Author: FINOGINA, TAMARA
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN APPLIED MATHEMATICS
Department: Department of Mathematics (MAT)
Mode: Normal
Deposit date: 26/07/2023
Reading date: 02/10/2023
Reading time: 11:00
Reading place: Sala d'Actes de la Facultat de Matemàtiques i Estadística, Edifici U, Campus Sud
Thesis director: HERRANZ SOTOCA, JAVIER
Committee:
     PRESIDENT: RONNE, PETER BROWNE
     SECRETARI: VILLAR SANTOS, JORGE LUIS
     VOCAL: HAENNI, ROLF
Thesis abstract: One of the most common fears regarding electronic voting is that a voting device will disregard the voter's intent and cast a different vote instead. An undetectable attack like that on a large scale will allow the adversary to control the election result completely. Therefore, the cast-as-intended verification, which ensures that the ballot contains the voter's choice and not something else, is crucial. Another common fear when introducing electronic voting is coercion, which captures a variety of ways the coercer can use to prevent voters from expressing their will. Hence, coercion resistance is a valuable property of electronic voting as well. One particularly challenging task is to find a trade-off between ensuring a voter cannot be coerced and, at the same time, preventing a malicious voting device from cheating. This thesis explores this trade-off to find how we can provide coercion-resistant cast-as-intended verification. The contributions can be roughly divided into three parts: (1) study in the standard settings, (2) exploration of post-quantum cryptography, and (3) practical constructions and search for the limitations of both properties. In the first part, we give an extensive overview of the current state of the art in electronic voting literature regarding those properties. Then, we put forward two formal definitions for achieving coercion-resistant cast-as-intended verification in settings without pre-exchanged data. After that, we present two practical constructions and prove their security under the proposed definitions. We also show the efficiency of our proposals by providing proof of the concept implementations. In the second part, we switch to post-quantum settings and identify the usability issues rooted in the lattice-based math affecting both proposed solutions. To address those issues, we present a generic transformation that departs from an interactive zero-knowledge system (that might require multiple re-runs to complete the protocol) and obtains a 3-move zero-knowledge system (without re-runs). The transformation combines the well-known Fiat-Shamir technique with several initially exchanged messages. The resulting 3-move system enjoys honest-verifier zero-knowledge and can be easily turned into a fully deniable proof using standard methods. In the final part, we focus on the practical aspects of the coercion-resistant cast-as-intended verification. First, we present the case of a computationally limited voter, which we consider the most realistic. We show that even a computationally limited voter can enjoy coercion-resistant cast-as-intended verification, but a help of a simple aid device for nonce generation is required. Also, we demonstrate that our generic definition easily adapts to the constraints of the limited voter. After that, we present ongoing work that focuses on the cases of extreme coercion based on new and unexplored mechanisms such as delay encryption and blockchain. We show an advanced coercive attack on our first construction and describe an improvement to the second solution that reduces the number of interactions to an optimal three rounds. To summarize, we start by studying coercion-resistant cast-as-intended verification in standard settings, which results in formal definitions and two practical solutions. Then we move into the post-quantum world, where we learn that an extra step is needed to preserve the usability of our previously proposed constructions, which results in the generic transformation to avoid protocol re-runs. After that, we concentrate on a computationally limited voter, which leads to another simple solution and shows the adaptability of our original definitions. Finally, we explore the extreme coercion threats, which result in a new coercion attack on the first construction and upgrade of the second solution.

Author: PALOMINO GAYETE, ARTURO
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN STATISTICS AND OPERATIONS RESEARCH
Department: Department of Statistics and Operations Research (EIO)
Mode: Normal
Deposit date: 23/05/2023
Reading date: 03/10/2023
Reading time: 11:30
Reading place: Sala de Teleensenyament. Edifici B3 - Campus Nord UPC, Carrer de Jordi Girona, 1-3, 08034 Barcelona.
Thesis director: GIBERT OLIVERAS, CARINA
Committee:
     PRESIDENT: SOLE CASALS, JORDI
     SECRETARI: MONTERO MERCADÉ, LIDIA
     VOCAL: HERRERA FERNANDEZ, MANUEL
Thesis abstract: The Internet has changed the way companies sell and users shop. Nowadays, having an online presence is an opportunity, while not doing so is a competitive disadvantage, and companies strive to appear as the first option when users are looking for a product and chained advertising is one of the most used tools currently used to achieve it.The key in these chained advertising campaigns is to correctly choose in which sequence of web domains the pieces of the campaign will be placed for it to be successful, and this is an open problem today, of a highly combinatorial nature, with a very expensive solution computationally initially. This thesis responds to the problem of how to ensure that users are really impacted by a chained advertising campaign by seeing all the banners that compose it in the right order, thus increasing the chances of them making an online purchase. The thesis proposes a method to discover the best sequences of websites for the purchase of optimal advertising spaces to place banners of chained advertising campaigns that lead the consumer towards a potential purchase in a less intrusive way.The proposed model has been constructed from user-centric clickstream data in an original and new way and it has been possible to formulate a data-driven method that finds these sequences in almost constant time. The proposed method, Quick Sequence Detector (QSD) is based on a strategic combination of mathematical formulas that allow the sequences to be transformed into numerical codes on which to operate efficiently, and very specific data engineering operations, which are part of almost all primitive sets of standard programming languages, and of very low algorithmic complexity. These two formal resources combine to articulate a very efficient computational time calculation of the joint probability function of the space produced by web domains, with a power equal to the length of the desired sequences, which is directly linked to the nature of the chained advertising campaign to be launched. The QSD proposal so significantly reduces runtimes in the calculation of this joint probability function that it goes from tens of years of computation to a few seconds in real applications of magnitude, for this reason the method is the subject of a patent currently being processed. Above the conceptualization of the solution, an additional improvement has been made that uses parallelization and programming of graphics processors to further accelerate data processing and obtain even more improved results. It should be said that the proposed solution is not limited to the field of digital marketing but presents a general formalization that allows the method to be transferred to other areas of application. In a second part of the thesis, the method extends to the introduction of conditioning factors on the sociodemographic characteristics of the user so that campaigns can be oriented to the appropriate target populations, and this extension of the method, which we call COND-QSD, is accompanied by an additional proposal on how to make the necessary inference to identify which pages are really significant and specific to the profile with respect to the general population.On the other hand, a software tool with a user interface has also been developed that allows a pilot test with a sample of real Internet users, collecting user-centric data, and comparing the results with the results of the algorithms to confirm the hypotheses. While the user is browsing the internet, they are also impacted by a chained advertising campaign designed for the occasion and in this way measure whether users of the target profile follow the expected itinerary, and how users of the target profile who have been impacted by the campaign make the purchase.

Author: ANGELATS COMPANY, EDUARD
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN AEROSPACE SCIENCE AND TECHNOLOGY
Department: Department of Physics (FIS)
Mode: Normal
Deposit date: 18/07/2023
Reading date: 04/10/2023
Reading time: 11:00
Reading place: CTTC, parc Mediterrani de la Tecnologia ¿ Edifici B6 Av. Carl Friedrich Gauss 7 08860 ¿ Castelldefels Barcelona, Sala de la defensa: Auditori Edifici B6
Thesis director: COLOMINA FOSCH, ISMAEL
Committee:
     PRESIDENT: CASAS PIEDRAFITA, JAIME OSCAR
     SECRETARI: DELGADO GARCIA, JORGE
     VOCAL NO PRESENCIAL: MARTÍNEZ SÁNCHEZ, JOAQUIN
Thesis abstract: Hybrid or multi-sensor-based positioning has been a research topic actively investigated in the last decade. In this context, the possibility of using information extracted from imaging sensors, for positioning, is very appealing to mitigate the problems that GNSS or INS/GNSS-based trajectories have in terms of robustness and accuracy. On the other hand, different processing workflows, sensor positioning quality or system calibration errors, may also produce even in GNSS-friendly conditions, that multiple geospatial datasets are not properly co-registered. This thesis proposes the use of non-semantic information, this is, the use of a set of geometric entities or features, to improve the trajectory estimation in a multi-sensor-based approach. This thesis covers the mathematical modelling of non-semantic information, implements several hybrid-based trajectory estimation approaches that use this kind of information with the appropriate modelling, and also explores the use of non-semantic features to model the trajectory error modelling.The implementation of combined models allowing to use of observations from camera or LiDAR sensors is the first contribution of this thesis. The proposed models have enabled improved trajectory determination in both urban post-processing and airborne environments with good accuracy (cm level). The implemented INS/GNSS trajectory error models are relatively simple but proved to be efficient. The combined models have been tested, in post-processing, using a bundle adjustment approach, with real data from metric cameras and aerial laser mapping systems as well as in Terrestrial Mobile Mapping systems (TMM). The second contribution of this thesis is the characterization of trajectory errors that TMM may have in GNSS urban scenarios. The non-semantic information extracted from the images has allowed, using an integrated sensor orientation approach, to model these errors in an urban environment. This modelling opens the door to the development of new, more advanced trajectory error models that go beyond the deterministic models currently used. The determination of trajectories in real time, in GNSS unfriendly environments, is also explored in this thesis using non-semantic features. An approach has been implemented based on a tightly coupling sequential nonlinear least squares using GNSS positions, image coordinates and raw inertial measurements. The proposed approach exploits a sliding window bundle adjustment technique to use the image coordinates of tie points and the positions and attitudes derived from the last epochs to determine the position and attitude parameters of the most recent epoch. The approach has been evaluated using both real and simulated data from a mobile mapping campaign over an urban area with long GNSS outage periods, with promising results.This thesis also presents an approach to improve the determination of Remotely Piloted Aircraft Systems (RPAS) trajectories using open aerial data obtained in the framework of a national mapping project (PNOA). The development of this methodology is another contribution aiming to ensure the geospatial coherence between the orthophotos, and digital elevation models obtained with an RPAS and the orthophotos and digital models of the PNOA. The results, applied in the context of a multi-temporal and multi-sensor high-resolution archaeological documentation, show that photogrammetric products can be generated with a similar accuracy (cm level accuracy) to the ones generated with more complex approaches. Last but not least, this thesis presents a seamless indoor-outdoor positioning approach with encouraging results (meter-level accuracy) in several scenarios. This contribution opens the door for enhanced tracking of members of civil protection and emergency teams. This is an open field of research with not widely accepted /adopted solution yet.

Author: CHO, MEEHWA
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN URBANISM
Department: (DUTP)
Mode: Normal
Deposit date: 26/06/2023
Reading date: 04/10/2023
Reading time: 11:00
Reading place: Sala de Graus ETSAB. Planta Baixa (E.T.S. Arquitectura de Barcelona)Presencialment
Thesis director: SABATE BEL, JOAQUIN
Committee:
     PRESIDENT: LLOP TORNE, CARLOS JUAN
     SECRETARI: BROSSA I BALCELLS, MARCOS
     VOCAL: SOARES RIBEIRO GOMES CAVACO, CRISTINA
Thesis abstract: After more than a century of existence, a district considered marginal in my city is tragically disappearing without a trace. This inspired me to examine its development and destruction. Busan, the first industrial and the second largest city in South Korea, has experieneced a population explosion three times since the beginning of the modern city. Evidently, urban planning comes after spatial expansion and occupation. Therefore, planned and unplanned districts now coexist within the city of Busan. Both districts have developed through a process of repeated exclusion, expulsion and eventually acculturation of the most underprivileged population at different periods. A spatial discrimination not only erases the ¿other¿ at the real space structure but also at the official history, both in its descriptions and drawings. Therefore, before tackling the morphological analysis, making its existence visible and incorporating it into the historical narrative is essential.This story is not limited to be a testimony of the suffering of the affected citizens, but also intended to reveal the turbulent modern history of the South Korean society that are condensed in the growth of the city: which entered modernity from the most peripheral corner of the world by 19th century imperialism; began the 20th century being occupied by a foreign force; later on in the middle of the last century, suffered an extreme poverty during the actual as well as the ideological war; however, within in a few decades later the country recovered to being a developed one; recently, it achieved a miraculous socioeconomic growth, which, however, hides part of its history.The discrimination that is born from imposing new urban patterns, has reproduced each time a new hegemony. Inevitably entailing a class conflict. In order to reveal the successive attempts at acculturation, or direct expulsion, we tried to discover the process followed by two different urban growths throughout five periods. In which, urban policies were repeatedly subordinated to the objectives of the dominant group and the government, either from an external force or from the Korean society itself. Observations of the dynamics of various agents involved in housing policies revealed the reciprocal relationship between two growths and as a result, the successive urban transformation.Few morphological studies have been addressed in South Korea. Therefore, this research represents a novel contribution to the knowledge of urban formation, which in turns reveals its contribution in building a dramatic inequality. The case study, Busan, has even fewer studies related to its urban history and to the morphological analysis despite its critical importance in the advancement of modern Korean history. We intended to address and to provide for future research projects, an intense morphological analysis that contains varied documentation of formation and transformation along with the historical description with fifteen representative axes.Apropos of the apparent irregularity of that "other city", the analysis revealed patterns, specific rules that resulted from a collective organization, based on a traditional spatial composition. Therefore, the analysis uncovered that this "other city" is not marginal, but rather historical. It is the only place where the traces of traditional space still remain today that is quite different from some heritage elements that are artificially preserved out of their urban context. If one day the South Korean society tries to recover the tradition that has been despised, covered , or considered inferior when it comes to pursuing belated modernization, and also tries to reconnect with the contemporary society, the "other city" would offer, if we arrive on time, the only physical testimony to achieve it.

Author: ROBLES GONZÁLEZ, ANTONIO
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN NETWORK ENGINEERING
Department: Department of Network Engineering (ENTEL)
Mode: Normal
Deposit date: 13/07/2023
Reading date: 05/10/2023
Reading time: 12:00
Reading place: Aula Màster C3-005, Campus Nord (ENTEL-MAT)
Thesis director: PARRA ARNAU, JAVIER | ARIAS CABARCOS, PATRICIA
Committee:
     PRESIDENT: ALMENARES MENDOZA, FLORINA
     SECRETARI: PALLARES SEGARRA, ESTEVE
     VOCAL: MENG, WEIZHI
Thesis abstract: Smart community services are reaching nearly every area of our daily life, often requiring private information from their users. The scope of all contributors to these services is to collaboratively share information for the benefit of all stakeholders, including citizens (user), organizations, schools, and governing institutions. User contribution can be participatory --flence intentionally given --0r smart community services can gather information opportunistically from user sensors and/or APls nearly automatically or with less user influence. The present dissertation focuses on participatory user contribution. SCS increasingly demand consciously undertaken participatory user contributions, that predominantly require a login-based user verification process based on an identification (I) and authentication (A) process. Throughout this process, the user logs in with a user identity, regardless of whether it is a real one or not. The verification process is associated with immanent privacy threats to users. The users can contribute with tagging, posting, or uploading information demanded by the SCS, which may need reliability guarantees linked to the trustworthiness of the users. Nowadays becoming more aware about their privacy and right to self-determine, users are not so willing to contribute unconditionally to the SCS, leading to a conflict between user privacy and the SCS requirements. The verification process for a user login as well as his contributions implies user privacy threats. Chapter 3 and 4 of the dissertation focus on the privacy threat analysis (PTA) of the user verification process and chapter 5 on user self-determined privacy aware contributions to the SCS. Chapter 3 focuses on the PTA of the verification process in the modelling phase. We extend the scientifically grounded LINDDUN PT A framework to be used systematically for modelling of the verification process to perform a user login. Our contribution includes the modelling of the I and A processes, considering IA methods, the extension of the trust boundary concept, and extensively extends the privacy threat mapping table. Our contributions are assembled in a systematic and reproducible step-by-step guide intended for privacy auditors including knowledge and decision support, whereby the results do not depend on the knowledge of the auditor or his intuition. The results provide the requirements for the authentication schemes (AS) to be implemented or selected. Chapter 4 focuses on the PTA of the verification process of realized AS. Bonneau et al. proposed a comparison framework to extensively evaluate AS for usability (U), deployability (D) and security (S), namely, the UDS framework. We extend it with a new defined privacy (P) category to become the UDSP framework. Our evaluation of the 38 AS -including biometrics -with the UDSP framework reveal inter alia fundamental privacy threats, for which we propose guidelines for more secure implementations. Chapter 5 focuses on self-determined and user accepted revocable privacy. The contributing user in particular is exposed to privacy threats when he contributes to a critical incident of a SCS, that requires evidence and trustworthiness for the contribution. That is the reason why we propose a taxonomy concept for classifying the criticality of incidents, including a mapping to enhanced privacy requirements and the cryptographic primitives that would support their realization in a privacy preserving fashion. The taxonomy for user self-determination comprising enforceable graded revocable privacy, which is nonetheless partially applicable to the right to be forgotten, is exemplified for two proofs-of-concepts applying cryptographic primitives alike blacklistable anonymous credentials and group signatures with distributed management.