www.osram.com LiDAR Teach-In June 20, 2018 Munich Light is OSRAM
Agenda Introduction Autonomous driving LIDAR technology deep-dive LiDAR@OS: Emitter technologies Outlook LiDAR Tech Teach-In June 20, 2018 2
OSRAM has extensive autonomous driving expertise Gesture recognition Interior lighting Ambient lighting Adaptive cruise control system (LIDAR) Pre crash sensing (LIDAR) Rain sensor Rear light LED/Laser headup display Headlamp (e.g. laser, matrix) Ambient light sensor Driver monitoring Side view and lane detection 3-D environment detection (LIDAR) Daytime running light LED Visible LED Non visible LiDAR Tech Teach-In June 20, 2018 3
OS is a long-time partner for lasers in the automotive industry First use of LiDAR since early 2000s Short-range-LiDAR Cam System OS with pioneer role and a proven track record LiDAR Laser supplier to automotive industry for over 10 years over 10 million LiDAR lasers in the field (~ 200 billion km ) without chip failures zero field failures for 7 million bare dies more than 20 LiDAR design-ins and -wins with OS lasers 905 nm pulse Laser, 75 W, 20 ns More than 10 car OEMs use laser based AEB 1 systems with OSRAM s 905nm laser already today Technology USP Highest Efficiency (30%) Best thermal resistance First automotive grade pulse laser on the market 1) AEB: Auto emergency breaking LiDAR Tech Teach-In June 20, 2018 4
Agenda Introduction Autonomous driving LIDAR technology deep-dive LiDAR@OS: Emitter technologies Outlook LiDAR Tech Teach-In June 20, 2018 5
There are 6 levels of autonomous driving, between no automation (L0) up to full control by system (L5) Humans always monitor driving functions Humans do not always monitor driving functions L0 Driver Only L1 L2 L3 L4 L5 Driving Assistance Partial Automation Conditioned Automation High Automation System Only Role of driver Eyes + hands on Eyes on + hands temporary off Eyes + hands temporary off Eyes + hands off Eyes + hands + mind off Driver always with full control Driver always in charge of monitoring Driver needs to be ready to regain control Driver only needs to act in specific driving conditions No driver needed Role of system None Some automation by electronics Driving assistance in specific and limited situations Perform driving functions and recognize limits Full control in defined driving conditions Full control LiDAR Tech Teach-In June 20, 2018 6
All of today s potential sensor solutions for ADAS have specific strengths and weaknesses; hence all are needed to guarantee full functionality Camera Radar LIDAR Traffic Sign recognition Emergengy braking Pedestrian detection Collision avoidance Environment mapping Lane departure warning Cross traffic alert Environment Environment Mapping Mapping Surround view Environment Mapping Surround view Blind spot detection Environment Mapping Rear collision warning Environment Mapping Surround view Value propositions of different technologies (selection) Camera Radar LiDAR Color vision and ability to recognise signs Speed detection and ability to function in unfavorable weather conditions / darkness 3D capability and ability to function in unfavorable weather conditions / darkness The vehicles ability to see (= sense its environment) strongly depends on the environmental conditions, such as day time or weather Full functionality at any environmental condition therefore requires different and redundant sensors In addition, redundancy is being generated using intelligent algorithms (sensor fusion) LiDAR Tech Teach-In June 20, 2018 7
Only the combination of Camera, Radar and LiDAR will provide the capabilities required for autonomous driving. Inherent technical capabilities Range Resolution / Accuracy 3D capability Detection of speed of objects Read signs and see colors Environmental conditions Immunity to weather (rain / fog / snow) Works in dark Works in very bright light Other Interference effects Price Size Detector systems Camera Radar LiDAR Status today Status today Combination of all three systems Only the combination of all three systems will provide all technical capabilities needed for autonomous driving In addition, redundancies are desired to increase accuracy and mitigate risk through potential technical failures LiDAR Tech Teach-In June 20, 2018 8
LiDAR will be required for autonomous driving; 6 or more LiDAR modules for L4/L5 Number of modules needed by automation level Radar Camera Level of automation L2 L3 L4/5 3 6 10 1 2 8 Co-existence of LIDAR, RADAR and camera sensors in Autonomous Driving (L3 upwards) LIDAR crucial for long-range detection in AD, no different substitute technology available, specifically for object identification Six or more LiDAR modules expected for L4/L5 LiDAR none 1 6 1 LiDAR will play a key role in autonomous driving (especially L3 upwards) 1) Optional for collision warning, AEB LiDAR Tech Teach-In June 20, 2018 9
Two different approaches to autonomous driving: Mobility providers starting with L5 short term, traditional OEMs evolve from L1 to L5 ADAS 1 Level ILLUSTRATIVE L5 MaaS - Mobility as a Service path (e.g. waymo) Robo cars Convergence Two completely different approaches to autonomous driving with different dynamics driven by two distinct groups of players mobility-as-a-service providers (MaaS) and traditional car manufacturers MaaS providers pushing for early introduction of fully autonomous L3 vehicles today in significant numbers Traditional car manufacturers moving to L3; introduction of fully autonomous cars expected for 2024+ L1-2 However, in the long run there will likely be a convergence of the two approaches 2000 2018 2020 2022 2024 2026 1) Advanced driver assistance systems LiDAR Tech Teach-In June 20, 2018 10 10
The LiDAR (emitter) market will see two waves of growth LiDAR emitter market total number of laser channels Traditional AEB 2 Existing LiDAR business since early 2000s Collision mitigation system Traditional OEMs Short range FLASH LiDAR ILLUSTRATIVE MaaS 1 : Short- and medium-term volume driver with early introduction of L5 cars Consumer-like business with steep ramps, driven by Tec companies AM qualification less important Long-range and short-range scanning LiDAR; large field of view, high resolution Traditional L3-L5: Long-term volume driver: Evolution of traditional car mass production from L3 to L5 Mainly traditional OEMs, quality and AM qualification very important Long-range scanning or Solid State LiDAR, short range FLASH 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 1) MaaS: Mobility as a Service 2) AEB: Auto emergency breaking LiDAR Tech Teach-In June 20, 2018 11
Agenda Introduction Autonomous driving LIDAR technology deep-dive LiDAR@OS: Emitter technologies Outlook LiDAR Tech Teach-In June 20, 2018 12
LiDAR uses reflected light to measure distance to objects; it can be used for short-range up to a distance of ~300m LiDAR (Light Detection and Ranging) a Time of Flight measurement Detection principle: A very short laser pulse travels from the LiDAR sensor to an object and back. The sensor measures the travelling time of the laser pulse and determines the distance and relative velocity of the object. 100 m distance 0,6 ms LiDAR Tech Teach-In June 20, 2018 13
Multiple module system to dominate combination of short / mid range and long range LIDARs Depending on L-level, different LiDAR systems are used Lidars can be used for different distances and use cases # SRL 1 # LRL 2 Short range (<20m) Stop & Go Automatic Emergency Breaking (AEB) Pre-crash sensing L1, L2 optional for collision warning, AEB 20m L3 Medium range (20 to 100m) Lane departure warning Blind spot detection 0-2 optional 0-1 100m L4 2-4 SRL for 360 sensing 1 LRL to the front Long range (>100m) Adaptive cruise control (ACC) Long Range LiDAR (LRL) for autonomous vehicles (L4/L5) L5 2-4 SRL for 360 sensing 1-2 rear LRL for fast on coming vehicles LiDAR Tech Teach-In June 20, 2018 14 1) Short range LIDAR 2) Long range LIDAR
A LiDAR system consists of different building blocks LiDAR system overview PROCESSING & CONTROL Data visualization or interpretation, analysis, filter, localization and control signals EMITTER Laser Illuminate the scene in front L e n s Key steps for signal acquisition The emitter is triggered by the ASIC (application specific integrated circuit) and sends out invisible light The light is reflected from objects in the field of view and the signal captured by the detector ASIC 1 Process captured signal and feedback time stamps to emitter DETECTOR Photo Detector Capture the reflected photons L e n s The signal is then processed by the ASIC 1 and forwarded to the processing unit Here, analysis of the signal takes place, potentially taking signals from other sensors (e.g. radar or camera) into account 1) Application specific integrated circuit LiDAR Tech Teach-In June 20, 2018 15
A LiDAR system consists of different building blocks LiDAR system overview PROCESSING & CONTROL Data visualization or interpretation, analysis, filter, localization and control signals EMITTER Focus today Laser Illuminate the scene in front L e n s Key steps for signal acquisition The emitter is triggered by the ASIC (application specific integrated circuit) and sends out invisible light The light is reflected from objects in the field of view and the signal captured by the detector ASIC 1 Process captured signal and feedback time stamps to emitter DETECTOR Photo Detector Capture the reflected photons L e n s The signal is then processed by the ASIC 1 and forwarded to the processing unit Here, analysis of the signal takes place, potentially taking signals from other sensors (e.g. radar or camera) into account 1) Application specific integrated circuit LiDAR Tech Teach-In June 20, 2018 16
There are two basic LiDAR technical implementations flash and scanning LiDAR Flash LiDAR EMITTER High to very high power needed to illuminate the whole scene with one laser beam In the future, laser bars (edge emitting laser) with high power for longer distances or potential use of VCSEL for SRL Scanning LiDAR EMITTER Highly directional beam with high power needed Good thermal performance for high repetition rate Today and in future edge emitting lasers operating at 905nm expected to dominate The whole FOV is illuminated at once using a wide-angle beam No moving parts in the LiDAR module Scanning, narrow emitter beam which is being moved across the FOV over time Mechanical solution or micro-mirrors used for beam steering LiDAR Tech Teach-In June 20, 2018 17
Agenda Introduction Autonomous driving LIDAR technology deep-dive LiDAR@OS: Emitter technologies Outlook LiDAR Tech Teach-In June 20, 2018 18
Edge emitting lasers at 905nm are the best overall fit for LiDAR requirements in terms of cost position, efficiency and output Comparison of basic laser technology and wavelength Laser technology VCSEL Edge emitting laser Fit to LiDAR requirements Output power Efficiency Temp. stability 1) Component cost Eye safety Emitting area Light emitted from surface of laser die typically <1000nm Light emitted from a small area in the side of the laser die <1000nm 2) >1000nm In most cases, edge emitting laser @905nm expected to be the winning technology 1550nm with better eye safety, but otherwise lagging behind and too expensive for customer requirements VCSEL potentially useful for short range LiDAR, but yet to be automotive qualified Conclusion Potential use for short range LiDAR Best overall fit to requirements Technology too expensive 1) With regard to power 2) typically 905nm LiDAR Tech Teach-In June 20, 2018 19
Osram: we offer solutions for all relevant LiDAR technologies based on 905 nm Lasers SELECTIVE Evolution of 905nm edge emitting laser Today 2018/19 OS product roadmap for different LiDAR technologies FLASH Scanning First AM grade pulse laser on market Best performance in the market Very short pulses (1ns demonstrated) Smart Laser Increased output power More laser channels AM qualified packages Beyond Customization Integrated solutions Multichannel versions / laser bars 4 channel SMT Highest Power: Laser Bars Chip- and package variations based on OS Platform OSRAM is in discussion with all relevant players from the LiDAR ecosystem and we further develop our portfolio based on their needs Mechanical bare laser dies 1 or 4 channel SMT Integrated versions MEMS / Solid state 1 channel SMT Customized laser bars LiDAR Tech Teach-In June 20, 2018 20
Agenda Introduction Autonomous driving LIDAR technology deep-dive LiDAR@OS: Emitter technologies Outlook LiDAR Tech Teach-In June 20, 2018 21
OSRAM is well positioned to participate in the LiDAR market Next to our strong position in emitter lasers for LiDAR systems, we are evaluating to extend our portfolio in the area of detector components and smart modules, concentrating on the essentials of data fusion: working closely together with LeddarTech in order to assess the industrialization of LiDAR modules for the automotive mass market. Specific focus is on affordable long-range LiDAR modules acquired a strategic stake in the LiDAR start-up company Blickfeld. Its particular strengths lie in beam steering technology, which permits a full scanning of the surrounding environment smaller, purely financial investment in the start-up company Tetravue, which uses a unique approach for high-resolution depth images Strong position in emitter lasers Signal processing Optics and beam control High resolution depth images Evaluating portfolio fit of detectors Additional application fields beyond automotive LiDAR Tech Teach-In June 20, 2018 22