To be able to do the correct action when a fault is detected, the fault isolation part must be precise and run in real time during operation of the process. In many cases can it be difficult to decide exactly where the fault is localized. In those cases, the isolation algorithm must rank the faults according to their probability to be the cause to the behavior. The masters thesis project aims at probabilistic methods and algorithms for fault isolation in embedded systems. Different kind of Bayesian Networks have been compared in this report and the comparison has been done on a literature defined “benchmark system”. Those Bayesian network models which have been implemented for fault isolation are: 1. Manually (on the basis of physical representations) 2. Two-layer structure continuous signals discreet signals 3. Via temporal causal graph (dynamical network) The algorithms should be compared in the following areas: computational complexity, isolation performance and degree of difficulty to construct the network on the basis of data. The evaluated algorithms showed good results. Even though the system data which have been used in the Bayesian Networks are not very accurate in the first place, it manage to give a fairly precise isolation of the faults. The continuous Bayesian Network manage to show a good isolation performance for different type of faults and the Dynamic Bayesian Network found most of the faults even for a rather complex network. / Detta examensarbete handlar om sannolikhetsbaserade metoder för felisolering. När ett fel uppstår ombord på en Scania lastbil kan man upptäcka det. I bästa fall kan en viss komponent pekas ut som orsak, men ofta kommer man att ha ett antal komponenter som kan vara orsaken. I många fall är det dock svårt att hitta var exakta felet finns. För att hantera dessa situationer vill man använda metoder för att beräkna sannolikheten att olika komponenter är trasiga. För att beräkna sannolikheten kan man använda en probabilistisk model, dvs. Bayesianska nätverk. I detta arbete har olika metoder för att skapa Bayesianska nätverk jämförts. Jämförelsen görs på ett litteratur väl definierat benchmark problem: diagnosar en två-tank system. De typer av Bayesiansk nätverks modeller som har implementerats för felisolering är: 1. Manuellt (ut ifrån fysikalisk modell) 2. Två-lagers struktur kontinuerliga signaler diskreta signaler 3. Via Bindningsgrafer (dynamiskt nätverk) Problemen som undersöktes var bland annat svårighet att bygga nätverket utifrån data, beräkningskomplexitet samt isolerings prestanda. En jämförelse mellan de Bayesianska metoderna för felisolering och samt dem befintliga standardmetoder har även gjorts. De undersökta algoritmerna visade goda resultat. Trots bristen på data, visade algoritmerna lovande resultat. Det Två-lagers Bayesianska nätverket visade en bra isoleringsprestanda på olika komponent fel och det Dynamiska Bayesianska nätverket upptäckte de flesta fel trots att det var ett ganska complext nätverk.

Model predictive control techniques are widely used in the process industry. They are considered methods that give good performance and are able to operate during long periods without almost any intervention. Model predictive control is also the only technique that is able to consider model restrictions. Almost all industrial processes have nonlinear dynamics, however most MPC applications are based on linear models. Linear models do not always give a sufficiently adequate representation of the system and therefore nonlinear model predictive control techniques have to be considered. Working with nonlinear models give rise to a wide range of difficulties such as, non convex optimization problems, slow processes and a different approach to guarantee stability . This project deals with nonlinear model predictive control and is written at the University of Seville at the department of Systems and Automatic control and at the department of Automatic Control at KTH. The first objective is to control the nonlinear Four Tank Process using nonlinear model predictive control. Objective number two is to investigate if and how the computational time and complexity can be reduced. Simulations show that a nonlinear model predictive control algorithm is developed with satisfactory results. The algorithm is fast enough and all restrictions are respected for initial state values inside of the terminal set as well as for initial state values outside of the terminal set. Feasibility and stability is ensured for both short as well as for longer prediction horizon, guaranteeing that the output reaches the reference. Hence the choice of a short respectively long prediction horizon is a trade off between shorter computational time versus better precision. Regarding the reduction of the computational time, penalty functions have been implemented in the optimization problem converting it to an unconstrained optimization problem including a PHASE-I problem. Results show that this implementation give approximately the same computational time as for the constrained optimization problem. Precision is good for implementations with penalty functions both for long and short prediction horizons and initial state values inside and outside of the terminal set.

Due to high competition on the market the industry aims at increasing the efficiency in their assembly production line. The production cycle time is determined by many factors such as operation time at each production step, the reliability of the tool and so forth. Many of the assembly applications are performed with hand held power tools which mean that a person, called operator, is holding the tool. Hence, physical stress is acting on the operator during tightening procedures and is therefore an important issue. As a result, development of better power tools is always needed. The goal with this thesis work is to build a simulation model of a hand held assembly tool. The simulation is to be use for development and evaluation of the present computer controlled tightening strategies. The model includes control unit, tool, joint and operator. Based on the model the aim is also the exploit the possibilities to improve the tightening algorithms in an automatic control fashion. The control structure investigated is MPC, (Model Predictive Control). The result is that the simulation model with good correlation describes the system and can therefore be used as an aid in the development of tool components and software. The proposed tightening control strategy, MPC, show excellent performance regarding operation tine and the reduction of operator energy input. The drawback is that the robustness of the controller is somewhat limited due to fast process time. Also the real-time implementation  as to be further developed.

The project aims to improve positive torque transient response through more advanced wastegate controllers than what are used today. All controllers are developed for a standard General Motors turbocharged engine. In many turbocharged SI engines, a wastegate is used for preventing the turbine to overrun and to decrease the pumping loss. Today, the wastegate is controlled by a PI controller, which tries to fulfill a compromise between fuel consumption and torque response by regulating the wastegate position. A nonlinear Mean Value Engine Model (MVEM) of this engine, with 13 states and linearized in 45 different working points, is used. The original model, implemented in Matlab/Simulink, has been enriched with new features, like lambda and spark advance efficiencies and the related exhaust temperature correction. The project aims to do a theoretical analysis to find the optimal control of wastegate position, investigating also spark retard and fuel enrichment during a positive torque transient. First a solution for achieving optimal wastegate control is designed, based on Linear Quadratic (LQ) approach. Since the optimal control strategy is expected to vary quite much for different working points, a gain scheduling architecture has been investigated. An independent lambda controller has been developed, in order to maximize the lambda efficiency and quicken the torque response during transient. Since the system operates near a constraint boundary, another solution based on Model Predictive Control (MPC) of the wastegate has been investigated. The MPC design has been extended also to a MIMO formulation, adding the throttle and the air to fuel ratio as control inputs, and the trade off between fast torque response and fuel economy is analyzed. A complete realtime MPC implementation, with the capability for automatic code generation in the dSpace microAutobox environment, requires the model, now with 13 states, to be reduced to a minimum state space order. The extent of model reduction that is required and the possible performance deterioration have been investigated.

This report describes an exemple of management systems based on Programmable Logic Controllers. Such systems are designed in order to help monitoring and controlling of a physical system. The system which is controlled in this project is the electrical network of two buildings used by the french parliament. This network consists of several electrical boards where each board is made up of several circuit breakers that protect the electrical equipments and other boards. The goal of the management system is to monitor the network, i.e acquire data about the state of the network, and to control the network, i.e control the position of the breakers (open/closed) according to power supplies availability and network load. The management system is based on a multiple levels structure. - Level 0 : is the operating part of the network which is made up of sensors and actuators. - Level 1 : is the part made up of the Programmable Logic Controllers (PLC). The sensors and actuators of the level 0 are directly connected to the PLC. - Level 2 : This level enables to acquire the information on a computer and to monitor/control it through an ergonomic graphical interface. Thanks to that level, the operator can keep an eye on the network remotely and control it. This project consists in programming the Programmable Logic Controllers, and the Human Machine Interface, according to the specifications of the customer. Only a part of the system has so far been developped since the total duration of the project is about 2 years. The main functions of the PLC program has been programmed, and tested successfully but not in real conditions since the buildings were still under construction. Simulation tools were used in order to test Modbus communication between the levels 0 and 1. The function that controls the breakers according to power supplies availability has been tested as well, by forcing variables in the PLC (to simulate a voltage loss for example).

The recent development of control applications overWireless Sensor Networks (WSNs) imposes new approaches to the protocol design. These networks are characterized by the scarcity of energy supply and processing capabilities. Furthermore, existing protocol solutions are often based on the traditional OSI model, where communication layers are not optimized to support efficiently the reliability and latency requirements imposed by control applications. The critical aspects of wireless transmission have lead to a lack of protocols that are able to guarantee latency and quality of service under unreliable channel conditions. In this thesis, we design and implement a cross-layer protocol for WSNs in industrial automation, the Extended Randomized Protocol, which considers jointly physical layer aspects (as power control and duty cycling strategies), randomized MAC and routing. The protocol can be considered and extension of an already existing Randomized Protocol, and it is designed with the objective to maximize the network lifetime under the constraints of error rate and end-to-end delay in the packet delivery. As a relevant part of our activity, we have provided a complete test bed implementation of the protocol building a WSN with TinyOS and a large number of Moteiv’s Tmote Sky wireless sensors. An experimental campaign has been conducted in order to test the validity of the protocol solution we propose. Experimental results show that the protocol achieves the required successful packet reception rate and the latency constraints while minimizing the energy consumption. Despite the fact that improving solutions are necessary to take into account the problem of duplicated packets, our protocol solution seems to be a good candidate for WSN in industrial automation.

An eye tracker makes it possible to record the gaze point of a person looking at for example a computer monitor. Modern techniques are very flexible and allow the user to behave naturally without the need of cumbersome equipment such as special contact lenses or electrical probes. This is valuable in psychological research, marketing research and Human Computer Interaction. Eye trackers also give people who are severely paralyzed and unable to type and speak means to communicate using their eyes. Measurement noise makes the use of digital filters necessary. An example is an eye-controlled cursor for a desktop environment such as Windows. The cursor has to be stable enough to allow the user to select folders, icons or other items of interest. While this type of application requires a fast real-time filter, others are less sensitive to processing time but demand an even higher level of accuracy. This work explores three areas of eye tracking filtration and aims at enhancing the performance of the filters used in the eye tracking systems built by Tobii Technology, Sweden. First, a post-processing algorithm to find fixations in raw gaze data is detailed. Second, modifications to an existing reading detection algorithm are described to make it more robust to natural irregularities in reading patterns. Third, a real-time filter for an eye-controlled cursor to be used in a desktop environment is designed using a low-pass filter in parallel with a change detector. The fixation filter produced fewer false fixations and was also able to detect fixations lying spatially closer together than the previously used filter. The reading detection algorithm was shown to be robust to natural irregularities in reading such as revisits to previously read text or skipped paragraphs. The eye-cursor filter proved to respond quicker than the previously used moving average filter while maintaining a high level of noise attenuation.

In this thesis we present ReDaR, a software and hardware environment for reliable data retrieval from multiple remotely located wireless sensor networks (WSNs). The software is programmed in Java and NesC over TinyOS, and allows for direct interaction with individual sensors using a graphical user interface running on a remote client. The hardware platform includes a WSN deployed with Telos wireless sensors. ReDaR implements a control algorithm over a real WSN, which allows to keep under control the overhead communication bandwidth between the sensors and the remote client. Specifically, we implement a control algorithm based on the extremum seeking control theory, which enables an adaptive forward error correction mechanism to ensure reliable real-time data retrieval. The amount of redundancy is controlled via a switching controller that uses a feedforward and a feedback mechanism. The switching controller ensures an optimal trade-off between maximizing the bandwidth and maximizing the quality of service. Experiments performed in our environment confirm theoretical results about performance of the controller, i.e. the feedback mechanism provides a robust controller that finds the optimum without model reliance, while the feed forward one improves the transient behavior.

Gene expression often reacts to intra- and extra-cellular signals in an ”on-off” switchlike manner. These switches, which display hysteresis, correspond to multiple steady-state solutions of the gene regulatory network and are created through saddle-node bifurcations at which a single real eigenvalue of the local Jacobian crosses the imaginary axis for changes in some system parameter. From control theory, it is well known that such an instability can be induced by a single positive feedback loop, possibly involving only a single gene and its corresponding mRNA. However, experimental results have suggested that bistable switches often involve several interlinked feedback loops. Since such a design is relatively costly, it is reasonable to assume that the excessive loops serve a purpose. A standing hypothesis is that the existence of multiple loops improves both the performance and the robustness of the switches. With performance is here understood the speed of switching, while robustness usually refers to the sensitivity of the switch with respect to noisy signals. The aim of this project is to consider these claims, and in particular the one concerning robustness, from a control theoretic perspective. First, an analysis of some specific biological examples will be performed, considering robustness both with respect to noise and with respect to model (systems) perturbations. In the second part, the aim is to propose a multiple loop design that provides optimal robustness.

IARES is a highly flexible planetary exploration demonstration rover developed by CNES (the French National Center for Space Studies) mainly for autonomous navigation and locomotion studies. It has 19 degrees of freedom, including six active, steerable wheels. The rover uses a software for autonomous navigation, including stereo camera perception, path planning and motion control. It is complemented by a visual simulator that can substitute the rover for practical purposes. The goal of this MSc thesis, carried out during the second semester 2006 at CNES in Toulouse, has been to make out the most of the localization capabilities of this rover using a recently implemented method : odometry. A previous study had been carried out at ONERA in Toulouse and the main goal of this thesis was to implement this new method into the environment used for the CNES rover and to test the performances of this method thanks to the simulator. All this work might be even tested on board at the very end of the internship. Given the hardware platform and the software environment, this new localization method had primarily to be studied from a theoretical point of view before being integrated into the CNES environment. The study was conducted on a Linux platform and code has been developed in C for the simulator whereas Scilab has been used for the validation tests.