Real-time, driver-centric notifications and recommendations algorithm for safe-driving and eco-driving

Development of novel machine learning algorithms for real-time, driver-centric notifications and recommendations to drivers for promoting safe-driving and eco-driving behavior. The algorithms will be fed with a sheer volume of real-time and historical data collected from multiple, heterogeneous sources such as (a) vehicle tracking systems capturing on-board diagnostic bus (OBD II) data and global positioning system (GPS) data, (b) weather data providers, (c) infrastructure-less traffic data collected from fleet vehicles, and (d) predefined eco-driving rules that appear in bibliography.

Fleet-centric and driver-centric intelligent metrics and analytics

Development of novel fleet-centric and driver-centric intelligent metrics and analytics for measuring the compliance of each and every driver with system-provided recommendations for eco-driving and thus enabling drivers’ competence classification (e.g. on a reward-penalty scheme basis). Furthermore, the proposed metrics will provide insights into the dependency of fuel consumption on the different fleet aspects (driver behavior, vehicle's technical condition, etc).

Cloud-based scalable, highly available and high performance infrastructure

The consortium will discard obsolete implementations and proceed with the deployment of a big data ready cloud based infrastructure safeguarding reliable operation of the system, and benefiting from all attributes and services cloud based infrastructures offer including resource utilisation dynamicity, reduced failure risks, enhanced scalability, multi-tenancy and more.

Efficient aggregation of multi-source Data

Development of a data collection gateway to receive data from vehicle tracking systems as well as retrieve data from open data providers (e.g. public transport data, weather data, traffic data etc.) that can be made available. Currently, there is no such a gateway, thus data from vehicles belonging to Navarchos FMS is transferred to vehicle tracking hardware provider’s cloud-based infrastructure and then using dedicated API calls are retrieved by the Navarchos FMS.

Facilitation of complex event processing over (potentially big) data warehousing

Design and development of a complex event processing pipeline capable of addressing the main challenges associated with the traditional CEP approaches, namely: 1) latency in the processing which has become very critical, so that CEP tasks have to be parallelized (scalability is an issue), and 2) complexity of the situations to be detected requires networks of CEP engines in order to process the real-time data.

Routing optimization and route planning (scheduling) engine

Routing optimization and scheduling engine becomes a must-have service in modern FMS. It offers remarkable cost savings due to routing optimization (between 10% and 25%), reduced carbon footprint, timely delivery service, improved management control by linking the sales process with warehousing and transportation systems, improved information flow - everyone knows the status of orders at any time, improved reporting and supports strategic logistics network design. The consortium will rely upon open source routing optimisation and scheduling engines and will customise them and deploy them in order to increase the platform efficiency.

Historical Vehicle Tracking System Data Analytics

Automatic region of interest (ROI) detection algorithm, driver proficiency metric in terms of optimal routing vs driver routing, driver routing and scheduling analysis, ROI analysis (i.e. to reveal unauthorized vehicle use), driver comparing and benchmarking including fleet utilization, performance metrics (fuel usage, idling activity, harsh acceleration/de-acceleration) will be extracted.