In this project we will study the fundamental properties of advanced industrial hard real-time control-systems and the, for these systems highly suitable, analysis techniques called Response-Time Analysis for tasks with Offsets (RTA-O). Our goal is to understand, and ultimately limit the effects of, the sources of unsustainability in RTA-O. We will focus on the classes of preemtable fixed-priority scheduled uniprocessor systems and server-based scheduled multiprocessor systems.
Sustainability is a property of a hard real-time scheduling algorithm (or scheduling test) that is highly useful from an engineering perspective. Sustainability intuitively implies that the scheduling test should behave monotonically with respect to the parameters used in the test. Equally useful is to have scheduling tests that have low pessimism and allow high system utilization. However, in the real-time systems community it has been concluded that there is an inherent contradiction in achieving both sustainability and low pessimism.
This project will provide fundamental understanding of the concept of sustainability. A concept studied too little in literature. The inherent contradiction between sustainable methods and exact and high-precision methods needs to be explored and understood and resolved or mitigated. If the scientific community continues to fail to deliver analysis methods with good precision; there is very little hope that our methods will be adopted in practical use.
FastStereoNet: A Fast Neural Architecture Search for Improving the Inference of Disparity Estimation on Resource-Limited Platforms (Nov 2021) Mohammad Loni, Ali Zoljodi, Amin Majd , Byung Hoon Ahn , Masoud Daneshtalab, Mikael Sjödin, Hadi Esmaeilzadeh IEEE Transactions on Systems, Man, and Cybernetics: Systems (SMCS)
On Sustainability for Offset Based Response-Time Analysis (May 2021) Jukka Mäki-Turja, Kaj Hänninen, Mikael Sjödin European Conference on the Engineering of Computer-Based Systems (ECBS)
A software implemented comprehensive soft error detection method for embedded systems (Sep 2020) Seyyed Amir Asghari , Mohammadreza Binesh Marvasti , Masoud Daneshtalab Elsevier journal of Microprocessors and Microsystems (MICPRO)
DenseDisp: Resource-Aware Disparity Map Estimation by Compressing Siamese Neural Architecture (Jul 2020) Mohammad Loni, Ali Zoljodi, Amin Majd , Masoud Daneshtalab, Mikael Sjödin, Ben Juurlink , Reza Akbari IEEE WORLD CONGRESS ON COMPUTATIONAL INTELLIGENCE (WCCI) 2020 (IEEE WCCI)
DeepMaker: A Multi-Objective Optimization Framework for Deep Neural Networks in Embedded Systems (Jan 2020) Mohammad Loni, Sima Sinaei, Ali Zoljodi, Masoud Daneshtalab, Mikael Sjödin Elsevier journal of Microprocessors and Microsystems (MICPRO)