WP1: Specifications & Requirements
This WP deals with the definition of the specifications, requirements and research frames, i.e. specifies the research questions that are of value to be answered and within the project’s reach given the resources and time. It does so by identifying relevant requirements, specifications and focusing on the work that needs to be done on top of existing assets (technology, tools, etc.) to ensure that the project will narrow down its effort’s scope to challenges that their solution will have a maximized impact from a societal, scientific and business perspective.
In short: This WP sets the scope to maximise the societal, scientific and business impact.
Leading partner: Konnect-able Technologies limited
WP2: Data Processing Architectures & Infrastructure (BDMI Toolbox)
Data Processing Architectures & Infrastructure aim to develop innovative methodologies to advance architectures for heterogeneous structured and unstructured data stream exploitation by also including batch and interactive data sources generated in an urban setting. These architectures will not only be limited to real-time data generated in an urban ecosystem but also based on archival and contextual data. Individual streams are noisy, unreliable and contain only partial information. Jointly analysing all streams together and merging information with contextual and archival data is the key issue to provide reliable information. The advances proposed in this task answer several challenges in heterogeneous stream and batch processing. More specifically, we will pursue the design of an optimized architecture and implementation of a system prototype allowing the seamless integration of data sources and processing components and the implementation of ready-to-go integration connectors that allow the interoperability of internal and external data sources.
In short: In this WP we develop a methodology to advance architectures, used for exploitation of heterogeneous data streams.
Leading partner: Intrasoft International SA
WP3: Big Data Processing Toolboxes Management (BDP Toolbox)
Track&Know relies on the consolidation of voluminous, high input rate data in different formats, from various sources, in different modalities into a Big Data management system. Incoming data include streams of mobility data as well as sensor measurements. Track&Know aims to implement data acquisition technology suitable to consume data from heterogeneous streaming and archival data sources, including sensors, social networks, and crowd-sourced data. To provide a uniform view over the collected data, data integrators for Big Data are going to be developed aiming at providing a common data representation, thus dealing with the variety dimension of Big Data. To complement the acquisition and integration of Big Data, Track&Know plans to deliver a parallel / distributed storage system for the efficient management of underlying data, combining different scalable NoSQL stores according to their suitability to the particular data at hand. Several fundamental issues related to highly distributed storage need to be revisited leading to innovative solutions, including: local and global indexing, data partitioning techniques, and load balancing. Last, but not least, primitive query operators will be implemented that facilitate data access, but at the same time remain abstracted in order to be readily implemented for different NoSQL stores. The query operators provide a software layer as close to the storage system as possible, which will allow easier integration to the software stack and flexible interaction with the Big Data processing architecture as well as the various Toolboxes.
In short: In WP3 we implement data acquisition technology to capture data of heterogeneous data sources.
Leading partner: University of Piraeus research center
WP4: Big Data Analytics (BDA Toolbox)
Big Data Analytics (BDA) Toolbox harvests huge volumes of trajectories from Floating Car Data (FCD). When these data are considered in the context of urban environments, analysis methods can facilitate the extraction of useful knowledge and the provision of sophisticated services towards vehicles’ drivers, citizens, stakeholders and city operators. The outcome of these methodologies can be exploited and combined with contextual information to automatically detect event occurrences within a system or a modelled domain, known as Complex Event Recognition (CER). An event recognition system accepts streams and archival data of low-level events (for example sensor data) and uses predefined patterns to recognize high-level events of interest, that is, events that satisfy/match these patterns. As an example, consider the real-time detection of crowd or vehicle mobility patterns in an urban area, or events related to traffic in areas at different hours (high-level events), via combining low-level events originating from in-vehicle or static sensors and probes, or user reports on social networks (e.g. tweets).
The need for novel approaches and data analytics poses severe strain on efficient transportation and planning, as the urban landscapes are evolving very fast in the Big Data era. Analysis tools and CER tasks can also assist the improvement of several environmental aspects and the quality of daily life as they can indicate insights regarding safety, risk assessment. The BDA Toolbox enables us to better understand key mobility factors, which drive automotive transportation or serve as hubs for accessibility to the respective urban zones of interest. They also enable us to characterize traffic flows, regular routes, drivers’ behavioural patterns in the space and time, as well as their relationship and mutual interaction with the contextual environment.
This WP plans to introduce novel methodologies within Toolboxes which will enable citizens and stakeholders to draw useful conclusions regarding the spatiotemporal distribution of traffic flows using data gathered and fused from a wide variety of sources. It also addresses the issue of recognizing geographic borders, which are formed instantly (i.e. vehicles’ masses, events, accidents, etc.) for situational awareness and risk assessment or periodically (daily, weekly, monthly itineraries). A particular emphasis will be given to distributed algorithmic approaches and unified view over vehicles’ data, aimed towards scalability, interoperability between different computing environments and adaptability to variable data rates. We also plan to provide event recognition technology capable of consuming massive data of high-speed. Using these methods and tools, forecasting models will be designed for the short- and long-term modalities that describe these mobility patterns in the future.
In short: Developing a methodology to analyse heterogeneous data and to draw conclusions about spatiotemporal distribution of traffic flows, is the focus of this WP.
Leading partner: Consiglio Nazionale Delle Richerche
WP5: Data Visualization & User Interaction (VA Toolbox)
Data Visualization and User Interaction Toolbox aims at developing interactive and scalable methodologies, which can efficiently handle both historical and streaming spatiotemporal data originating from different sources, with varying levels of resolution and quality. All potentially useful possible transformations between the types of spatiotemporal data [12] will be considered and supported. Appropriate VA methods will be developed for selected types of spatiotemporal data. FRHF leads this work package, being responsible to deliver the main results, associated with its tasks.
In short: In this WP we develop an interactive and scalable methodology to visualize spatiotemporal data.
Leading partner: Fraunhofer Gesellschaft Zur Foerderung Der Angewandten Forschung E.V.
WP6: Use cases and domain validation
The concept of the pilots is to provide the test-bench to the stakeholders, in order to be able to evaluate the Track&Know innovations and Toolboxes with realistic data (in terms of content, but also in terms of using large data volumes, to test non-functional parameters of the system, such as performance, scalability, responsiveness, usability). The process in developing the pilot use cases starts with the involvement of business stakeholders in WP1. Development of tools and front-end applications and dashboards will be agile, so that a constant interaction cycle of progress will be delivering the results incrementally as services. To this end, the pilots will be following Plan – Do – Check – Act cycles. There will be a constant interplay between the pilots’ progress and the technology developments in WP2-WP5.
In short: To put theory intro practice, we develop pilot cases that serve as a test-bench for the stakeholders to evaluate developments using realistic data. Therefore a constant interaction between testing (WP6) and developing (WP’s 2 to 5) occurs.
Leading partner: SISTEMATICA
WP7: Dissemination & Exploitation
This WP will: (a) Specify and apply user acceptance criteria and an inclusive stakeholder engagement strategy, to attract extended participation of key players wishing to join forces with the Track&Know platform, (b) Make policy, standardisation and research recommendations particularly regarding simplification and harmonisation of laws associated with data privacy and data sharing, (c) Plan and act exploitation, addressing both individual partner’s plans and collaborative arrangements to provide a solution, (d) Provide a communications programme linked to the adoption/exploitation plan. Key elements will be techniques for awareness raising, knowledge transfer and early participation in Track&Know.
In short: Disseminating project progress and results, exploiting and commercialising project outcomes, collaborating with stakeholders, establishing dialogue and collaboration activities with other projects to identify commonalities and outcomes … are the core activities of this WP.
Leading partner: Hasselt University
WP8: Management & Coordination
WP8 ensures the completion of all deliverables in time, within budget and to the required quality standard. This includes administrative, technical, innovation and quality management.
Leading partner: Inlecom Group BVBA
