Abstract

div"Scientific problems are solved by finding the optimal solution for a specific task. Some problems can be solved analytically while other problems are solved using data driven methods. The use of digital technologies to improve the transportation of people and goods, which is referred to as intelligent mobility, is one of the principal beneficiaries of data driven solutions. Autonomous vehicles are at the heart of the developments that propel Intelligent Mobility. Due to the high dimensionality and complexities involved in real-world environments, it needs to become commonplace for intelligent mobility to use data-driven solutions. As it is near impossible to program decision making logic for every eventuality manually. While recent developments of data-driven solutions such as deep learning facilitate machines to learn effectively from large datasets, the application of techniques within safety-critical systems such as driverless cars remain scarce."/div""div"Autonomous vehicles need to be able to make context-driven decisions autonomously in different environments in which they operate. The recent literature on driverless vehicle research is heavily focused only on road or highway environments but have discounted pedestrianized areas and indoor environments. These unstructured environments tend to have more clutter and change rapidly over time. Therefore, for intelligent mobility to make a significant impact on human life, it is vital to extend the application beyond the structured environments. To further advance intelligent mobility, researchers need to take cues from multiple sensor streams, and multiple machine learning algorithms so that decisions can be robust and reliable. Only then will machines indeed be able to operate in unstructured and dynamic environments safely. Towards addressing these limitations, this thesis investigates data driven solutions towards crucial building blocks in intelligent mobility. Specifically, the thesis investigates multimodal sensor data fusion, machine learning, multimodal deep representation learning and its application of intelligent mobility. This work demonstrates that mobile robots can use multimodal machine learning to derive driver policy and therefore make autonomous decisions."/div""div"To facilitate autonomous decisions necessary to derive safe driving algorithms, we present an algorithm for free space detection and human activity recognition. Driving these decision-making algorithms are specific datasets collected throughout this study. They include the Loughborough London Autonomous Vehicle dataset, and the Loughborough London Human Activity Recognition dataset. The datasets were collected using an autonomous platform design and developed in house as part of this research activity. The proposed framework for Free-Space Detection is based on an active learning paradigm that leverages the relative uncertainty of multimodal sensor data streams (ultrasound and camera). It utilizes an online learning methodology to continuously update the learnt model whenever the vehicle experiences new environments. The proposed Free Space Detection algorithm enables an autonomous vehicle to self-learn, evolve and adapt to new environments never encountered before. The results illustrate that online learning mechanism is superior to one-off training of deep neural networks that require large datasets to generalize to unfamiliar surroundings. "/div""div"The thesis takes the view that human should be at the centre of any technological development related to artificial intelligence. It is imperative within the spectrum of intelligent mobility where an autonomous vehicle should be aware of what humans are doing in its vicinity. Towards improving the robustness of human activity recognition, this thesis proposes a novel algorithm that classifies point-cloud data originated from Light Detection and Ranging sensors. The proposed algorithm leverages multimodality by using the camera data to identify humans and segment the region of interest in point cloud data. The corresponding 3-dimensional data was converted to a Fisher Vector Representation before being classified by a deep Convolutional Neural Network. The proposed algorithm classifies the indoor activities performed by a human subject with an average precision of 90.3%. When compared to an alternative point cloud classifier, PointNet[1], [2], the proposed framework out preformed on all classes. "/div""div"The developed autonomous testbed for data collection and algorithm validation, as well as the multimodal data-driven solutions for driverless cars, is the major contributions of this thesis. It is anticipated that these results and the testbed will have significant implications on the future of intelligent mobility by amplifying the developments of intelligent driverless vehicles."/div

Abstract

OpenStreetMap (OSM) is an openly-licensed geospatial database created and edited by millions of volunteers worldwide. This database is used in various humanitarian projects, such as mapping during disaster relief operations or malaria elimination campaigns. In some parts of the world, for example certain areas in Mozambique, no official cartographic data is available and hence OSM can play a crucial role. A mapathon (literally “map marathon”) is a collaborative effect that is an effective method of collecting map data in unmapped areas. Through the YouthMappers global initiative (http://www.youthmappers.org), students are actively engaging in OSM-based collaborative mapping efforts internationally. We will briefly discuss the process followed by three YouthMappers chapters; Politecnico di Milano (Italy), Texas Tech University (United States of America), and University of Pretoria (South Africa), and lessons learned. In Milano, students’ mapathons are organized monthly and they are focused either on field-mapping (e.g. to improve the OSM map of the university) or remote mapping on humanitarian tasks; learning sessions on OSM editing tools are also held periodically. In Texas, students organize mapathons several times per year, and they can take a dedicated service-learning course for credit where they work on teams with other students around the world to map the same areas and research the humanitarian or development issue affecting the region. In Pretoria, monthly mapathons are held and each mapathon focusses on a different area. Extra credits are offered to encourage participation. From our experiences, it became evident that even though OSM tools are user-friendly, students do need some training, and a short training video as part of the introduction presentation was developed. Additionally, understanding the purpose for a mapathon in an area, motivates and encourage active participation and ensures that quality data is captured. Internationally students participate and learn from other YouthMappers chapters through regular organized web conferences. In conclusion, incorporating mapathons into the curriculum encourages students to contribute needed data required for humanitarian projects. Simultaneously, students get practical experience in data capturing and awareness is raised of humanitarian challenges in other parts of the world. 

Abstract

Spatial decision-support tools can guide decision-makers towards knowledge-based spatial planning. Their application for multiple-objective zoning is beneficial for supporting the on-going Maritime Spatial Planning (MSP) process, in particular in areas of high density of uses like the Adriatic-Ionian region (AIR)."/p" "p"In this work, we give contribution to the sustainable development of the aquaculture, as one of the main maritime activities for sustainable Blue Growth in the AIR. Aquaculture expansion should occur alongside the protection of biodiversity and the conservation of sites where aquaculture and its synergic uses can coexist. Planning for these objectives independently from one another has been shown to produce inefficiencies in costs, area, and conflicts across sectors. The objective of this research is to operationalize multi-objective zoning for aquaculture expansion and biodiversity protection (seabed habitats, marine mammals, seabirds and turtles species distribution, nursery and spawning areas of commercially important fish species). The area of analyses is the coastal area of the Emilia Romagna Region (Italian Northern Adriatic Sea). This area is the main producer of mussels in Italy and it is currently intensively crowded by several maritime uses, which are expected to grow over next years resulting in increase of conflicts and pressures on sensitive environments (Natura 2000 sites; Biological Protection zone)."/p" "p"The decision support tool Marxan and its advanced version Marxan with Zones were applied to develop planning scenarios. Firstly, we prioritise biodiversity conservation, through Marxan, while minimizing conflicts with other anthropic uses. Secondly, we carry out an aquaculture suitability analysis using a multicriteria evaluation approach based on environmental, biological and socio-economic conditions. Next, we simultaneously prioritize both aquaculture and biodiversity, through Marxan with Zones, while minimizing impacts on seven other industries operating in the Emilia Romagna Region."/p" "p"We discuss an integrated zoning plan that meets targets for multiple objectives, and which adheres to several constraints regarding the location of aquaculture farms offshore in order to both allow the use of bigger technologies for mussel catch and minimize the impacts on coastal environments (pollution impact, visual impact)."/p" "p"Our analysis can guide multiple stakeholders and decision-makers towards cost-efficient solutions that are scientifically based, thus contributing to the marine spatial planning process and sustainable growth in the Adriatic-Ionian region.

Abstract

This presentation was aimed to explore the growing opportunities to merge two booming fields: deep learning and autonomous vehicles, from a technical point of view. It addressed some Intelligent Systems Laboratory (Universidad Carlos III de Madrid, Spain) developments in this line of research, such as an obstacle detection framework using convolutional neural networks (CNNs). Furthermore, it presented a large number of challenging driving-related tasks that were expected to become tractable through this new approach, with the focus on the strong requirements posed by the upcoming self-driving systems."/p" "p"This presentation was part of the 6th LSI Ph.D. Meeting, which was held on 14 Jun 2016 at the Escuela Politécnica Superior of the Universidad Carlos III de Madrid. It was published on Zenodo as an exercise within the THOR Bootcamp on Open Data, organized on 16 Nov 2016.

Abstract

These slides were presented at State of the Map 2017 conference in Aizu-Wakamatsu, Japan (http://2017.stateofthemap.org). They aim to provide an overview on some current research trends on OpenStreetMap (OSM), focusing in particular on three research works I have carried out over the last few years: 1) an automated procedure for quality assessment of OSM road networks; 2) an analysis of OSM tagging patterns; and 3) the use of OSM to generate Land Use/Land Cover maps.

Abstract

These slides (in Italian) were presented during the FOSS4G-IT 2017 conference, held in Genova (Italy) on February 9-10, 2017 (http://www.dicca.unige.it/geomatica/foss4git_2017/index.html). The presentation is about the exploitation of OpenStreetMap (OSM) database in the emergency stage and focuses on the activities of the Italian OSM community after the earthquakes in Central Italy happened in late 2016.