580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How can real-time operating systems optimize task scheduling and resource management for real-time embedded systems to guarantee deadline adherence?
This research area focuses on developing scheduling models, algorithms, and resource management techniques in real-time operating systems (RTOS) that ensure timely execution of tasks in embedded systems. It addresses challenges like dynamic task arrivals, multiprocessor scalability, unpredictable workloads, and meeting hard deadlines without compromising system stability. The aim is to provide efficient scheduling, minimal overhead, and guaranteed temporal correctness critical for reliability in safety-critical real-time embedded systems.
Key finding: Proposes a dynamic scheduling model recognizing the stochastic and irregular arrival of real-time processes in embedded systems. It distinguishes periodic and aperiodic processes and emphasizes the importance of time-valued... Read more
Key finding: Introduces HIPPEROS, a scalable, configurable real-time micro-kernel designed from scratch for multi-core embedded platforms, supporting parallel execution of kernel operations and fine-grained resource management. The... Read more
Key finding: Demonstrates the design and benchmarking of open embedded RTOS real-time controllers based on Xenomai and low-cost hardware platforms. It evaluates critical real-time metrics such as periodicity, schedulability, task... Read more
Key finding: Presents an enhanced fixed-priority scheduling framework that integrates dynamic priority adjustments based on task criticality and temporal constraints, reducing the chances of deadline misses in embedded real-time... Read more
Key finding: Develops a Constraint Programming-based systematic approach to generate worst-case input scenarios stressing real-time embedded systems for task deadline misses. Compared with genetic algorithm methods, CP guarantees... Read more
2. What are the impacts and challenges of integrating real-time operating systems with embedded multiprocessor and reconfigurable platforms?
This theme investigates the design and adaptation of RTOS and scheduling techniques for embedded platforms using multi-core processors and Field Programmable Gate Arrays (FPGAs). The research addresses resource allocation, task migration between software and hardware, real-time guarantees on heterogeneous platforms, and predictable execution with limited overhead. It focuses on how to leverage multiprocessing and reconfiguration while still providing deterministic real-time behavior.
Key finding: Proposes a novel on-line admission control and task relocation scheme that enables tasks to migrate dynamically between software running on CPUs and hardware implemented on partially reconfigurable FPGAs in embedded systems.... Read more
Key finding: As above, the micro-kernel’s distributed asymmetric design enables scalable parallel execution and resource management on multi-core platforms, directly addressing challenges in embedded systems that require strict real-time... Read more
Key finding: Develops two lightweight and effective control-flow error (CFE) detection and recovery techniques that consider inter-thread dependencies in multi-threaded programs on multi-core processors. By inserting minimal additional... Read more
Key finding: Analyzes the Data Distribution Service (DDS) middleware standard focusing on real-time guarantees for distributed embedded systems involving multi-core and networked platforms. It integrates MARTE concepts for schedulability... Read more
Key finding: As above, its ability to generate worst-case scenarios for tasks scheduling is especially important in complex multi-core and distributed real-time embedded systems where concurrency and unpredictable interactions complicate... Read more
3. How can formal methods and model-driven engineering improve design, verification, and validation in real-time embedded system development?
This area targets the adoption of formal languages, model checking, model-driven engineering (MDE), and standardized profiles like UML MARTE to model, specify, and verify real-time embedded system behavior accurately. It deals with overcoming challenges such as state-space explosion, timing correctness, and integration of hardware/software co-design to ensure systems meet temporal and functional correctness before deployment, thus reducing testing costs and improving reliability.
Key finding: Presents the TAXYS tool that integrates statistical synchronous programming language ESTEREL with timed automata model checking (KRONOS) for design and validation of real-time embedded applications. The approach captures time... Read more
Key finding: Proposes an operational method for translating UML sequence diagrams annotated with MARTE to Time Petri Nets with Action Duration (DTPN) specifications, endowed with true concurrency semantics modeled as Durational Action... Read more
Key finding: Surveys model-driven engineering approaches using UML/MARTE profiles for high abstraction level design of adaptive real-time embedded systems. It highlights the lack of comprehensive MDE methods supporting dynamic adaptation... Read more
Key finding: Introduces an approach where real-time operating systems (RTOS) are automatically generated from high-level system descriptions (Codesign Finite State Machines). This bridges hardware/software co-design with RTOS generation,... Read more
Key finding: Though primarily educational, this work emphasizes integrated training based on theoretical and practical models such as model-driven engineering approaches and real-time system theories. It reflects the growing academic... Read more