Bayerische Motoren Werke Ag · XETR:BMW

Bayerische Motoren Werke — BMW Group — Barclays Investor Conference, London 2015

14 slides from Bayerische Motoren Werke's conference presentation, Sep 4, 2015

  1. Slide 1: Highly Automated Driving

    Dr. René Grosspietsch, Head of Driver Assistance Systems. BMW Group Research and Technology. HIGHLY AUTOMATED DRIVING. BARCLAYS INVESTOR CONFERENCE, LONDON. SEPTEMBER 4TH, 2015.

  2. Slide 2: Highly automated driving will improve driving comfort and still admit driving pleasure

    HIGHLY AUTOMATED DRIVING WILL IMPROVE DRIVING COMFORT AND STILL ADMIT DRIVING PLEASURE. • Is HAD consistent with BMW‘s aspiration to provide customers with sheer driving pleasure? • Yes! Progressing urbanization and increasing car densities in cities lead to traffic congestion on an unprecedented sc

  3. Slide 3: From research to reality: long-time experience in vehicle automation

    FROM RESEARCH TO REALITY: THE BMW GROUP HAS LONG-TIME EXPERIENCE IN VEHICLE AUTOMATION. BMW Track Trainer Emergency stop Highly automated driving (2006) assistant (2009) on the motorway (2011) HAD on the vehicle‘s limit 360° collision avoidance Fully automated remote (2014) (2015) valet parking (201

  4. Slide 4: New 7 Series already provides functions very close to highly automated driving

    NEW 7 SERIES ALREADY PROVIDES FUNCTIONS VERY CLOSE TO HIGHLY AUTOMATED DRIVING. Active Side Collision Protection and Cross Traffic Alert. Pedestrian Protection & Night Vision. Traffic Jam Assistant up to 60km/h Remote Controlled Parking Highly automated driving, Dr. René Grosspietsch, 04.09.2015. Pa

  5. Slide 5: True automated driving still means a big step because of responsibility switch

    HOWEVER TRUE AUTOMATED DRIVING STILL MEANS A BIG STEP BECAUSE OF RESPONSIBILITY SWITCH. New 7 series Man Switch of Responsibility Machine Driver Readiness No tasks / “Brain-off” No driver task to take over / “Eyes-off” Autonomous Attention to traffic / “Hands-off” Full control Vertical or lateral co

  6. Slide 6: Many technologies to be mastered on the way to highly automated driving

    MANY TECHNOLOGIES TO BE MASTERED ON THE WAY TO HIGHLY AUTOMATED DRIVING. Validation / Testing System System itself BACKEND: DIGITAL SERVICES AND MAP SENSORS ENVIRONMENT MODEL E/E ARCHITECTURE & CONTROL UNITS Fusion AND OBJECTS and extraction FREE SPACE SELF-LOCALIZATION MOTION CONTROL HAF DRIVING FU

  7. Slide 7: Many technologies to be mastered on the way to highly automated driving

    MANY TECHNOLOGIES TO BE MASTERED ON THE WAY TO HIGHLY AUTOMATED DRIVING. Validation / Testing System System itself BACKEND: DIGITAL SERVICES AND MAP SENSORS ENVIRONMENT MODEL E/E ARCHITECTURE & CONTROL UNITS Fusion AND OBJECTS and extraction FREE SPACE SELF-LOCALIZATION MOTION CONTROL HAF DRIVING FU

  8. Slide 8: Backend: HD map with dynamic information will provide live road preview

    BACKEND: HD MAP WITH DYNAMIC INFORMATION WILL PROVIDE LIVE ROAD PREVIEW AS ESSENTIAL DIGITAL SERVICES. VARIABLE HIGH- EN-ROUTE TRAFFIC ROAD WEATHER TRAFFIC PRECISION APPROVAL CONDITIONS WORKS CONDITIONS Cooperative ITS, SIGNS DIGITAL MAPS 3rd Party Info > 200 m Far Range Sensors ~ 200 m Short/Mid-Ra

  9. Slide 9: Sensors: price level and package of current lidar sensors is a challenge

    SENSORS: PRICE LEVEL AND PACKAGE OF CURRENT HIGH PERFORMANCE LIDAR SENSORS IS A CHALLENGE. Example: Velodyne 64-beam laser – rotates 360-degrees and takes up to 1.3 million readings per second, making it a truly versatile sensor on the car. Mounting it on top of the car ensures its view isn’t obstru

  10. Slide 10: Cost efficient alternatives on the horizon and expected to become mass market compatible

    BUT COST EFFICIENT ALTERNATIVES ALREADY ON THE HORIZON AND EXPECTED TO BECOME MASS MARKET COMPATIBLE AS IN THE PAST. Parts costs* Lidar 600 HC 400 2 Mechanical, Front radar Sample volume 200 Mechanical 0 Cameras Mass volume Antenna array Volume Solid-state 2006 2015 2021 *actual cost degression base

  11. Slide 11: Onboard sensor fusion generates a real-time environmental model

    ONBOARD SENSOR FUSION GENERATES A REAL-TIME ENVIRONMENTAL MODEL OF VEHICLE SURROUNDINGS. ENVIRONMENTAL MODELLING BASED ON GRIDS Highly automated driving, Dr. René Grosspietsch, 04.09.2015. Page 11

  12. Slide 12: Perception: machine learning seen as an essential building block

    PERCEPTION: FOR BOTH SENSING OF THE ENVIRONMENT AND CONTEXTUAL UNDERSTANDING MACHINE LEARNING SEEN AS AN ESSENTIAL BUILDING BLOCK. Green light will last for the next 10sek Object-based models 30 Zone v = 3 m/s Occupancy maps/free space v = 0 km/h Oppo- v = 7,5m/s site lane v = 7m/s v = 0,5m/s ttc

  13. Slide 13: Remaining challenges: industrialisation, technological, social and legal aspects

    REMAINING CHALLENGES COMPRISE INDUSTRIALISATION AS WELL AS TECHNOLOGICAL, SOCIAL AND LEGAL ASPECTS. Challenges Description • Field operational tests with a heterogeneous fleet to gather all kinds of situational Industrialization functional validation data for a later . • Standardization of methods f

  14. Slide 14: Thank you for your attention!

    THANK YOU FOR YOUR ATTENTION!

BMW Group — Barclays Investor Conference, London 2015 on 6ix