京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21918号 / 情博第701号 / 新制||情||120(附属図書館) / 京都大学大学院情報学研究科システム科学専攻 / (主査)准教授 西原 修, 教授 大塚 敏之, 教授 加納 学 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

The development of connected vehicle (CV) technologies enables advanced management of individual vehicles and traffic signals to improve urban mobility and energy efficiency. In this thesis, a cooperative vehicle-signal control system will be developed to integrate an Eco-driving system and a proactive signal control system under a mixed connected environment with both connected vehicles (CVs) and human-driven vehicles (HDVs). The system utilizes CVs to conduct an accurate prediction of queue length and delay at different approaches of intersections. Then, a queue-based optimal control strategy is established to minimize the fuel usage of individual CVs and the travel time delay of entire intersections. The system applies the model predictive control to search for the optimal signal timing plan for each intersection and the most-fuel efficient speed profiles for each CV to gain the global optimum of the intersection. In this thesis, a simulation platform is designed to verify the effectiveness of the proposed system under different traffic scenarios. The comparison with the eco-driving only and signal control only algorithms verifies that the cooperative system has a much more extensive reduction range of the trip delay in the case of medium and high saturation. At low saturation, the effect of the system is not much different from that of the eco-driving algorithm, but it is still better than the signal control. Results show that the benefits of CVs are significant at all different market penetration rates of CVs. It also demonstrates the drawback of the system at high congestion levels. / Thesis / Master of Applied Science (MASc)

Public spaces, especially educational institutions like universities, encounter challenges with their room booking and access control systems. These challenges commonly manifest as overlapping bookings and unauthorized entry. The latter issue, unauthorized access, specifically stems from inadequate integration between the respective systems. This bachelor thesis introduces a proof-of-concept for a cohesive room booking and access control system to address these issues. The proposed solution encompasses two mobile applications, one as the room reservation platform and the other as the access control mechanism. By integrating the management of bookings and access control, this proof-of-concept aims to overcome the prevalent shortcomings in existing systems. Halmstad University's IT department was consulted during the requirement definition phase to ensure a comprehensive understanding of the common problems, their underlying causes, and possible solutions. The proposed system utilizes common technologies such as NodeJS, Android Studio, and PostgreSQL. Additionally, Mobile BankID is integrated as a unique feature for secure user authentication, providing a trusted and widely-accepted method to verify users' identities. The final results were tested in a simulated environment and indicate that the developed system satisfies the initial requirements, addressing the problems of double bookings and unauthorized access identified during the consultation with the IT department.

Industrial Control Systems (ICS) play a critical part in our world’s economy, supply chain and critical infrastructure. Securing the various types of ICS is of the utmost importance and has been a focus of much research for the last several years. At the heart of many defense in depth strategies is the signature based intrusion detection system (IDS). The signatures that define an IDS determine the effectiveness of the system. Existing methods for IDS signature creation do not leverage the information contained within the PLC ladder logic file. The ladder logic file is a rich source of information about the PLC control system. This thesis describes a method for parsing PLC ladder logic to extract address register information, data types and usage that can be used to better define the normal operation of the control system which will allow for rules to be created to detect abnormal activity.

Den agila arbetsmetoden har vuxit fram i takt med en volatil marknad som ställer krav på flexibilitet och innovation. Arbetsmetoden utmärker sig som en teamledningsmetod med ett dynamiskt och autonomt arbetssätt. Det råder ofullständig förståelse för vilka utmaningar arbetssättet medför. Syftet med denna studie är att bidra till en djupare förståelse för fenomenet och kunna bidra med insikter för hur organisationens val av styrmekanismer influerar medarbetare i det agila arbetssättet. Totalt genomfördes sju semistrukturerade intervjuer som analyserades med en tematisk analys vilket genererade tre teman: agilt mindset, teamwork och kultur som meningsskapande. Studien visar att de agila arbetsmetoderna och modellerna omfattar en bredare komplexitet än många tidigare forskare framställt. Företagskulturen visar kunna gagna arbetsmetoden när den upplevs meningsfull, transparent, inkluderande och stimulerar engagemang. En internaliserad företagskultur kan frigöra potential att stimulera arbetsmetodens fördelar. Vid implementering av en agil arbetsmetod bör organisationer beakta att det kan vara en tidskrävande process och följer av andra utmaningar än vid traditionella arbetsmetoder.

The Left Ventricular Assist Device (LVAD) is a rotary mechanical pump that is implanted in patients with congestive heart failure to help the left ventricle in pumping blood in the circulatory system. The rotary type pumps are controlled by varying the pump motor current to adjust the amount of blood flowing through the LVAD. One important challenge in using such a device is the desire to provide the patient with as close to a normal lifestyle as possible until a donor heart becomes available. The development of an appropriate feedback controller that is capable of automatically adjusting the pump current is therefore a crucial step in meeting this challenge. In addition to being able to adapt to changes in the patient's daily activities, the controller must be able to prevent the occurrence of excessive pumping of blood from the left ventricle (a phenomenon known as ventricular suction) that may cause collapse of the left ventricle and damage to the heart muscle and tissues. In this dissertation, we present a new suction detection system that can precisely classify pump flow patterns, based on a Lagrangian Support Vector Machine (LSVM) model that combines six suction indices extracted from the pump flow signal to make a decision about whether the pump is not in suction, approaching suction, or in suction. The proposed method has been tested using in vivo experimental data based on two different LVAD pumps. The results show that the system can produce superior performance in terms of classification accuracy, stability, learning speed, iv and good robustness compared to three other existing suction detection methods and the original SVM-based algorithm. The ability of the proposed algorithm to detect suction provides a reliable platform for the development of a feedback control system to control the current of the pump (input variable) while at the same time ensuring that suction is avoided. Based on the proposed suction detector, a new control system for the rotary LVAD was developed to automatically regulate the pump current of the device to avoid ventricular suction. The control system consists of an LSVM suction detector and a feedback controller. The LSVM suction detector is activated first so as to correctly classify the pump status as No Suction (NS) or Suction (S). When the detection is “No Suction”, the feedback controller is activated so as to automatically adjust the pump current in order that the blood flow requirements of the patient’s body at different physiological states are met according to the patient’s activity level. When the detection is “Suction”, the pump current is immediately decreased in order to drive the pump back to a normal No Suction operating condition. The performance of the control system was tested in simulations over a wide range of physiological conditions.

This work presents the Czochralski growth dynamics as time-varying delay based model, applied to the growth of La3Ga5.5Ta0.5O14 (LGT) piezoelectric crystals. The growth of high-quality large-diameter oxides by Czochralski technique requires the theoretical understanding and optimization of all relevant process parameters, growth conditions, and melts chemistry. Presently, proportional-integral- derivative (PID) type controllers are widely accepted for constant-diameter crystal growth by Czochralski. Such control systems, however, do not account for aspects such as the transportation delay of the heat from crucible wall to the crystal solidification front, heat radiated from the crucible wall above the melt surface, and varying melt level. During crystal growth, these time delays play a dominant role, and pose a significant challenge to the control design. In this study, a time varying linear delay model was applied to the identification of nonlinearities of the growth dynamics. Initial results reveled the benefits of this model with actual growth results. These results were used to develop a long-range model predictive control system design. Two different control techniques using long range prediction are studied for the comparative study. Development and testing of the new control system on real time growth system are discussed in detail. The results are promising and suggest future work in this direction. Other discussion about the problems during the crystal growth, optimization of crystal growth parameters are also studied along with the control system design.

This thesis presents the integration of an acoustic tracking system within an autonomous underwater AUV (AUV) to enable real-time tracking of sharks tagged with artificial acoustic sources. The tracking system consists of two hydrophones and a receiver unit that outputs a measurement of the relative angle to the tagged shark. Since only two hydrophones are used, the sign of the relative angle measurement is unknown. To overcome this ambiguity, a particle filter algorithm was developed to estimate the position of the acoustic source. When combined with an active control system that drives vehicle to obtain different orientations with respect to the acoustic source, real-time autonomous localization, tracking, and following of a tagged shark is shown to be possible. Four types of ocean experiments were used to validate the system including: 1) AUV tracking of a stationary tag, 2) AUV tracking of a tagged kayak, 3) AUV tracking of a tagged AUV, and 4) AUV tracking of a tagged shark. These experiments were analyzed with respect to the localization error, associated error variance, and distance between the AUV and the tag. The final shark tracking experiments took place in SeaPlane Lagoon, Los Angeles, CA, where the AUV was able to autonomously track and follow a tagged Leopard Shark for several hours.

Critical Infrastructures (CIs) such as water treatment plants, power grids and telecommunication networks are critical to the daily activities and well-being of our society. Disruption of such CIs would have catastrophic consequences for public safety and the national economy. Hence, these infrastructures have become major targets in the upsurge of cyberattacks. Defending against such attacks often depends on an arsenal of cyber-defence tools, including Machine Learning (ML)-based Anomaly Detection Systems (ADSs). These detection systems use ML models to learn the profile of the normal behaviour of a CI and classify deviations that go well beyond the normality profile as anomalies. However, ML methods are vulnerable to both adversarial and non-adversarial input perturbations. Adversarial perturbations are imperceptible noises added to the input data by an attacker to evade the classification mechanism. Non-adversarial perturbations can be a normal behaviour evolution as a result of changes in usage patterns or other characteristics and noisy data from normally degrading devices, generating a high rate of false positives. We first study the problem of ML-based ADSs being vulnerable to non-adversarial perturbations, which causes a high rate of false alarms. To address this problem, we propose an ADS called DAICS, based on a wide and deep learning model that is both adaptive to evolving normality and robust to noisy data normally emerging from the system. DAICS adapts the pre-trained model to new normality with a small number of data samples and a few gradient updates based on feedback from the operator on false alarms. The DAICS was evaluated on two datasets collected from real-world Industrial Control System (ICS) testbeds. The results show that the adaptation process is fast and that DAICS has an improved robustness compared to state-of-the-art approaches. We further investigated the problem of false-positive alarms in the ADSs. To address this problem, an extension of DAICS, called the SiFA framework, is proposed. The SiFA collects a buffer of historical false alarms and suppresses every new alarm that is similar to these false alarms. The proposed framework is evaluated using a dataset collected from a real-world ICS testbed. The evaluation results show that the SiFA can decrease the false alarm rate of DAICS by more than 80%. We also investigate the problem of ML-based network ADSs that are vulnerable to adversarial perturbations. In the case of network ADSs, attackers may use their knowledge of anomaly detection logic to generate malicious traffic that remains undetected. One way to solve this issue is to adopt adversarial training in which the training set is augmented with adversarially perturbed samples. This thesis presents an adversarial training approach called GADoT that leverages a Generative Adversarial Network (GAN) to generate adversarial samples for training. GADoT is validated in the scenario of an ADS detecting Distributed Denial of Service (DDoS) attacks, which have been witnessing an increase in volume and complexity. For a practical evaluation, the DDoS network traffic was perturbed to generate two datasets while fully preserving the semantics of the attack. The results show that adversaries can exploit their domain expertise to craft adversarial attacks without requiring knowledge of the underlying detection model. We then demonstrate that adversarial training using GADoT renders ML models more robust to adversarial perturbations. However, the evaluation of adversarial robustness is often susceptible to errors, leading to robustness overestimation. We investigate the problem of robustness overestimation in network ADSs and propose an adversarial attack called UPAS to evaluate the robustness of such ADSs. The UPAS attack perturbs the inter-arrival time between packets by injecting a random time delay before packets from the attacker. The attack is validated by perturbing malicious network traffic in a multi-attack dataset and used to evaluate the robustness of two robust ADSs, which are based on a denoising autoencoder and an adversarially trained ML model. The results demonstrate that the robustness of both ADSs is overestimated and that a standardised evaluation of robustness is needed.

This thesis investigates the development of an animatronic robot with the objective of showing how modern animatronic models created as special effects have roots in models created during the scientific and mechanical revolution of the 17th and 18th centuries. It is noted that animatronic models that are available today have not been described in any great detail and most are covered by industrial secrecy. This project utilises technologies developed during the latter part of the 20th century and into the beginning of the 21st century to create the design of the animatronic robot. The objective of the project is to bring effective designs for animatronic robots into the public domain. The project will investigate a large variety of different mechanisms and apply them to various functioning parts of the model, with the design and method of each of these functions discussed. From this, one main part of the project, the jaw, will receive the focus of construction. Once the construction is complete this will be evaluated against what improvements and changes could be made for future iterations, with a revised design produced based on what has been learned.