Security assurance in a computer system can be viewed as distinguishing between self and non-self. Artificial Immune Systems (AIS) are a class of machine learning (ML) techniques inspired by the behavior of innate biological immune systems, which have evolved to accurately classify self-behavior from non-self-behavior. This work aims to leverage AIS-based ML techniques for identifying certain behavioral traits in high level hardware descriptions, including unsafe or undesirable behaviors, whether such behavior exists due to human error during development or due to intentional, malicious circuit modifications, known as hardware Trojans, without the need fora golden reference model. We explore the use of Negative Selection and Clonal Selection Algorithms, which have historically been applied to malware detection on software binaries, to detect potentially unsafe or malicious behavior in hardware. We present a software tool which analyzes Trojan-inserted benchmarks, extracts their control and data-flow graphs (CDFGs), and uses this to train an AIS behavior model, against which new hardware descriptions may be tested.

Étant donné G = (V, E) un graphe non orienté connexe et un entier positif β (n), où n est le nombrede sommets de G, le problème du séparateur (VSP) consiste à trouver une partition de V en troisclasses A, B et C de sorte qu'il n'y a pas d'arêtes entre A et B, max {| A |, | B |} est inférieur ou égal àβ (n) et | C | est minimum. Dans cette thèse, nous considérons une modélisation du problème sous laforme d'un programme linéaire en nombres entiers. Nous décrivons certaines inégalités valides et etdéveloppons des algorithmes basés sur un schéma de voisinage.Nous étudions également le problème du st-séparateur connexe. Soient s et t deux sommets de Vnon adjacents. Un st-séparateur connexe dans le graphe G est un sous-ensemble S de V \ {s, t} quiinduit un sous-graphe connexe et dont la suppression déconnecte s de t. Il s'agit de déterminer un stséparateur de cardinalité minimum. Nous proposons trois formulations pour ce problème et donnonsdes inégalités valides du polyèdre associé à ce problème. Nous présentons aussi une heuristiqueefficace pour résoudre ce problème. / Given G=(V,E) a connected undirected graph and a positive integer β(n), where n is number ofvertices, the vertex separator problem (VSP) is to find a partition of V into three classes A,B and Csuch that there is no edge between A and B, max{|A|,|B|}less than or equal β(n), and |C| isminimum. In this thesis, we consider aninteger programming formulation for this problem. Wedescribe some valid inequalties and using these results to develop algorithms based onneighborhood scheme.We also study st-connected vertex separator problem. Let s and tbe two disjoint vertices of V, notadjacent. A st-connected separator in the graph G is a subset S of V\{s,t} such that there are no morepaths between sand tin G[G\S] and the graph G[S] is connected . The st-connected vertex speratorproblem consists in finding such subset with minimum cardinality. We propose three formulationsfor this problem and give some valid inequalities for the polyhedron associated with this problem.We develop also an efficient heuristic to solve this problem.

The demand for collaborative, easy to use robots has increased during the last decades in hope of incorporating the use of robotics in smaller production scales, with easier and faster programming. Artificial intelligence (AI) and Machine learning (ML) are showing promising potential in robotics and this project has attempted to automatically solve a specific assembly task with Behavior trees (BTs). BTs can be used to elegantly divide a problem into different subtasks, while being modular and easy to modify. The main focus is put towards developing a Genetic algorithm (GA), that uses the fundamentals of biological evolution to produce BTs that solves the problem at hand. As a comparison to the GA result, a so-called Automated planner was developed to solve the problem and produce a benchmark BT. With a realistic physics simulation, this project automatically generated BTs that builds a tower of Duplo-like bricks and achieved successful results. The results produced by the GA showed a variety of possible solutions, a portion resembling the automated planner's results but also alternative, perhaps more elegant, solutions. As a conclusion, the approach used in this project shows promising signs and has many possible improvements for future research.

This thesis investigates the use of Maximum Likelihood Sequence Estimation (MLSE) in the presence of unknown channel parameters. MLSE is a fundamental problem that is closely related to many modern research areas like Space-Time Coding, Overloaded Array Processing and Multi-User Detection. Per-Survivor Processing (PSP) is a technique for approximating MLSE for unknown channels by embedding channel estimation into the structure of the Viterbi Algorithm (VA). In the case of successful acquisition, the convergence rate of PSP is comparable to that of the pilot-aided RLS algorithm. However, the performance of PSP degrades when certain sequences are transmitted. In this thesis, the blind acquisition characteristics of PSP are discussed. The problematic sequences for any joint ML data and channel estimator are discussed from an analytic perspective. Based on the theory of indistinguishable sequences, modifications to conventional PSP are suggested that improve its acquisition performance significantly. The effect of tree search and list-based algorithms on PSP is also discussed. Proposed improvement techniques are compared for different channels. For higher order channels, complexity issues dominate the choice of algorithms, so PSP with state reduction techniques is considered. Typical misacquisition conditions, transients, and initialization issues are reported. / Master of Science

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第18530号 / 情博第534号 / 新制||情||95(附属図書館) / 31416 / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 岡部 寿男, 教授 松山 隆司, 教授 阿久津 達也 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

No description available.

Multilevel inverters (MLI) have been widely applied in medium and high voltage applications for their advantages in high quality of output waveforms. Among various multilevel topologies, cascaded H-bridge (CHB) inverters have attracted more attentions for its modular structure, which simplifies the design and implementation. In addition, the modularity of CHB also expands diverse power ratings without many changes in the hardware setup. In a CHB inverter, the AC output voltage can be produced at different voltage levels depending on the number power cells that are cascaded at the output. To produce the AC output voltage, different modulation schemes and control algorithms have been studied and applied to the CHB inverter. Model predictive control (MPC) has been widely employed among all control algorithms in multilevel topologies due to their advantages such as good dynamic performance, multiple control targets, inclusion of nonlinearity, and flexibility to add more performance objectives. However, one disadvantage of the MPC is that the switching frequency is variable compared with other modulation schemes. Therefore, a new MPC method called modulated model predictive control (M2PC) has been researched to obtain a fixed switching frequency, which improves the harmonic spectrum of load currents and simplifies the filter design. In the modulated model predictive control, the mathematical model is obtained by electrical model of the system. It means that the operation of the M2PC algorithm relies on the accuracy of the given parameters and model. If there is an error in parameters and model, the performance of the control will be affected negatively. To solve this problem, modulated model-free predictive control (M2FPC) algorithm has been introduced. With this method, the mathematical model is established with measured values instead of given values and model. Reliability is one of the most important issues in the design of power converters. However, the failure of power switches will lead to the distortion of load currents and voltage waveforms. Also, the distortion in load currents and voltage waveforms causes power imbalance between faulty and healthy phases. To reduce the negative effects of IGBT failure in power converters, the faulty power cells should be found and isolated. Therefore, fault detection and localization algorithm (FDL) should be introduced to detect the fault in power converters and localize the faulty power switches. FDL algorithm based on the given M2PC scheme is proposed in this thesis for the CHB inverter to make the system more reliable. The FDL algorithm utilizes the phase voltages and load currents to detect the open fault in the CHB inverter and localize the single and multiple open switches by measuring the expected and actual phase voltages. With the faulty information, the faulty power cell can be isolated, and the fault-tolerant control can be applied to make the system work normally even though there is an open fault. In this thesis, without losing the generality, a seven-level CHB inverter is considered where there are three power cells in each phase. The M2PC algorithm was introduced to obtain the fixed switching frequency with the design of possible voltage vector set and carrier phase-shifting modulation. Based on the proposed M2PC algorithm, the FDL algorithm is designed to detect and localize the open switches to improve the system reliability. The theoretical analysis and simulation results validate the feasibility of the proposed M2PC algorithms and open fault diagnosis scheme. All possible open-circuit scenarios in power cells are discussed and the M2PC-based FDL algorithm has been verified. Experimental results verify the feasibility of the proposed M2PC. The experimental result of M2PC algorithm is presented to verify its operation. Also, diverse open scenarios can be diagnosed in the experiments. / Thesis / Master of Applied Science (MASc)

Thermal systems often feature composite regions that are mechanically mated. In general, there exists a significant temperature drop across the interface between such regions which may be composed of similar or different materials. The parameter characterizing this temperature drop is the thermal contact resistance, which is defined as the ratio of the temperature drop to the heat flux normal to the interface. The thermal contact resistance is due to roughness effects between mating surfaces which cause certain regions of the mating surfaces to loose contact thereby creating gaps. In these gap regions, the principal modes of heat transfer are conduction across the contacting regions of the interface, conduction or natural convection in the fluid filling the gap regions of the interface, and radiation across the gap surfaces. Moreover, the contact resistance is a function of contact pressure as this can significantly alter the topology of the contact region. The thermal contact resistance is a phenomenologically complex function and can significantly alter prediction of thermal models of complex multi-component structures. Accurate estimates of thermal contact resistances are important in engineering calculations and find application in thermal analysis ranging from relatively simple layered and composite materials to more complex biomaterials. There have been many studies devoted to the theoretical predictions of thermal contact resistance and although general theories have been somewhat successful in predicting thermal contact resistances, most reliable results have been obtained experimentally. This is due to the fact that the nature of thermal contact resistance is quite complex and depends on many parameters including types of mating materials, surface characteristics of the interfacial region such as roughness and hardness, and contact pressure distribution. In experiments, temperatures are measured at a certain number of locations, usually close to the contact surface, and these measurements are used as inputs to a parameter estimation procedure to arrive at the sought-after thermal contact resistance. Most studies seek a single value for the contact resistance, while the resistance may in fact also vary spatially. In this thesis, an inverse problem (IP) is formulated to estimate the spatial variation of the thermal contact resistance along an interface in a two-dimensional configuration. Temperatures measured at discrete locations using embedded sensors appropriately placed in proximity to the interface provide the additional information required to solve the inverse problem. A superposition method serves to determine sensitivity coefficients and provides guidance in the location of the measuring points. Temperature measurements are then used to define a regularized quadratic functional that is minimized to yield the contact resistance between the two mating surfaces. A boundary element method analysis (BEM) provides the temperature field under current estimates of the contact resistance in the solution of the inverse problem when the geometry of interest is not regular, while an analytical solution can be used for regular geometries. Minimization of the IP functional is carried out by the Levenberg-Marquadt method or by a Genetic Algorithm depending on the problem under consideration. The L-curve method of Hansen is used to choose the optimal regularization parameter. A series of numerical examples are provided to demonstrate and validate the approach.

No / This paper presents a two-stage hybrid optimization algorithm based on a modified genetic algorithm. In the first stage, a global search is carried out over the design search space using a modified GA. The proposed modifications on the basic GA includes dynamically changing the population size throughout the GA process and the use of different forms of the penalty function in constraint handling. In the second stage, a local search based on the genetic algorithm solution is executed using a discretized form of Hooke and Jeeves method. The hybrid algorithm and the modifications to the basic genetic algorithm are examined on the design optimization of reinforced concrete flat slab buildings. The objective function is the total cost of the structure including the cost of concrete, formwork, reinforcement and foundation excavation. The constraints are defined according to the British Standard BS8110 for reinforced concrete structures. Comparative studies are presented to study the effect of different parameters of handling genetic algorithm on the optimized flat slab building. It has been shown that the proposed hybrid algorithm can improve genetic algorithm solutions at the expense of more function evaluations.

k-means is a simple and flexible clustering algorithm that has remained in common use for 50+ years. In this thesis, we discuss the algorithm in general, its advantages, weaknesses and how its ability to locate clusters can be enhanced with a suitable initialization method. We formulate appropriate requirements for the (batched) UnifRandom, k-means++ and Kaufman initialization methods and compare their performance on real and generated data through simulations. We find that all three methods (followed by the k-means procedure) are able to accurately locate at least up to nine well-separated clusters, but the appropriately batched UnifRandom and the Kaufman methods are both significantly more computationally expensive than the k-means++ method already for K = 5 clusters in a dataset of N = 1000 points.