Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - Track and Know (Big Data for Mobility Tracking Knowledge Extraction in Urban Areas)

Teaser

There is a strong demand for efficient and scalable smart services using mobility data, but this imposes new requirements to better exploit the immense and continuously rising amounts of data. The business impact of Big Data is starting to take form in many business fields...

Summary

There is a strong demand for efficient and scalable smart services using mobility data, but this imposes new requirements to better exploit the immense and continuously rising amounts of data. The business impact of Big Data is starting to take form in many business fields. However, innovation is held back by limitations of current Big Data processing methods and infrastructures as in most cases smart services remain uncoupled and isolated.
Track & Know researches, develops and exploits modern software frameworks that aims to increase the efficiency of Big Data. A variety of toolboxes (that contain specific methods/functions/algorithms for various types of data aggregation, manipulation and further analysis) are developed within the project and integrated in a software platform. A Big Data Processing (BDP) toolbox is developed to implement data acquisition technology that captures data from heterogeneous data sources. The BDP toolbox extends the current solutions and delivers a tool for efficient access, indexing, partitioning and load balancing for Big spatiotemporal data. A Complex Event Recognition (CER) toolbox detects complex event occurrences by analysing patterns in simple events. To do that, it uses contextual information and results from the Big Data Analytics (BDA) toolbox. For example: the toolbox may infer a complex event (such as dangerous driving or non-economical driving) by analysing patterns based on vehicle speed, direction, driver events, fuel consumption and other contextual information such as weather etc.
The BDA toolbox is developed to analyse heterogeneous data and to draw conclusions about the spatiotemporal distribution of mobility patterns. The BDA toolbox delivers scalable data mining techniques (such as clustering, sequence mining, hot-spot analysis) for voluminous offline and streaming trajectory data. A Visual Analytics (VA) toolbox develops interactive and scalable methodologies to visualise data at all steps of analysis. The VA toolbox can efficiently handle both historical and streaming spatiotemporal data originating from different sources, with varying levels of resolution and quality. To put theory into practice, we integrate the toolboxes in a platform and test them in pilot cases. We organise pilot cases in the three domains with two common links: service optimisation and driver behaviour.
Within the Insurance sector we are looking at: (a) using historic telematics, environmental, demographic and geographic information to gain in-depth and accurate crash probability estimation; (b) Electric Cars adoption by studying the cost-benefit of a switching to an electric car mobility, matching global charging times and charging points to drivers’ habits; (c) and finally Car Pooling opportunities by analysing for parking decreasing due to sharable routes, cost-benefit of switching to a sharing mobility paradigm, and likelihood of finding a proper sharable route that matches time and geographical zone.
Within the Health Service Track & Know aims to (a) improve the response times (increasing new patients, follow-up patients; (b) reduce unnecessary travel (reduce patient travel distances, courier costs, CO2 emissions); (c) generate Cost efficiency gains; (d) New methods of OSA diagnoses based on driver behaviour.
For Fleet Management, Track & Know’s Business objectives include Predictive maintenance, Anomaly detection and reduction of false alarms, Correlation of Fleet Data with external Weather and Traffic services, Fleet costs reduction, Fleet downtime reduction, Fleet response time improvement, and Improve driver behaviour and reduce accidents.

Work performed

The project has thus far:
- Established an Online Observatory comprising a literature survey of different facets of mobility analytics, including driver profiling, bad driving identification, journey clustering etc. and a detailed analysis of interoperability standards for Big Data from a mobility context
- Created a highly scalable interoperable Big Data platform for streaming and at rest data in continuous operation since February 2019
- Developed a data cleaning, map matching and enrichment pipeline to deal with streaming and historic heterogeneous mobility data, in the form of different spatio-temporal resolutions, noisy signals, with identification of events within the data to be augmented with weather, point of interest and traffic data
- Creation of data access operators over NoSQL stores has enabled the Track&Know platform to store data in multiple formats and using different storage systems along with interfacing with other remote services using just a common framework
- Individual mobility network discovery with applications to crash prediction
- Developments in future location prediction improving the accuracy of prediction of trajectories, congestion and routing, ultimately leading to CO2 and accident reduction
- The complex event recognition algorithms are able to leverage the enriched data pipeline to identify high-speed driving, dangerous or non-eco driving, and refuelling opportunities using inputs such as road network information, abrupt acceleration/deceleration/cornering, points of interest information, and weather information
- Contextual Analysis of Movement Events through visual analytics has enabled greater insights and understanding into the pilot’s data. This work has resulted in the Best Paper award at the 10th EuroVA 2019 conference
- Customisable analytics dashboard to support cross-scale analysis and location-allocation analysis has been designed for the pilots
- Creation of an app for OSA patients to use during their diagnosis phase under the healthcare pilot

Final results

Track & Know foresees several socio-economic impacts from the application of the toolboxes. For the pilots themselves:
- Reduction of patient travel times, resulting in CO2 reductions, and operating cost reduction have already been realised through the results and trials
- Reduction in accidents by approaching all sectors: Insurance, Health, and Fleet Management
- Reduction of CO2 in everyday mobility applications
Track & Know has made several key innovations, and exploitable developments within the first period. The second period will see some of these innovations taken to market.
Implemented Prototypes:
1. Big Mobility Data Integrator Platform – the Track&Know platform capable of processing Big Data streams and achieves
2. Data Cleaning, map matching and Enrichment Pipeline – integrated into the Track&Know platform to pre-process at scale large amounts of data and augment with POI, Weather, and Map data
3. Future Location Prediction components for online / streaming processing, specifically the P-RMF* and ARIMA variants (including data preparation & filtering)
4. The project is involved in testing its Big Data Analytics and Complex Event Recognition toolboxes that leverage the use of deep/machine learning to provide beyond the start of the art insights into large scale Big mobility data
5. The project is utilising visual analytics to identify complexities with data at scale and integrating such approaches (usually manual) into a dashboard system
Unique Value Propositions
1. Identification of the potential for vehicle electrification
2. Identification of driver risk leading to accidents
3. New approach to remote and distributed service delivery for healthcare and other sectors

Website & more info

More info: https://trackandknowproject.eu/.