The Mobile Conduit: Delivery of Advanced Automotive Services through - - PowerPoint PPT Presentation

the mobile conduit delivery of advanced
SMART_READER_LITE
LIVE PREVIEW

The Mobile Conduit: Delivery of Advanced Automotive Services through - - PowerPoint PPT Presentation

The Mobile Conduit: Delivery of Advanced Automotive Services through the Phone Jacek Serafiski Sen Murphy Zylia, Poland UCD, Ireland sean.murphy@iname.com jacek@zylia.pl http://zylia.pl Automotive Lunix Summit, Edinburgh 24.10.2013 Zylia


slide-1
SLIDE 1

Jacek Serafiński Seán Murphy Zylia, Poland UCD, Ireland jacek@zylia.pl sean.murphy@iname.com http://zylia.pl

Automotive Lunix Summit, Edinburgh 24.10.2013

The Mobile Conduit: Delivery of Advanced Automotive Services through the Phone

slide-2
SLIDE 2

Zylia profile

 R&D consulting company with strong practical focus

 Prototype development our primary activity

 Expertise

 Embedded systems  Speech recognition systems  Audio processing and coding  Mobile-car integration

 Work on mix of public- and private-funded projects  Interested in leveraging expertise in embedded

systems in Connected Car context

slide-3
SLIDE 3

Outline

 Connected Car vision, concept, issues  The role of mobile in the Connected Car  Mobile-car integration technologies  Carmesh Implementation work

 Platform-specific experiences  Demo applications

 Conclusions

slide-4
SLIDE 4

Market forecast for Connected Car

 Automotive sector sees the connected car market as a big

  • pportunity to offer new services

 Automotive applications market should reach US$ 1.2bn by

2017 (Juniper research)

 GSMA forecasts that:

 More than 20% of vehicles sold worldwide in 2015 to include

embedded connectivity solutions;

 More than 50% of vehicles sold worldwide in 2015 to be

connected (either by embedded, tethered or smartphone integration);

 Every new car to be connected in multiple ways by 2025.

slide-5
SLIDE 5

What is the Connected Car?

 Two main aspects to the Connected Car:

 Integration with existing services (social networks,

calendar, entertainment)

 Tailored automotive experience

 Right data, right time  Advanced Uis – mix of speech, in car controls, etc.

 New data making new services possible

 Driving behaviour analysis  Eco-driving  UBI – Usage-based Insurance  Dashboard apps (Torque, Waze)

 But it’s much more messy of course…

slide-6
SLIDE 6

Challenges realizing Connected Car

 The expected lifetime of a car is 10 years thus the

deployment of new services to older cars might not be possible

 Diversity in capabilities of different cars

 There are problems to keep the car connected while

moving at high speed and/or in congested areas

 Privacy issues regarding usage of vehicular data  Driver distraction issues for new services need to be

resolved

 Application developers don’t know how to design for the

automotive context

 New user interfaces  Present the right data at the right time

slide-7
SLIDE 7

Increasing influence of mobile

 Power of mobile increasing at phenomenal rate

 Increasing wireless interfaces’ capacity

 OBU will not be able to compete with the mobile in terms of power,

connectivity or release cycles

 Mobile already knows much about you

 Home, work, apps, locations, etc.

 Why not leverage Internet on the mobile?

slide-8
SLIDE 8

Why not replace OBU with a mobile?

 Heterogeneous mobile world makes neat integration hard

(size, computation power, device connectors)

 Security issues - some car data and functionality needs to

be locked from mobile developers (e.g. remote engine start-up, etc.)

 Safety regulations, strict QA processes  Auto manufacturers want to monetize automotive

applications

 want to have some control over the ecosystem

slide-9
SLIDE 9

OBU and Mobile – harmonious marriage

 Natural solution leverages strength of different systems

 OBU provides better integration with Car and driving

experience

 Mobile provides compute power and network connectivity

 As in every marriage there are some tensions…

slide-10
SLIDE 10

Tensions

 Some obvious touchpoints

 Speech  Navigation  Control of the application store/service delivery

platform

 Need holistic view in which functionality split is clear

 System design should allow some flexibility  Standard communication protocols for orchestration

required

 Car manufacturers will not be tied to single mobile

platform

slide-11
SLIDE 11

Mobile-car integration

 Solutions are not yet mature…

slide-12
SLIDE 12

Mobile-car integration

 There are different approaches towards mobile-car

integration

 Basic integration – contacts, call support, no data connection  Single in-house mobile app (e.g. Mercedes, Lexus)  Mirroring solutions (e.g. MirrorLink, iOS 7)  Application level solutions (Ford Applink, OnStar,

SmartDeviceLink)

slide-13
SLIDE 13

Application level solutions for mobile-car integration

 Applications run on the phone and connect wirelessly

to the car

 Multiple applications connected to the car can run

independently from each other

 Car provides UI components for mobile app

 Mobile app is able to control the output components

(display, audio)

 Callbacks trigger mobile applications on input from car

 Applications can access car data  Open SDK for application developers for different

mobile platforms

slide-14
SLIDE 14

Why were we interested in SDL?

 Started looking at it early this year  Was the most open phone-car interaction solution

 Ford are championing it and doing quite good

developer outreach

 Although Mirrorlink has more industry support, developer

  • utreach less mature

 We like the vision of brains in phone and car

provides UI components

 very compatible with SDL

slide-15
SLIDE 15

Carmesh Implementation

 Working with new platforms is challenging…

 Experiences with SDL  Experiences with Tizen IVI

 We are working on three demonstration

applications using SDL in Tizen IVI

 Google Calendar integration  Facebook integration  OBDII data collection via RPi

slide-16
SLIDE 16

Overview of SDL

 Phone-car interaction protocol

 Evolution of Ford’s Applink

 Enables individual mobile applications to interact with car  Being pushed (somewhat) within GENIVI consortium  On Tizen IVI v3.0 roadmap  SDL ‘richer’ view of mobile-car interaction than mirroring

 iOS, MirrorLink

 There is a quite well defined protocol right now

 although it is somewhat missing an architecture

slide-17
SLIDE 17

SDL - the working parts

Mobile

SDL Library

(iOS, Android)

OBU

SDL Proxy

SDL Javascript Library Web Sockets

HMI

WiFi or Bluetooth

slide-18
SLIDE 18

SDL working parts

Mobile

 iOS/Android libs  to be included in SDL

compatible applications

 understands SDL

primitives and can maintain communications with OBU

OBU

 Proxy running as OS process  C++ implementation  maintains communications with

phone and sends to browser

 quite robust  Browser based component  Websockets interface to proxy  Javascript library for

registering for notifications, message composition etc

 mostly designed around

callbacks

slide-19
SLIDE 19

Getting it going - OBU

 cmake build process  Did not compile out of the box on either Tizen or

Ubuntu

 Had some issues around Bluetooth and websockets SHA

calculation

 Build generates binary which runs proxy and

launches HMI

 chromium by default

 HMI offers main automotive UI and emulates

automotive buttons

slide-20
SLIDE 20

Getting it going - Android

 System comprises of library and application  Straightforward to build and deploy for Android  Small issue with default settings for WiFi mode being

incorrect

 No discovery mechanism for WiFi

slide-21
SLIDE 21

Limitations of SDL (current)

 Control over in car display is very limited

 Logic required on OBU side to know how to render

message

 No real graphics support

 Little navi integration

 not possible to send POI to in car navi  some TurnByTurn updates available, though

 Features supported by protocol not yet implemented

 vehicle data, application types/templates, file transfer

slide-22
SLIDE 22

Tizen IVI

 Application development and deployment process

 We are using Tizen SDK 2.2 to build Web Application

 HTML5-based UI components

 Downloading Tizen IVI snapshot image from the

repository

 New revisions uploaded on monthly basis

 Sign the application and upload to running Tizen IVI

image in VM Player

 Install and run applications

 wrt-installer –i your_app.wgt  wrt-launcher –l (to get the ID of the installed application)  wrt-client –l your_app_ID

slide-23
SLIDE 23

Demonstration applications

slide-24
SLIDE 24

Technologies used in demo apps

 Android 4.3  SmartDeviceLink (from 24.04.2013 -

http://git.projects.genivi.org/)

 Tizen IVI 3.0-M2-Jul  VMWare Player 5.0.2 build-1031769  WiFi for phone-car communication

 Manual configuration

slide-25
SLIDE 25

Google Calendar with SDL

 Basic application to provide GCal integration in car  Can see next upcoming appointments and navigate

to them

 assuming location data in GCal event

 Use in car controls to control application

slide-26
SLIDE 26

GCal app architecture

GCal App Logic

(callbacks on button press, update UI)

Display and car controls

Car Mobile

SDL

slide-27
SLIDE 27

Facebook application

 Fetch locations from Facebook account

 Show these locations relevant to current position  Provide locations of friend’s check-ins, places tagge in,

etc.

 FB-based user management (including authentication)

 Architecture similar to the previous application  Technologies used

 Android Embedded browser  REST API  Facebook Graph API

slide-28
SLIDE 28

OBDII data collection with RPi

 Connection to ODBII interface

 Mobile application connects to the RPi and stores the

data and sends them to the OBU

 OBDII to USB serial interface  PyOBD library  Raspberry Pi powered by car  RPi as an access point

slide-29
SLIDE 29

Conclusions

 Connected Car will change how we use a car

 Routing, eco-driving, looking for gas station, etc.

 Mobile is the key part of the Connected Car

 Mobile becomes context-aware

 There is no single view of relationship between car

and mobile

 Development tools and platforms are still basic

 Little understanding of how to develop for automotive

context

 GENIVI is developing good technologies to make

Connected Car happen

slide-30
SLIDE 30

Future activities

 Leveraging open in-car systems such as SDL  Tizen IVI prototypes and applications  Back-end for vehicular cloud-based services  Vehicle-specific speech recognition systems  Novel vehicular applications  Over-the-air vehicle software updates  In-vehicle (surround) audio coding  Looking at EU-funded collaborations in the area

slide-31
SLIDE 31

Obligatory cat…

slide-32
SLIDE 32

OK, Let’s be serious…

slide-33
SLIDE 33

http://carmesh.eu http://blog.carmesh.eu/

@carmesh_fp7

jacek@zylia.pl

THANKS!