Søren Top

I/O Sharing in a Multi-core Kernel for Mixed-Criticality Applications

In a mixed-criticality system, applications with different safety criticality levels are usually required to be implemented upon one platform for several reasons( reducing hardware cost, space, power consumption). Partitioning technology is used to enable the integration of mixed-criticality applications with reduced certification cost. In the partitioning architecture of strong spatial and temporal isolation, fault propagation can be prevented among mixed-criticality applications (regarded as partitions). However, I/O sharing between partitions could be the path of fault propagation that hinders the partitioning. E.g. a crashed partition generates incorrect outputs to shared I/Os, which affects the functioning of another partition. This paper focuses on a message-based approach of I/O sharing in the HARTEX real-time kernel on a multi-core platform. Based on a simple multi-core partitioning architecture, a certifiable I/O sharing approach is implemented based on a safe message mechanism, in order to support the partitioning architecture, enable individual certification of mixed-criticality applications and thus achieve minimized total certification cost of the entire system.

Trajectory tracking for autonomous turf-care vehicle using Liouvillian approach

Autonomous mobile robots and vehicles are a longstanding but recently reinvigorated research area, due in part to the commercialization of sensors technologies and processing power. In this work, a fully autonomous turf-care robot is used as a basis for development of a trajectory control solution with application to differential-drive robots with displaced end-effectors. A kinematic model for the vehicle is derived as well as expressions for the movement of the tool (end-effector), shown to be a system of the Liouvillan type. The Liouvillan model allows for the development of a feed-forward controller and feed-back controller, which are combined to allow trajectory tracking based on an arbitrary linear-segmented path. The path will ideally be a covering path and can be generated by an algorithm described in previous work. To validate the model and controller, the control solution is tested by numerical simulation against a kinematic model in MATLAB/Simulink and additionally against a dynamic 6-DOF model in OSRF Gazebo software. The developed controller enables sufficient tool trajectory-tracking in the kinematic model, but there are significant oscillations and deviations when used with a dynamic model, warranting further work on the feedback controller.

Covering path generation for autonomous turf-care vehicle

A covering path generation algorithm is developed to generate a lengthwise pattern based on a polygon describing the outer boundary and obstacles (polygon holes) of a geographical area. The algorithm is applied to an autonomous lawn-care robot for application to large grass turfs, for example golf-courses, which require structured and precise cutting patterns. The geographical polygon is recorded by manually driving the vehicle around the contour, resulting in a polygon given as geographical (latitude, longitude) coordinates of the vertices, which together with machine parameters are used to generate a suitable toolpath. The algorithm has been tested on a recorded polygon from a local park turf which is non-convex and has holes, illustrating the algorithm functionality and limitations wrt. optimality. In particular, the algorithm can generate a tool-path for any polygon orientation.

A distributed multi-agent linear bi-objective algorithm for energy flow optimization in microgrids

A distributed linear bi-objective optimization algorithm for management of energy flow in a microgrid with individual agents is introduced. In contrast to widely used centralized approaches for energy management a distributed multi-agent scheme for optimization of energy flow within a microgrid consisting of local energy resources and storage capacities is presented which is based on the auction algorithm for assignment problems originally introduced by Bertsekas in 1979 [1]. It is shown that the topology of a microgrid can be represented as a bipartite graph and mathematically be described as a classical transportation problem. This allows applying an auction algorithm scheme in a distributed way where each energy supply system node is either a source or a sink and is represented by an individual acting agent. The single-objective approach is extended towards bi-objectivity to build a framework which gives each agent freedom and authority to intelligently influence the global decision making (optimization) with respect to its own individual objectives and therefore a distributed bi-objective optimization for the energy flow in the microgrid with individual agent objectives is achieved. In the introduced framework computational intelligent agents enhance the intelligence of the global decision making based on their individual agents intelligence.

Towards spatial isolation design in a multi-core real-time kernel targeting safety-critical applications

In mixed-criticality systems, applications naturally have different safety criticality levels. Partitioning technology is usually used to enable the integration of such mixed criticality applications upon one platform, aiming at reducing hardware, power consumption and especially certification cost. Partitioning can prevent fault propagation among mixed-criticality applications, if spatial and temporal isolation are adequately ensured. This paper focuses on the solution of spatial isolation in the HARTEX kernel on a multi-core platform in terms of memory, communication between applications and I/O sharing. According to formulated isolation requirements, a simple partitioning multi-core hardware architecture is proposed using SoC and memory protection units, and the kernel is extended to support spatial isolation between the kernel and applications as well as between applications. Combined design of hardware and software can easily achieve this isolation. At last, the spatial isolation is evaluated using a statistical sampling method and its performance is tested in terms of task switch, system call and footprint.

Component-based analysis of embedded control applications

The widespread use of embedded systems requires the creation of industrial software technology that will make it possible to engineer systems being correct by construction. That can be achieved through the use of validated (trusted) components, verification of design models, and automatic configuration of applications from validated design models and trusted components. This design philosophy has been instrumental for developing COMDES—a component-based framework for distributed embedded control systems. A COMDES application is conceived as a network of embedded actors that are configured from instances of reusable, executable components—function blocks (FBs). System actors operate in accordance with a timed multitasking model of computation, whereby I/O signals are exchanged with the controlled plant at precisely specified time instants, resulting in the elimination of I/O jitter. The paper presents an analysis technique that can be used to validate COMDES design models in SIMULINK. It is based on a transformation of the COMDES design model into a SIMULINK analysis model, which preserves the functional and timing behaviour of the application. This technique has been employed to develop a feasible (light-weight) analysis method based on runtime observers. The latter are conceived as special-purpose actors running in parallel with the application actors, while checking system properties specified in Linear Temporal Logic. Observers are configured from reusable FBs that can be exported to SIMULINK in the same way as application components, making it possible to analyze system properties via simulation. The discussion is illustrated with an industrial case study—a Medical Ventilator Control System, which has been used to validate the developed design and analysis methods.

A Heterogeneous Multi-core Architecture with a Hardware Kernel for Control Systems

Abstract Rapid industrialisation has resulted in a demand for improved embedded control systems with features such as predictability, high processing performance and low power consumption. Software kernel implementation on a single processor is becoming more difficult to satisfy those constraints. This paper presents a multi-core architecture incorporating a hardware kernel on FPGAs, intended for high performance applications in control engineering domain. First, the hardware kernel is investigated on the basis of a component-based real-time kernel HARTEX (Hard Real-Time Executive for Control Systems). Second, a heterogeneous multi-core architecture is investigated, focusing on its performance in relation to hard real-time constraints and predictable behavior. Third, the hardware implementation of HARTEX is designated to support the heterogeneous multi-core architecture. This hardware kernel has several advantages over a counterpart kernel implemented in software: higher-speed processing capability, parallel computation, and separation between the kernel itself and the applications being run. A microbenchmark has been used to compare the hardware kernel with the software kernel, and compare the kernel on the multi-core platform with the kernel on the single-core platform.

Simulink analysis of component-based embedded applications

The widespread use of embedded systems requires the creation of industrial software technology, which will make it possible to engineer systems that are correct by construction. That can be achieved through the use of validated (trusted) components, verification of design models and automatic configuration of applications from validated design models. These guidelines have been instrumental for developing COMDES - a component-based framework for real-time embedded control systems. In this framework, an application is conceived as a network of distributed embedded actors that communicate with one another by means of labeled messages (signals), whereby I/O signals are exchanged with the controlled plant at precisely specified time instants, resulting in the elimination of I/O jitter. The paper presents an analysis method that can be used to validate COMDES design models using the Simulink environment. It is based on a semantics-preserving transformation of a COMDES design model into a Simulink analysis model, which preserves both the functional and timing behaviour of the original design model. The discussion is illustrated with an industrial case study -- a Medical Ventilator Control System, which has been used to validate the developed design and analysis methods.