A Smart Mobility Platform for Electric Vehicles with

A Smart Mobility Platform for Electric Vehicles with Event Processing Müge Kural Fatih Kaan Tuncer Deniz Memiş M. Naci Dai

Meet us Müge Kural Software Developer @ Eteration Fatih Kaan Tuncer Software Developer & Instructor @ Eteration

Future with electric vehicles Figure: Evolution of the global electric car stock, 2013 -2017

… 2019

New Policies EV 30@30 Scenarios Figure: Global EV stock in the New Policies and EV 30@30 scenarios, 201730

A Hybrid Platform Cloud OR Embedded

Event Processing

Event Processing

Event Processing on EV • Io. T board, which receives/transmits data transporting at CAN interface on EV. • Io. T adapter, provides the conversion between raw CAN data and events, • Io. T engine, which is wrapped around Siddhi library, Hardware Software includes processing logic, takes the events coming from Io. T adapter as an input and sends the derived output based on processing logic to the CEP adapters back.

Cloud Integration

Hardware Design (front) 1. Power Supply Module 2. Adapter Module 3. Networking Module (back)

OSGI Gateway UI enables: Io. T Adapter • Necessary information for Io. T adapter such as Vehicle ID, CAN Bus interface selection, publish rate and car- related data channels’ ID through web UI of gateway. • Plan management for Io. T Engine to easily install/update/start/stop plans Io. T Engine

Use case: Reaching to charging stations The unit keeps track of ETA of closest charging stations by a third party traffic API. It also computes the decrease rate of state of charge (SOC) by processing the following parameters: pack capacity (k. Wh), economy* (Wh/km) and trip meter (km). Given the decrease rate in SOC remaining battery life is estimated. Ultimately, given ETAs by traffic API are compared with the estimated remaining life and the unit creates a warning whethere is a risk that a charging station is unreachable. * While calculating economy (Wh/km), all data points within the last one hour range are processed and this res

Use case: Reaching to charging stations Figure: Execution plan diagram for risk detections

Results Until the warning event is triggered, it can be seen that over 2000 events per second are processed in different data streams justifying the functionality of CEP.

Results 1 3 2

Thank you ! info@eteration. com muge. kural@eteration. com fatih. tuncer@eteration. com