Filetype
Theory for Distributed Model Predictive Control (DMPC) is developed based on dual decomposition of the convex optimization problem that is solved in each time sample. The process to be controlled is an interconnection of several subsystems, where each subsystem corresponds to a node in a graph. We present a stopping criterion for the DMPC scheme that can be locally verified by each node and that g
We approach the simultaneous localization and mapping problem by using an ultrasound sensor and wheel encoders on a mobile robot. The measurements are modeled to yield a conditionally linear model for all the map states. Moreover, we implement a Rao-Blackwellized particle smoother (RBPS) for jointly estimating the position of the robot and the map. The method is applied and successfully verified b
A resource management framework for media applications modelled as dataflow streams is described. The framework is aimed at multi-core applications and the main resource managed is the CPU time. The management is implemented using reservation-based scheduling in combination with optimization and feedback. Experiences of applying the framework to media applications are reported. A simulation model
During the recent years, researchers in control engineering have proposed more resource-efficient control strategies than the traditional periodic control paradigm, resulting in two main control approaches: event-based and self-triggered control. Such nonperiodic, state-based control methods, although subject to many open research problems, can provide similar control performance as periodic contr
Recurrent neural networks (RNNs) are efficient for classification of sequential data such as speech and audio due to their high precision on tasks. However, power efficiency, the required memory capacity and bandwidth requirements make them less suitable for battery powered devices. In this work, we introduce FLoPAD-GRU: a system on a chip (SoC) for efficient processing of gated recurrent unit (GR
There is a large variability between individuals in the response to anesthetic drugs, that seriously limits the achievable performance of closed-loop controlled drug dosing. Full individualization of patient models based on early clinical response data has been suggested as a means to improve performance with maintained robustness (safety). We use estimation theoretic analysis and realization theo
Today, the majority of control applications in embedded systems, e.g., in the automotive domain, are implemented as software tasks on shared platforms. Ignoring implementation impacts during the design of embedded control systems results in complex timing behaviors that may lead to poor performance and, in the worst case, instability of control applications. This article presents a methodology for
The European Spallation Source will, once fully operational in 2025, be the world's brightest neutron source. The neutrons will be generated by bombarding a tungsten target with protons accelerated to 96\% the speed of light by electromagnetic fields confined in 155 radio-frequency cavities along the world's most powerful linear accelerator.This thesis has been motivated by the strict control spec
Having multiple control tasks concurrently running on a single computing platform increases the processor utilization but degrades the control performance due to delay and jitter. In scheduling and control codesign, the objective is to optimize the combined performance of all the controllers, subject to a schedulability constraint. The codesign procedure consists of selecting task parameters, e.g.
We give closed-form expressions for H-infinity optimal state feedback laws applicable to linear time-invariant discrete time systems with symmetric and Schur state matrix. This class includes networked systems with local dynamics in each node and control action along each edge. Furthermore, the structure of the controllers mimics that of the system, which makes them suitable for distributed contro
In this paper, we address distributed convergence to fair allocations of CPU resources for time-sensitive applications. We propose a novel resource management framework where a centralized objective for fair allocations is decomposed into a pair of performance-driven recursive processes for updating: (a) the allocation of computing bandwidth to the applications (resource adaptation), executed by t
Shift-map image processing is a new framework based on energy minimization over a large space of labels. The optimization utilizes -expansion moves and iterative refinement over a Gaussian pyramid. In this paper we extend the range of applications to image registration. To do this, new data and smoothness terms have to be constructed. We note a great improvement when we measure pixel similarities
In the study of distributed systems, it is of fundamental interest to understand how specifications on local components and interconnections influence global properties of the system. In this paper, we consider linear continuous time systems described by sparse matrices. Properties of interest are stability and quadratic performance specifications such as passivity and input-output gain. In partic
It has been shown that optimal controller synthesis for positive systems can be formulated as a linear program. Leveraging these results, we propose a scalable iterative algorithm for the systematic design of sparse, small gain feedback strategies that stabilize the evolutionary dynamics of a generic disease model. We achieve the desired feedback structure by augmenting the optimization problems w
For interconnected systems and systems of large size, aggregating information of subsystems studied individually is useful for addressing the overall stability. In the Lyapunov- based analysis, summation and maximization of separately constructed functions are two typical approaches in such a philosophy. This paper focuses on monotone systems which are common in control applications and elucidates
