1. | EXECUTIVE SUMMARY AND CONCLUSIONS |
1.1. | Purpose of this report |
1.2. | SAE levels of automation in land vehicles |
1.3. | Thirteen primary conclusions |
1.3.1. | The dream and the basics for getting there |
1.3.2. | Specification of a robot shuttle |
1.3.3. | Very different from a robotaxi |
1.3.4. | Smart shuttles will address megatrends in society |
1.3.5. | Robot shuttle business cases from bans and subsidies |
1.3.6. | Robot shuttle business cases from exceptional penetration of locations |
1.3.7. | Intensive use business cases are compelling |
1.3.8. | Campuses are not a quick win |
1.3.9. | The robot shuttle opportunity cannot be addressed by adapting existing vehicles |
1.3.10. | The leaders so far |
1.3.11. | Upfront cost and other impediments |
1.3.12. | Dramatic technical improvements are coming |
1.3.13. | Autonomous vehicle use of 5G and 6G communications |
1.4. | Two generations of robot shuttle |
1.4.1. | Envisaged applications compared |
1.4.2. | Second generation robot shuttle 2025-2040 |
1.5. | Robot shuttles: the good things |
1.5.1. | Many benefits |
1.5.2. | Building on the multi-purposing of the past |
1.5.3. | Heart of the ideal robot shuttle |
1.5.4. | Archetypal large robot shuttles |
1.5.5. | Archetypal small robot shuttles |
1.6. | Robot shuttles: the bad things |
1.7. | Analysis of 39 robot shuttles and their dreams |
1.8. | Geographical, size, deployment distribution of 39 robot shuttles |
1.8.1. | Manufacture by country |
1.8.2. | Manufacture by major region |
1.8.3. | Designs by size |
1.8.4. | Number of robot shuttles deployed globally by manufacturer 2021 |
1.9. | Timelines and forecasts |
1.9.1. | Technology and launch roadmap 2020-2041 |
1.9.2. | Predicting when the robot shuttle has lower up-front price than a legal diesel midibus 2020-2041 |
1.9.3. | Hype 2018-2041 |
1.9.4. | Robot shuttles total market size in unit numbers thousand 2019-2041 |
1.9.5. | Robot shuttles total market size in US$ million 2019-2041 |
2. | INTRODUCTION |
2.1. | Bus and robot shuttle types compared |
2.2. | Pure electric buses for lowest TCO |
2.3. | Peak car coming: global passenger car sales forecast 2020-2040 - moderate scenario (unit numbers) |
2.4. | Background to robot shuttles |
2.5. | Tough for robot shuttles to compete |
2.6. | Second generation robot shuttles |
2.7. | Trials in Japan |
2.8. | Schaeffler: mechanically repurposed shuttle |
2.9. | Einride Sweden: not quite a robot shuttle |
2.10. | Rinspeed dreams embrace robot shuttles |
3. | ROBOT SHUTTLES IN ACTION - 37 TYPES IN 15 COUNTRIES |
3.1. | 2getthere Netherlands |
3.2. | ANA collaboration Japan |
3.3. | Apollo Apolong: Baidu King Long China |
3.4. | Apple VWT6 USA |
3.5. | Astar Golden Dragon China |
3.6. | Aurrigo UK |
3.7. | AUVE Tech. Estonia |
3.8. | BlueSG/ Nanyang France Singapore |
3.9. | Capri AECOM UK |
3.10. | Coast Autonomous |
3.11. | Cruise Origin USA |
3.12. | DeLijn Belgium |
3.13. | e-BiGO Dubai |
3.14. | eGo Mover Germany |
3.15. | E-Palette Toyota |
3.16. | EZ10 EasyMile France |
3.17. | GACHA Sensible4 Finland |
3.18. | Hino Poncho SB Drive Japan |
3.19. | IAV HEAT Germany |
3.20. | iCristal Torc Robotics USA |
3.21. | KAMAZ shuttles Russia |
3.22. | KTI Hyundai Korea |
3.23. | LG Korea |
3.24. | Myla: May Mobility USA |
3.25. | Navya France |
3.26. | NEVS Sweden/ China |
3.27. | Ohmio Automation New Zealand |
3.28. | Olli: Local Motors USA |
3.29. | Optimus Ride USA |
3.30. | Ridecell Auro USA |
3.31. | Scania NXT - a second generation robot shuttle Sweden |
3.32. | Sedric Germany |
3.33. | ST Engineering Land Systems Singapore |
3.34. | Tony: Perrone Robotics USA |
3.35. | Volkswagen ID Buzz Germany |
3.36. | Yutong Xiaoyu China |
3.37. | Zoox USA |
4. | TOOLKIT FOR NEW EARNING STREAMS FROM NEW TECHNOLOGIES IN SECOND GENERATION |
4.1. | Challenges being addressed |
4.2. | How eight key enabling technologies for robot shuttles are improving to serve 10 primary needs |
4.3. | How to reduce diesel shuttle parts by 90% with advanced electrics |
4.4. | Future electric vehicle powertrains - relevance to robot shuttles |
4.5. | Platform evolution |
4.5.1. | Overview |
4.5.2. | Toyota REE chassis: huge advances |
4.6. | Voltage trends |
4.7. | Electric motors |
4.7.1. | Overview |
4.7.2. | Synchronous or asynchronous |
4.7.3. | Operating principles for most EV uses |
4.7.4. | Electric motor choices for robot shuttles and their current EV uses |
4.7.5. | Electric motors for pure electric cars, vans: lessons for shuttle buses |
4.7.6. | Company experience and designer preferences |
4.7.7. | Motor material cost trends spell trouble |
4.8. | In-wheel motors |
4.9. | Sideways steerable wheels |
4.10. | 360 degree wheels with in-wheel motor: Protean and Productiv |
4.11. | Energy storage for pure electric buses |
4.11.1. | Conventional buses see batteries shrink |
4.11.2. | Robot shuttles stay battery hungry |
4.11.3. | Even better batteries and supercapacitors a real prospect: future W/kg vs Wh/kg |
4.11.4. | Location and protection of batteries |
4.11.5. | Bus battery type, performance, future for 31 manufacturers |
4.12. | Charger standardisation: bus/truck commonality |
4.13. | Energy Independent Electric Vehicles EIEV |
4.14. | Stella Vie showing the way to an energy positive robot shuttle? |
5. | ENABLING TECHNOLOGIES: LIDARS, RADARS, CAMERAS, AI SOFTWARE AND COMPUTING PLATFORM, HD MAP, TELEOPERATION, CYBERSECURITY, 5G AND V2X |
5.1. | Chess pieces: autonomous driving tasks |
5.2. | Typical toolkit for autonomous cars |
5.3. | Anatomy of an autonomous car |
5.4. | Evolution of sensor suite from Level 1 to Level 5 |
5.5. | What is sensor fusion? |
6. | LIDARS |
6.1. | 3D Lidar: market segments & applications |
6.2. | 3D Lidar: four important technology choices |
6.3. | Comparison of Lidar, Radar, Camera & Ultrasonic sensors |
6.4. | Automotive Lidar: SWOT analysis |
6.5. | Automotive Lidar: operating process & requirements |
6.6. | Emerging technology trends |
6.7. | Comparison of TOF & FMCW Lidar |
6.8. | Laser technology choices |
6.9. | Comparison of common Laser type & wavelength options |
6.10. | Beam steering technology choices |
6.11. | Comparison of common beam steering options |
6.12. | Photodetector technology choices |
6.13. | Comparison of common photodetectors & materials |
6.14. | 106 Lidar players by geography |
6.15. | Lidar hardware supply chain for L3+ vehicles |
6.16. | Beam steering technology |
6.17. | Mechanical Lidar players, rotating & non-rotating |
6.18. | Micromechanical Lidar players, MEMS & other |
6.19. | Pure solid-state Lidar players, OPA & liquid crystal |
6.20. | Pure solid-state Lidar players, 3D flash |
6.21. | Players by technology & funding secured |
6.22. | Lidars per vehicle by technology & common configurations |
6.23. | Lidar configuration diagrams: L3, L4 & L5 vehicles |
6.24. | Average Lidar cost per vehicle by technology |
6.25. | L3 private vehicle market share by Lidar technology |
6.26. | L4 & L5 private vehicle market share by Lidar technology |
6.27. | L4 & L5 shared mobility market share by Lidar technology |
6.28. | Global Lidar unit sales by L3+ vehicle type |
6.29. | Global Lidar market size by L3+ vehicle type |
6.30. | Global Lidar unit sales by technology |
6.31. | Global Lidar market size by technology |
7. | RADARS |
7.1. | Towards ADAS and autonomous driving: increasing sensor content |
7.2. | Towards ADAS and autonomous driving: increasing radar use |
7.3. | SRR, MRR and LRR: different functions |
7.4. | The evolving role of the automotive radar towards full 360 degree imaging |
7.5. | Automotive radars: role of legislation in driving the market |
7.6. | Automotive radars: frequency trends |
7.7. | Radar: which parameters limit the achievable KPIs |
7.8. | Impact of frequency and bandwidth on angular resolution |
7.9. | Why are radars essential to ADAS and autonomy? |
7.10. | Towards autonomy: Increasing semiconductor use |
7.11. | Performance levels of existing automotive radars |
7.12. | Radar players and market share |
7.13. | Radar market forecasts 2020-2040 in all levels of autonomy/ADAS in vehicles and trucks (unit numbers) |
7.14. | Radar market forecasts 2020-2040 in all levels of autonomy/ADAS in vehicles and trucks (market value) |
7.15. | Radar market forecasts 2020-2040 in all levels of autonomy/ ADAS in vehicles and trucks (market value) - moderate |
7.16. | Radar market forecasts 2020-2040 in all levels of autonomy/ ADAS in vehicles and trucks (market value) - aggressive |
7.17. | Radar semiconductor market share forecast (GaAs, SiGe, Si) |
7.18. | Ten year (unit number) market forecasts for automotive radars |
7.19. | Benchmarking of semiconductor technologies for mmwave radars |
7.20. | The choice of the semiconductor technology |
7.21. | SiGe: current and emerging performance levels |
7.22. | SiGe: overview and comparison of manufacturers |
7.23. | SiGe BiCMOS: Infineon Technology |
7.24. | SiGe BiCMOS: NXP |
7.25. | SiGe BiCMOS: ST Microelectronics |
7.26. | A closer look at SiGe vs Si CMOS |
7.27. | Emerging all Si CMOS radar IC packages: NXP |
7.28. | Emerging all Si CMOS radar IC packages: ADI |
7.29. | Emerging all Si CMOS radar IC packages: TI |
7.30. | Many chip makers are on-board |
7.31. | Packaging trends: from discreet bare die (COB) to wafer-level packaging and beyond? |
7.32. | Packaging trends: AiP goes commercial? |
7.33. | Packaging trends: from discreet bare die (COB) to wafer-level packaging and beyond? |
7.34. | Comparison of die vs packaged options |
7.35. | eWLP vs flip chip and BGA in terms of insertion loss |
7.36. | Radar packaging: Material opportunities |
7.37. | Glass and panel level packaging of radars? |
7.38. | Function integration trend: from discreet to full chip-level function integration |
7.39. | Function integration trends: towards true radar-in-a-chip |
7.40. | Evolution of radar chips towards all-in-one designs |
7.41. | Evolution of radar chips: all-in-one designs |
7.42. | Board trends: from separate RF board to hybrid to full package integration? |
7.43. | Hybrid board is the norm |
7.44. | Hybrid board: what is it |
8. | CAMERAS |
8.1. | How many camera needed in various levels of autonomy |
8.2. | CMOS image sensors vs CCD cameras |
8.3. | Key components in a CMOS image sensor (CIS) |
8.4. | Front vs backside illumination |
8.5. | Process flow for back-side-illuminated CMOS image sensors |
8.6. | Global vs Rolling Shutter |
8.7. | Global shutter: pixel size limitation and read-out mechanism |
8.8. | TPSCo: leading foundry for global shutter FSI CMOS on 65nm node |
8.9. | TPSCo: its best-in-class performance and partners |
8.10. | Sony: pixel architecture and performance of FSI global-shutter CMOS |
8.11. | Sony: back-side-illuminated stacked global shutter CMOS (breakthrough?) |
8.12. | Sony: BSI global shutter CMOS with stacked ADC |
8.13. | Omnivision: 2.2um GS CIS for automotive |
8.14. | Hybrid organic-Si global shutter CIS with high-res and low-noise |
8.15. | Hybrid QD-Si GS CIS at IR and NIR: achieving small pixels by physical separation of charge conversion and storage |
8.16. | Why one needs NIR sensing in machine vision |
8.17. | NIR sensing: limitation of Si CMOS |
8.18. | OmniVision: making silicon CMOS sensitive to NIR |
8.19. | Deep trench isolation: innovation to reduce cross-talk |
8.20. | What is SWIR or short-wave-infra-red? |
8.21. | Why SWIR in autonomous mobility |
8.22. | Other SWIR benefits: better animal or on-road hazard detection |
8.23. | SWIR sensitivity of different materials (PbS QDs, Si, polymers, InGaAs, HgCdTe, etc) |
8.24. | SWIR: incumbent and emerging technology options |
8.25. | The challenge of high resolution, low cost IR sensors |
8.26. | Silicon based SWIR sensors: innovation |
8.27. | Why colloidal quantum dots? |
8.28. | Quantum dots: choice of the material system |
8.29. | Advantage of solution processing: ease of integration with ROIC CMOS? |
8.30. | How is the QD layer applied? |
8.31. | Other ongoing challenges |
8.32. | Emberion: QD-graphene SWIR sensor |
8.33. | Emberion: QD-Graphene-Si broadrange SWIR sensor |
8.34. | SWIR Vision Sensors: first QD-Si cameras and/or an alternative to InVisage? |
8.35. | SWIR Vision Sensors: first QD-Si cameras and/or an alternative to InVisage? |
8.36. | QD-ROIC Si-CMOS integration examples (IMEC) |
8.37. | QD-ROIC Si-CMOS integration examples (ICFO) |
9. | AI SOFTWARE AND COMPUTING PLATFORM |
9.1. | Terminologies explained: AI, machine learning, artificial neural networks, deep neural networks |
9.2. | Artificial intelligence: waves of development |
9.3. | Classical method: feature descriptors |
9.4. | Typical image detection deep neutral network |
9.5. | Algorithm training process in a single layer |
9.6. | Towards deep learning by deepening the neutral network |
9.7. | The main varieties of deep learning approaches explained |
9.8. | There is no single AI solution to autonomous driving |
9.9. | Application of AI to autonomous driving |
9.10. | End-to-end deep learning vs classical approach |
9.11. | Imitation learning for trajectory prediction: Valeo (1) |
9.12. | Imitation learning for trajectory prediction: Valeo (2) |
9.13. | Hybrid AI for Level 4/5 automation |
9.14. | Hybrid AI for object tracking |
9.15. | Hybrid AI for sensor fusion |
9.16. | Hybrid AI for motion planning |
9.17. | Autonomous driving requires different validation system |
9.18. | Validation of deep learning system? |
9.19. | The vulnerable road user challenge in city traffic |
9.20. | Multi-layered security needed for vehicle system |
9.21. | Aurora: building the full-stack AD solution |
9.22. | Argo AI: fully integrated AD driving system for OEMs |
9.23. | Drive.ai: AD retrofitting kit |
9.24. | Momenta: the Chinese AD solution provider |
9.25. | Sensor fusion for Mpilot Highway and Parking |
9.26. | HoloMatic: the Xuanyuan platform |
9.27. | The coming flood of data in autonomous vehicles |
9.28. | Computing power needed for autonomous driving |
9.29. | Horizon Robotics: the Chinese embedded AI chip unicorn |
9.30. | The paradigm shift of AI computing |
9.31. | Horizon Robotics: software and hardware roadmap |
9.32. | By-wire and AV domain computer |
9.33. | Waymo open dataset |
9.34. | Pandaset by Hesai and Scale |
9.35. | Oxford radar Robotcar dataset |
9.36. | Astyx Dataset HiRes2019 |
9.37. | Berkeley DeepDrive or BDD100K |
9.38. | Karlsruhe Institute of Technology and Toyota dataset |
9.39. | Cityscapes dataset presented in two 2015 and 2016 papers |
9.40. | Mapillary dataset presented in a 2017 paper |
9.41. | Apolloscape dataset by Baidu |
9.42. | Landmarks and Landmarks v2 by Google |
9.43. | Level 5 dataset by Lyft |
9.44. | nuScenes dataset by Aptiv |
9.45. | Datasets by University of Michigan and Stanford University |
9.46. | Sydney Urban Objects by the University of Sydney |
10. | HIGH-DEFINITION (HD) MAP |
10.1. | Lane models: uses and shortcomings |
10.2. | Localization: absolute vs relative |
10.3. | RTK systems: operation, performance and value chain |
10.4. | Sensors (GPS): price and market adoption (in unit numbers) evolution of GPS sensors |
10.5. | HD mapping assets: from ADAS map to full maps for level-5 autonomy |
10.6. | Many layers of an HD map for autonomous driving |
10.7. | HD map as a service |
10.8. | Who are the players? |
10.9. | Key business model differentiation between HD mapping players |
10.10. | Campines relying on vertical integration to build HD maps (TomTom. AutoNavi, Google, Here Technologies, etc.) |
10.11. | Campines relying on camera to build HD maps (IvI5, Atlatec, Carmera, Mapper) |
10.12. | Campines relying on camera to build HD maps (IvI5, Atlatec, Carmera) |
10.13. | Companies building a map for specific firms: DeepMap |
10.14. | Enabling edge-level calculations |
10.15. | Semi- or fully automating the data-to-map process |
11. | TELEOPERATION |
11.1. | Ottopia's advanced teleoperation for autonomous cars |
11.2. | Features of Ottopia's teleoperation platform |
11.3. | Business model of Ottopia |
11.4. | Phantom Auto's teleoperation as back-up for AVs |
11.5. | Phantom Auto gaining momentum in logistics |
12. | CYBERSECURITY |
12.1. | Cybersecurity risks for autonomous cars |
12.2. | Typical attack surfaces of a CAV |
12.3. | Vulnerable targets for hackers - connected ECUs |
12.4. | 5StarS - consortium for cybersecurity assurance |
12.5. | Arilou's in-vehicle cybersecurity solutions |
12.6. | Argus's multi-layered cybersecurity solutions |
12.7. | TowerSec's intrusion detection and prevention solution |
12.8. | C2A Security's in-vehicle cybersecurity protection |
12.9. | Regulus's cyber defense for GNSS sensors |
13. | 5G, 6G AND V2X FOR ROBOT SHUTTLES |
13.1. | Overview |
13.2. | Why Vehicle-to-everything (V2X) is important for future autonomous vehicles |
13.3. | Two type of V2X technology: Wi-Fi vs cellular |
13.4. | Regulatory: Wi-Fi based vs C-V2X |
13.5. | C-V2X assist the development of smart mobility |
13.6. | How C-V2X can support smart mobility |
13.7. | C-V2X includes two parts: via base station or direct communication |
13.8. | C-V2X via base station: vehicle to network (V2N) |
13.9. | 5G technology enable direct communication for C-V2X |
13.10. | Architecture of C-V2X technology |
13.11. | Use cases and applications of C-V2X overview |
13.12. | C-V2X for automated driving use case |
13.13. | Use cases of 5G NR C-V2X for autonomous driving |
13.14. | Other use cases |
13.15. | Case study: 5G to provide comprehensive view for autonomous driving |
13.16. | Case study: 5G to support HD content and driver assistance system |
13.17. | Timeline for the deployment of C-V2X |
13.18. | Progress so far |
13.19. | Landscape of supply chain |
13.20. | 5G for autonomous vehicle: 5GAA |
13.21. | Ford C-V2X from 2022 |
14. | MICROLED DISPLAYS: INTERNAL/ EXTERNAL VIEWING IN ROBOT SHUTTLES |
14.1. | Existing large mini-/micro-led display announcements |
14.2. | Expectation of future displays |
14.3. | Characteristic comparison of different display technologies |
14.4. | Horizontal comparison |
14.5. | Core value propositions of µLED displays |
14.6. | Micro-LED display types |
14.7. | Micro-LED application roadmap |
14.8. | Emerging displays enabled by micro-LED technology |