Future of Autonomous Driving An Overview of Market

Future of Autonomous Driving An Overview of Market & Technology Roadmaps, Changes in Architecture, Design / Testing / Validation of Automated Vehicles Prana Natarajan, Team Leader Nick Ford, Sr. Consultant Automotive & Transportation A Joint Presentation with Harald Barth Product Marketing Manager Dr. Anthony Baxendale Research Manager © 2013 Frost & Sullivan. All rights reserved. This document contains highly confidential information and is the sole property of Frost & Sullivan. No part of it may be circulated, quoted, copied or otherwise reproduced without the written approval of Frost & Sullivan.

Today’s Presenters Prana T Natarajan Nick Ford Team Leader, Sr. Consultant, Frost & Sullivan Mr. Harald Barth Product Marketing Manager Driving Assistance Valeo Bietigheim-Bissingen, Germany Dr. Anthony Baxendale Research Manager Future Transport Technologies MIRA Ltd • Prana leads a team of research analysts covering the chassis & safety systems market, besides tracking industry trends such as functional safety, 48 V power net and automated driving • Nick is a senior consultant with prior experience of having served a leading chassis & safety systems supplier as a Global Product Planning Director • Harald is a highly experienced product marketing manager with immense product-related knowledge in commercially launching various safety products, now related to automated driving. • Anthony is a highly experienced research manager with a proven record of developing and delivering high technology solutions to the automotive, defence, security and telecommunications sectors. Prana Natarajan : uk. linkedin. com/in/pranat/ Nick Ford: uk. linkedin. com/pub/nick-ford/1/254/566 @FS_automotive Anthony Baxendale: uk. linkedin. com/pub/anthony-baxendale/7/754/989 2

Agenda No. Topic Page 1 Overview of Automated Driving 4 2 Global Analysis of Industry Activity on Automated Driving 5 3 Comparative Analysis: Functional roadmap, technology preferences & alternatives 6 4 Pricing implications: Consumers, OEMs & suppliers 9 5 The Curious Case of the Google Car: What, how & why? 10 6 Case Study: Valeo Park 4 U – the solution & challenges addressed 12 7 Design, Testing & Validation of Automated cars 15 3

Macro Level Outlook of Automated Driving – Driver Involvement While capabilities exist with automakers to create semi- or highly-automated vehicles today, the biggest challenge in taking the driver out of the loop is to make the car think like a human driver Automated Driving Market: Levels of Automated Driving, Europe & North America, 2013 Level of Automation Semi-Automated Module Highly-Automated Fully-Automated Steering, braking, acceleration, monitoring, access Application EPS, EBS+ Electric Parking Brake, Electronic Throttle Control, ADAS, keyless entry Intersection assist, redundancy back-up for connectivity, self-driving capability until driver takes over control (~10 seconds) Multiple redundancies (hardware) and Artificial Intelligence (Software) Length/ Duration of Automation (miles, seconds) Low Medium High Medium NIL Driver Involvement (miles, seconds) Source: Frost & Sullivan analysis. 4

OEM Comparative Analysis – Overall Automated Driving Strategy While traditional OEMs tend to target an identical launch plan across functions, segments and regions, Google’s model agnostic approach could be the game changer Automated Driving Market: Comparative Analysis of Sensor Suite, Europe & North America, 2013 Automation Level Launch Year First Models BMW Semi. Highly 2014 2017 X 5 5 - / 7 -series Europe & North America Mercedes Benz Semi. Highly 2014 2017 E-Class S-Class Europe, North America, Australia Audi Semi. Highly 2016 A 8 Europe & North America Volkswagen Semi- 2017 Passat Europe & North America (Cadillac) 2018 >2020 SRX, ATS & XTS North America Highly- Ford Semi. Highly 2017 2020 Fusion Explorer Europe & North America Lexus Highly 2020 LS Japan, Europe & North America Nissan Highly 2020 Leaf Europe & North America Volvo Semi 2015 XC 90 Europe & North America Google Fully- 2018 Model / OEM agnostic North America OEM General Motors Semi- TJA – Traffic Jam Assist Top Level Functions TJA APA – Autonomous Parking Assist Region of Introduction Hwy – Automated Highway Driving Source: Frost & Sullivan analysis. 5

Sensor Portfolio for Automated Driving: Comparative Analysis Clear divide exists between OEMs more so in terms of periphery-fitted Laser versus roof-mounted LIDAR; Mercedes-Benz leads the way in leveraging stereo-vision to create 6 D-vision for future vision-sensor fusion Automated Driving Market: Comparative Analysis of Sensor Suite, Europe & North America, 2013 OEMs Radar LIDAR* Camera Laser# Ultrasonic Comments Volkswagen Mono-Vision Stereo vision not a current priority. BMW Mono-Vision Focus on peripheral laser, rather than internalization of Lidar Audi Stereo-Vision Bumper mounted Lidar offers lesser “field of view” Mercedes Benz Stereo-Vision 6 D vision leverages stereo camera for traffic movement GM Mono-Vision Current sensor suite not sufficient for AD Volvo Stereo-Vision Cautious approach in spite of being technologically astute Google Car Mono- / Stereo-Vision Looks at fully automated driving only * LIDAR refers to the typical roof-mounted multi-laser unit #- refers to lasers seamlessly integrated with vehicle periphery Source: Frost & Sullivan analysis. 6

Autonomous Driving: Application Roadmap While ADAS functionalities form the basis for vehicle autonomy, connected solutions ensure getting the most out of automated driving Automated Driving Market: Functional Roadmap, Europe & North America, 2010 - 2020 Semi-automated Highly-automated Autonomous emergency Lane Change Traffic Jam braking system assist Automated (Partial Parking Assist Semi Smart Navigation valet retrieval Steering Highway (Autonomous (Vehicle deciding autonomy) on routes) systems. Emergency parallel parking) steering Steer Assist 2010 ACC 2016 Intersection assist (powered by V 2 V V 2 X) Platooning Co-operative cruise control Assisted Cooperative driving 2020 Driverless car Occupant specific driving dynamics, alerts to the environment (horn, dim dip) Automated vehicle taking complete control of navigation, transmission, steering, braking and parking. Towards fully-automated driving Source: Frost & Sullivan analysis. 7

OEM Comparative Analysis While traditional OEMs are focussed on offering semi-automated features in the near future, Google’s to create an ecosystem dedicated to automated vehicles including retro-fit hardware High inclination towards going “driverless” Google Nissan Key Market for overall vehicle sales: US Volvo General Motors Ford Mercedes. Benz Capability needs improvement Highly capable Key Market for overall vehicle sales: global BMW Audi Fiat Renault VW Lexus Low inclination towards going “driverless” Note: capability is evaluated based on a weighted analysis of current vehicle line-up and demonstrator projects already showcased Source: Frost & Sullivan analysis. 8

OEM Comparative Analysis – Automated Driving Cost & Volume Analysis Ford is likely to lead the market offering automated modes of driving at a disruptive price-range, despite not having firm launch plans; Audi lags behind BMW & Mercedes-Benz who have very competently priced their offering Comparative Estimates of ADAS Option Cost with Automated Driving Option, US, 2014 10000 Google 4000 2800 Volvo 3500 2400 Lexus 2500 1700 Ford GM 3220 Audi 5000 6600 2300 MB 1920 BMW 4000 2600 0 4500 5000 AD Package 10000 ADAS Package Price ($) 15000 Source: Frost & Sullivan analysis. 9

Google Car: Sensors and Technologies Used The Google X automation module employs some of the finest and expensive sensors available today, enabling a transition to ADAS-enabled maps with accuracy of 10 cm, a dire need for autonomous driving 360 Degree Laser detection (LIDAR) GPS helps navigation of the vehicle and also transmits vehicle data for the purpose of diagnostics & prognostics Lidar senses the driving environment around, including tracing lane markings and road-width Stereo Camera (HD) A Stereo- or a Mono-vision camera forward-looking functions such as traffic sign recognition, traffic signal recognition, peer vehicle movement & trajectory projection, besides pedestrian and obstacle detection. RADAR Sensor 3 RADAR Sensors (Bow-mounted stereoscopic sensors) Forward sensing radar monitors the distance of the vehicles and obstacles in front Position Estimator Data from various sensors and vision inputs are fed into a central control unit that controls vehicle dynamics module and drive commands. Algorithm to be sophisticated to understand driving environment’s direct and indirect messages in terms of traffic and maneuvering Ultrasonic sensors may be used to measure the position of objects very close to the vehicle, such as curbs and other vehicles when parking Source: Frost & Sullivan analysis. 10

Business Models Around Automated Driving Besides the obvious ownership-based model, a new subscription-based driving-as-a-service business model can allow for autonomous driving, reinterpreting the human-availability clause Traditional OEM Approach Driving-as-a-service Model Manually driven • Every trim to have standard fitment of DAS warning functions & relevant hardware Basic DAS • This helps OEMs and suppliers to achieve economies of scale for enabling technologies for automated driving. Standard: LDW, BSD, Park assist, FCW, Rear vision camera and driver-monitoring • Every trim gets automated driving as an option, pursuing a pull-strategy. • Automated mode shall be activated only upon receiving a command from the driver Extended ADAS • Driver controls the vehicle when needed. Option: Lane Change / Keeping Assist, Autonomous Parking, Traffic Jam Assist, highway chauffeur and driver-monitoring Simulator Driverless Tele-operated vehicle control • Legislation around driver controlling a vehicle is open for interpretation, allowing for a potentially new business model on a subscription basis. • Vehicles can be tele-operated through a remotely located human driver. The driver controls the vehicle when needed, but monitors it constantly just as any human would. • Subscribers can hop-on and hop-off while a virtual driver always ensures the safety of the vehicle in the driving environment. Stakeholders involved: Legislators, Road-operators, telecom & telematics service providers, insurance providers, V 2 X service providers, mobility integrator etc. Source: Frost & Sullivan analysis 11

® Park 4 U Valeo – The solution & challenges addressed Intuitive Driving Valet Park 4 U® Harald Barth Product Marketing Manager Driving Assistance Nov 2013 I 12

Valeo Park 4 U® – The Solution Control Unit Push button Sensors Affordable luxury for everyday’s life 78 vehicle models from 18 brands From Semi- to Fully-automated Parking The future: Park 4 U® Remote / Valet Park 4 U® Source: Valeo Nov 2013 I 13

Valeo Park 4 U® – The Challenges addressed In a customer clinic over 70% of participants stated a strong interest to purchase Park 4 U® with their next car. Determine the vehicle’s movement / position End-user perception / acceptance System’s understanding of the situation Embed software code into a small µ-processor Source: Valeo Nov 2013 I 14

Meeting the Design Challenges of Autonomous Vehicles Balancing a high degree of system authority with a complex operating environment and security aspects requires a ‘System of systems’ approach MIRA’s slide can only be viewed by watching the analyst briefing. Click the below link to access the on-demand recording. http: //www. frost. com/prod/servlet/analyst-briefing-detail. pag? mode=open&sid=286359044

The Need for System Resilience A paradigm change in the design of highly automated vehicle systems requires a “system of systems” approach to design and assurance MIRA’s slide can only be viewed by watching the analyst briefing. Click the below link to access the on-demand recording. http: //www. frost. com/prod/servlet/analyst-briefing-detail. pag? mode=open&sid=286359044

Requirements for the Design Process of Autonomous Vehicles To implement the resilience approach, a development process incorporating disruption management is required MIRA’s slide can only be viewed by watching the analyst briefing. Click the below link to access the on-demand recording. http: //www. frost. com/prod/servlet/analyst-briefing-detail. pag? mode=open&sid=286359044

Systems Engineering Framework for Autonomous Vehicles Based on a traditional automotive V-model lifecycle process, additional parallel stages are needed to cope with disruptions MIRA’s slide can only be viewed by watching the analyst briefing. Click the below link to access the on-demand recording. http: //www. frost. com/prod/servlet/analyst-briefing-detail. pag? mode=open&sid=286359044

Towards Autonomous Driving Strategic Review and Assessment of Future Opportunities and Implications for the Automotive Industry with the Advent of Autonomous Driving Workshop Proposal November 2013

Follow Frost & Sullivan on Facebook, Linked. In, Slide. Share, and Twitter http: //www. facebook. com/Frostand. Sullivan http: //www. linkedin. com/groups? gid=4480787&trk=hb_side_g http: //www. slideshare. net/Frostand. Sullivan www. twitter. com/FS_Automotive 20

For Additional Information Prana Natarajan Team Leader – Chassis &Safety Automotive & Transportation +44 (0) 20 7915 7871 pranan@frost. com Nick Ford Consultant Automotive & Transportation +44 (0) 1454 880096 Nick. Ford@frost. com Harald Barth Product Marketing Manager Driving Assistance Mobile: +49(0)176 -3000 4025 harald. barth@valeo. com Dr. Anthony Baxendale Research Manager Future Transport Technologies +44 (0) 24 7635 5559 Anthony. baxendale@mira. co. uk 21