on process terms in a labeled transition system. Each kind of garbage collector step is captured as an internal transition and each kind of external heap access (read, write, and allocate) is captured as a labeled transition. We prove correctness and usefulness of the algorithm. We also deploy the garbage collector in a real-time system, to wit, the runtime kernel of Timber. Timber is a strongly typed, object-oriented, purely reactive, real-time programming language based on reactive objects. We show how properties of the language can be used in order to accomplish very efficient and predictable garbage collection. / <p>Godkänd; 2007; 20071121 (ysko)</p>

The complexity of software running on vehicular embedded systems is constantly increasing and this negatively affects its development costs and time to market. One way to deal with these issues is to boost abstraction in the form of models to (i) ease the reasoning about the system architecture, (ii) automate certain stages of the development, (iii) early detect flaws in the system architecture through fundamental analysis and (iv) take appropriate countermeasures before the system is implemented. Considering the importance of timing requirements in the design of software for vehicular embedded systems, in this licentiate thesis we leverage Model-Driven Engineering for realizing a semi-automatic approach which allows the developer to perform end-to-end delay timing analysis on design models, without having to manually model timing elements and set their values. The proposed approach, starting from a design model of an automotive software functionality, automatically generates a set of models enriched with timing elements whose values are set at generation time. End-to-end delay timing analysis is run on the generated models and, based on the analysis results, the approach automatically selects the generated models which better meet a specific set of timing requirements.

Today, wireless solutions for industrial networks are becoming more and more appealing since they increase flexibility and enable the use of additional wireless sensors, but also bring such advantages as mobility and weight reduction. Wired networks, on the other hand, are reliable and, more importantly, already existing in most distributed control loops. Heterogeneous networks consisting of wireless as well as wired sub-networks are gaining attention as such networks combine the advantages of both approaches. However, wireless communication links are more vulnerable to security breaches because of their broadcast nature. For this reason, industrial heterogeneous networks require a new type of security solutions, since they have different system assets and security objectives. This thesis aims to secure industrial heterogeneous networks. Such networks have real-time requirements due to interaction with some physical process, and thus have a schedule with one or more deadlines for data delivery in order to comply with the timing requirements of the application. The necessity to follow the schedule implies that all network participants should share the same notion of time and be synchronized. This fact makes clock synchronization a fundamental asset for industrial networks. The first step towards developing a security framework for industrial heterogeneous networks with real-time requirements is therefore to investigate ways of breaching clock synchronization. Once the vulnerabilities of this asset have been identified, the next step is to propose solutions to detect malicious attacks and mitigate their influence. The thesis provides a vulnerability analysis of the asset synchronization based on the widely deployed IEEE 1588 standard, and identifies a possibility to break clock synchronization through a combination of a man-in-the-middle attack and a delay attack. This attack is appealing to an adversary as it can target any network requiring synchronization. Next, several mitigation techniques, such as a relaxed synchronization condition mode, delay bounding and using knowledge of existing environmental conditions, are identified, making the network more resilient against these kinds of attacks. Finally, a network monitor aiming to detect anomalies introduced by the adversary performing attacks targeting clock synchronization is proposed as a mean to detect the delay attack.

An experiment was carried out, attempting to ascertain whether the Leap Motion Controller can be a useful input device for dynamically controlling graphic visualizations, e.g. by artists who use video and interactive visual arts to enhance music performances. The Leap Motion Controller was found to be too unreliable to be used as the primary controller in a professional visual arts performance.

In this thesis a novel approach is developed to detect urethral stones based on a computer-aided process. The input data is a CT scan from the patient, which is a high-resolution 3D grayscale image. The algorithm developed extracts the regions that might be stones, based on the intensity values of the pixels in the CT scan. This process includes a binarizing process of the image, finding the connected components of the resulting binary image and calculating the centroid of each of the components selected. The regions that are suspected to be stones are used as input of a CNN, a modified version of an ANN, so they can be classified as stone or non-stone. The parameters of the CNN have been chosen based on an exhaustive hyperparameter search with different configurations to select the one that gives the best performance. The results have been satisfactory, obtaining an accuracy of 98,3%, a sensitivity of 99,5% and a F1 score of 98,3%.

Web applications are becoming more and more complex and bigger and bigger. In the case of a web game, it can be as big as a software. As well, as these applications are run indirectly through a web browser they need to be quite well optimized. For these reasons, performance and architecture are becoming a crucial point in web development. This report covers some method to improve the architecture of a javascript application by using design patterns. It will also present some methods to optimize the performance of javascript code and some way to do some of these changes automatically. Applying these changes allows us to increase the framerate of the web game by 50% and to reduce the loading time of the game by more than 72%. The usage of design patterns such as the observer, the composite and the builder pattern permits to improve the maintainability, extendibility and flexibility of the code but also allows to implement new features.

Tamper-resistance is the implementation of some kind of security against physical attacks on a device. Most tamper-resistance methods are focused on preventing the theft of data or keys from a device, while others are focused on detecting or responding to an attack. The focus of our review is mainly on resistance against a number of attack-types that are regarded as the main threats in the area. In our review we first start by categorising the different tampering attacks and later on listing different methods of resistance that have been of interest in the last decade. We also describe how the different methods of resistance work and which different attacks they counter. Lastly there is a table that contains the various attacks and defences that will give the reader an easy overview of our review.

This thesis introduces a system to identify objects into a 3D reconstructed model. In particular, this is applied to automatize the inspection of an engine of a truck by detecting some parts in an online reconstructed 3D model. In this way, the work shows how the use of the augmented reality and the computer vision can be applied into a real application to automatize a task of inspection. To do this, the system employs the Signed Distance Function for the 3D representation which has been proven in other research as an efficient method for 3D reconstruction of environments. Then, some of the common processes for the recognition of shapes are applied to identify the pose of a specific part of the 3D model. This thesis explains the steps to achieve this task. The model is built using an industrial robot arm with a depth camera attached to the end effector. This allows taking snapshots from different viewpoints that are fused in a same frame to reconstruct the 3D model. The path for the robot is generated by applying translations to the initial pose of the end effector. Once the model is generated, the identification of the part is carried out. The reconstructed model and the model to be detected are analysed by detecting keypoints and features descriptors. These features can be computed together to obtain several instances over the target model, in this case the engine. Last, these instances can be filtered by the application of some constrains to get the true pose of the object over the scene. Last, some results are presented. The models were generated from a real engine truck. Then, these models were analysed to detect the oil filters by using different keypoint detectors. The results show that the quality of the recognition is good for almost all of the cases but it still presents some failures for some of the detectors. Keypoints too distinctive are more prune to produce wrong registrations due to the differences between the target and the scene. At the same time, more constrains make the detection more robust but also make the system less flexible.

The code size of single page web applications are constantly growing which do have an negative effect on the load time. Previous research have shown that load time are important to users and that a slow application will lose potential customers even before it has loaded. In this paper three architecturally far-reaching techniques are measured to see how they can improve the load time and help to decide if an application should be built with one or more of the tested techniques which are HTTP2 push, Code Splitting and Isomorphism. The experiment shows that Isomorphism can provide a big improvement for the time to first paint and that Code Splitting can be a useful technique for large code bases on mobile phones.