《Yole:2023年全球汽车激光雷达市场报告:竞争动态、技术变革和营收趋势(英文版)(28页).pdf》由会员分享,可在线阅读,更多相关《Yole:2023年全球汽车激光雷达市场报告:竞争动态、技术变革和营收趋势(英文版)(28页).pdf(28页珍藏版)》请在三个皮匠报告上搜索。
1、INTELLIGENCE TO SHAPE YOUR TOMORROW|Yole Intelligence 2023Automotive LIDAR Market:Competitive Dynamics,Technology Evolution,and Revenue TrendsAutomotive LiDAR conference Automotive LiDAR conference 2023,October 32023,October 3-5 5 Pierrick Boulay,Senior AnalystPierrick Boulay,Senior Analyst1.Context
2、2.Integration in passenger cars3.Partnerships and supply chain4.Market view5.Partnerships and supply chain6.Technology trends7.ConclusionAGENDA2Automotive LiDAR 2023|Conference|2023CONTEXT3Automotive LiDAR 2023|Conference|2023How automated driving has been achievedThe combination of multiple sensors
3、 is the basis of advanced driving functionalities.Adaptive Cruise ControlLane Keep AssistLKA&LCATraffic-Jam AssistALKS 60 km/hALKS 130 km/hAdaptive Cruise Control with Stop&GoFuture functionalitiesAutomated drivingCONTEXT4Automotive LiDAR 2023|Conference|2023Eyes-off applications:an incremental evol
4、utionEach use case is enabled by a new generation of sensors.A good example is the Valeo Scala 2,which enables applications up to 90 km/h,while the Scala 3 enables applications up to 130 km/h.HighwayHighway 60km/hHighwayHighway 130km/hHighwayHighwayOn-/Off rampArterialArterialJunctions and traffic l
5、ightsRuralRuralAll junctionsUrbanUrbanFull autonomyUse casesEach use case is an incremental evolution enabled by a new generation of hardware(sensors and computing)and software.SensorsComputingSensorsComputingSensorsComputingSensorsComputingSensorsComputingSensorsComputingScala 2Scala 3Example:20222
6、0222024202420262026-2028?2028?20352035-2040?2040?20302030-2035?2035?20402040-2050?2050?CONTEXT5Automotive LiDAR 2023|Conference|2023Necessary hardware for eyes-off applicationsThe complexity of eyes-off applications is following a logarithmic scale.HHANDSANDS-OFFOFFE EYESYES-OFFOFFHW-60 kphHW-130 kp
7、hHW-rampArterialRuralUrbanCameraRadarLiDAR800Tesla promises7 75 51 111 115 53 320206 64 425256 64 4New sensor generationNew sensor generationNew sensor generationNew sensor generationComplexity is following a Complexity is following a logarithmic scalelogarithmic scaleSensorsSensors3Neces
8、sary hardware to reach Necessary hardware to reach advanced eyesadvanced eyes-off applicationsoff applicationsCONTEXT6Automotive LiDAR 2023|Conference|2023OEM choices:Electrification,automated driving,or both?Chinese OEMs have already made the EV transition allowing them to integrate more automated
9、driving features.TimeChinas plan to invest in EV technologies started in the early 2000s.Between 2009 and 2022,the government poured over$29 billion into relevant subsidies and tax breaks to help local OEMs transition from ICE to EV.For a decade now,new OEMs have emerged as pure EV players.After the
10、 EV transition,they are now ready to push for automated driving.European and US OEMs are starting this transition with incentives and regulations pushing the transition from ICE to EV.This means that OEMs will face penalties or be unable to sell cars if they are not pure EVs.So,with the development
11、of automated driving features,most OEMs have to make some choices as the transition will be associated with a high development cost.Few OEMs,like Mercedes,Volvo,and Cadillac,are developing both features,but this is limited to high-end vehicles at a high selling price.CONTEXT7Automotive LiDAR 2023|Co
12、nference|2023Passenger car market:distribution of cars with LiDAR(2018-2022)First cars with LiDAR were in the F-segment.In 2022,several cars with LiDAR are now in the D-segment,with a much lower price.Car segmentCar segmentStarting priceStarting price$100,000$100,000$75,000$75,000$50,000$50,000Sourc
13、e:Car segment classification from MarklinesE-F segments are where ADAS innovations are initially introduced.Then,ADAS systems progressively spread into other segments.INTEGRATION IN PASSENGER CARS8Automotive LiDAR 2023|Conference|2023Evolution of LiDAR positionTo enable highway driving at 130km/h,th
14、ere is a critical need to improve LiDAR performance,especially the detection range.LiDAR positionUse caseLimited to highways at a speed up to 60km/h.No need to see far away.Stopping distance is 36m at 60km/h.Typical LiDAR position:grille or bumperTypical LiDAR position:roof or behind the windshieldL
15、imited to highways at a speed up to 130km/h.Need to see far away(200m).Stopping distance is 170m at 130km/h.Good position to see over the road separation to detect cars on a curve.With higher speeds,increasing the detection range is necessary,as well as the vertical resolution,in order to identify m
16、ore rapidly other cars as well as other objects on the road.This directly impacts the behavior of the car in automated driving mode.INTEGRATION IN PASSENGER CARS9Automotive LiDAR 2023|Conference|2023Evolution of LiDAR position Case of ON/OFF Ramp on highwayEnabling On/Off Ramp use case may require u
17、p to three additional LiDARs in a carLiDAR positionUse caseIn the case of highway driving with on/off ramps,there will be a need to add multiple LiDARs around the vehicle.Short-range LiDARs will be needed on the sides to monitor other road users.The rear LiDAR could be mid-to long-range to monitor o
18、n-coming cars in On-ramps.”Typical long-range LiDAR position:roof or behind the windshield for highway driving.Typical short-range LiDAR position:fender,B-pillar,or roof.On the rear side,the range has still to be defined,as a mid-to long-range LiDAR could be integrated.The LiDAR could be integrated
19、on a high mounting position(roof or rear window)or on the bumper.Nonetheless,a higher mounting position would enable better detection of objects and users.10Automotive LiDAR 2023|Conference|2023INTEGRATION IN PASSENGER CARSNumber of LiDARs per car:robotaxi vs.passenger car comparison#of LiDARs#of Li
20、DARs2026e2026e20020202320232024e2024e2028e2028e2030e2030e2032e2032e1 LiDAR1 LiDAR2 LiDAR2 LiDAR3 LiDAR3 LiDAR4 LiDAR4 LiDAR6 LiDAR6 LiDAR5 LiDAR5 LiDARPassenger car:Passenger car:11-2 LiDAR 2 LiDAR RobotaxiRobotaxi55-6 LiDAR 6 LiDAR Legend:Long-rangeShort-rangePassenger car:Passenger car:
21、3 LiDAR 3 LiDAR RobotaxiRobotaxi55-6 LiDAR 6 LiDAR Eyes-off urbanEyes-off highway$15$15-20k of LiDAR 20k of LiDAR per robotaxiper robotaxi$1,000 per$1,000 per LiDARLiDAR$1,000 of LiDAR$1,000 of LiDAR per carper car$6$6-7k of LiDAR 7k of LiDAR per robotaxiper robotaxi00002500Hei
22、ght(cm)Volume(cm3)LiDAR volumeINTEGRATION IN PASSENGER CARS11Automotive LiDAR 2023|Conference|2023LiDAR volume vs locationLiDAR placed behind the windshield needs to be compact and to have a low height.GrilleRoofBumperWindshieldThe location of LiDAR is related to the use cases developed by the OEM.B
23、ut other parameters like volume and height of the LiDAR can enter into consideration.High-volume LiDAR,especially those based on fiber lasers,are too bulky to be placed in locations other than the roof.For a position behind the windshield,recent research from Cepton and Hesai shows that the volume n
24、eeds to be limited,as well as the height,to avoid obstructing the drivers vision.Due to their compact size,LiDARs from RoboSense or Hesai can be placed either in the grille,behind the windshield,or on the roof.Headlamps?RoboSense and Hesai have a cumulative 74%of the design wins.That illustrates the
25、 intense drive of Chinese OEMs and how quickly they are implementing LiDAR compared to the rest of the world.Valeo has 8%of the design wins.That does not reflect the reality as they have partnered with various OEMs for multiple car models.It is the case for Stellantis;it has been counted as one desi
26、gn win,but many car models,number unknown,will use LiDAR.The 1-billion-dollar cumulative order for the Scala 3 bears witness to the excellent shape of Valeo.A similar understanding can be made for Innoviz with VW;many car models are expected to embed LiDAR.PARTNERSHIPS AND SUPPLY CHAIN12Automotive L
27、iDAR 2023|Conference|2023Supply chain for passenger cars analysis:known design wins since 2018(total of 144)(1/2)There are two markets:China and the rest of the world.Chinese OEMs are adopting LiDAR at a very rapid pace.Robosense43%Hesai31%Valeo8%Innovusion5%Luminar3%Huawei2%Innoviz2%Denso1%Continen
28、tal1%Ibeo1%Livox1%Cepton1%Opsys Technologies1%Design wins-Split by LiDAR supplierCN79%FR8%US6%DE3%JP1%IS3%Design wins-Split by country of origin905nm-based LiDAR clearly dominates the market.LiDAR performance at this wavelength is still improving,both on the emitter side with EELs or VCSELs and on t
29、he receiver side,where SiPMs and SPADs are replacing APDs.Some new LiDARs use 940nm instead of 905nm,but this slight shift in wavelength is not significant enough to drastically change the LiDARs performance or the underlying components.Regarding imaging technology,MEMS and hybrid solid-state are al
30、most at the same level.The share of MEMS-based LiDAR is due mainly to RoboSense,one of the only players with Innoviz to use this technology.PARTNERSHIPS AND SUPPLY CHAIN13Automotive LiDAR 2023|Conference|2023Supply chain for passenger cars analysis:known design wins since 2018(total of 144)(2/2)Whil
31、e the 905nm-based LiDAR share is still increasing,MEMS technology is almost at the same level as hybrid solid-state LiDAR.905 nm89%1550 nm8%1064 nm2%885 nm1%Design wins-Split by wavelengthHybrid Solid-State51%MEMS44%Flash5%Design wins-Split by technologyThe LiDAR market for passenger cars increased
32、rapidly from 2021,with a 285%growth.Since its inception,the passenger car market has been dominated by Valeo,with partnerships with European OEMs like Audi and Mercedes.Innovusion,a Chinese supplier,is leading the market thanks to its partnership with the Chinese automaker Nio.Chinese OEMs are rapid
33、ly implementing LiDAR and are using local suppliers.In total,Chinese LiDAR manufacturers control 73%of the market.As Hesai and Robosense have many design wins,their 2022 revenue should be significantly higher as they are supplying OEMs with B-samples and C-samples for production validation.MARKET VI
34、EW14Automotive LiDAR 2023|Conference|20232022 passenger car LiDAR market shareInnovusion,Valeo,Hesai,and RoboSense are leading the LiDAR market for automotive.79%79%6%6%7%7%8%8%InnovusionInnovusion28%28%ValeoValeo24%24%HesaiHesai23%23%RobosenseRobosense14%14%LivoxLivox8%8%OthersOthers3%3%PC&LCV LiDA
35、R market shares 2021(Inner)vs 2022(Outer)Inner:$44MOuter:$169MMARKET VIEW15Automotive LiDAR 2023|Conference|2023LiDAR market dynamicsThere are two distinct LiDAR markets:China and the rest of the world(RoW).106 car models to be released with LiDAR by Chinese OEMs in 2023 or soon after.3+car models*w
36、ith LiDAR to be released by US OEMs in 2023 or soon after.Chinese LiDAR ASP:$600-$800Rest of the world LiDAR ASP:$800-$1,0006+car models*with LiDAR to be released by EU OEMs in 2023 or soon after.4+car models*with LiDAR to be released by Japanese and Korean OEMs in 2023 or soon after.*For these OEMs
37、,design wins are mostly related to a car platform rather than a dedicated model.MARKET VIEW16Automotive LiDAR 2023|Conference|2023Overview of PC&LCV and robotaxi marketsThe LiDAR market for PC&LCV and robotaxis is expected to reach$4.5B in 2028 with a 55%CAGR from 2023.2022202220282028$3,920M$3,920M
38、CAGR:69%CAGR:69%$557M$557MCAGR:25%CAGR:25%$317M$317M$4,477M$4,477MCAGRCAGR2222-2828:55%:55%PC&LCV LiDARPC&LCV LiDARRobotaxi LiDARRobotaxi LiDAR$169M$169M$147M$147MPARTNERSHIPS AND SUPPLY CHAIN17Automotive LiDAR 2023|Conference|2023Supply chain for passenger cars:LiDAR shipments in 2022The big four a
39、re composed of Innovusion,Valeo,Hesai,and RoboSense.225,000 LiDARs were shipped in passenger cars in 202225%25%12%24%10%1%1%1%1%Last year,we were expecting that ValeoValeo would keep its leadership in 2022.They are ranked 2nd,but the difference in volume with Innovusion is not significant.Thanks to
40、its partnership with Nio,with LiDAR mounted in series,InnovusionInnovusion put more than 50,000 units on the road in 2022.HesaiHesai is directly entering the market in 3rd position thanks to partnerships with local OEMs.RoboSenseRoboSense is also a significant player,with 12%of the market in 2022.Li
41、voxsLivoxs market share was 10%in 2022,thanks to the XPeng P5.The future may not be bright as XPeng chose RoboSense for its following car models.othersPARTNERSHIPS AND SUPPLY CHAIN18Automotive LiDAR 2023|Conference|2023Supply chain for passenger cars:What can we expect for 2023?In 2023,the market is
42、 expected to increase by a factor of 2.7.The big four are still the same but in a different order.Hesai could lead the market.632,000 LiDARs are expected to be shipped in passenger cars in 202341%29%12%10%1%4%1%1%1%OthersIn 2023,we expect a big push by the two leading Chinese LiDAR manufacturers,Hes
43、aiHesai and Robosense,Robosense,that have signed partnerships with many OEMs on multiple platforms.The two companies could ship 68%of the total number of LiDARs for passenger cars.InnovusionInnovusion has only partnered with Nio,so volumes cannot grow as rapidly as the leading players.ValeoValeo wil
44、l expect new shipments to Kia and Hyundai,but it should be limited to high-end cars and,therefore,in limited volumes.CeptonCepton could make its entry with GM and InnovizInnoviz with BMW;after several delays,they could see initial volumes.LuminarLuminar,among others,is still not shipping in high vol
45、umes,as the Volvo EX90 has been delayed until 2024.When looking at cumulated LiDAR volume,there is a clear difference between the first player,Valeo,and the others.Valeo is benefiting from partnerships with Audi,Honda,and Mercedes.Since 2018,Valeo has shipped close to 200,000 LiDAR units,a mixture o
46、f generations 1 and 2 of the Scala.Followers are newcomers,as they have only been shipping since 2021,at best.In this category,Innovusion,Hesai,RoboSense,and Livox have shipped between 27,000 and 57,000 LiDAR.PARTNERSHIPS AND SUPPLY CHAIN19Automotive LiDAR 2023|Conference|2023Cumulated volumes:a dif
47、ferent vision of the marketRegarding cumulated LiDAR volumes,Valeo is still,by far,the leader.-50,000 100,000 150,000 200,000 250,000Volume(units)Cumulated LiDAR volumes for automotive,2018 to 2022PARTNERSHIPS AND SUPPLY CHAIN20Automotive LiDAR 2023|Conference|2023Recent partnerships for passenger c
48、arsA position on the roof is promising,as seen in many new car releases.Source:Webasto Group SEContinental is licensing the technology from AEye.A SOP is scheduled for 2024.Baraja partnered with Veoneer.An SOP is scheduled for 2024/2025.The roof is an attractive position for detecting objects at a l
49、ong distance without being blinded by other cars.Roof supplier Webasto Group SE is developing a new solutions to integrate LiDAR and possibly other sensors in this position.Multiple partnerships with headlamp manufacturers have been made recently.These may be related to integration in headlamps but
50、also to manufacturing LiDAR sensors.LiDAR manufacturers could be looking for partners that have mastered automotive-grade products.One of the first cars with eyes-off functionality was the S-Class,released in 2021 with 13 sensors embedded.The XPeng G9,produced in 2022,with 19 sensors,was the first u
51、nmodifiedunmodified EV car qualified for autonomous robotaxi testing.This car embeds more sensors than the Waymo 4th Gen cars that were robotaxis in operation in the US.The car has received its testing permit from the Guangzhou Municipality.The BYD YangWang U8,scheduled in 2023,will embed even more
52、sensors,24 in total.This shows how fast Chinese OEMs are moving toward autonomous driving.TECHNOLOGY TRENDS21Automotive LiDAR 2023|Conference|2023The race for advanced eyes-off applicationsThe XPeng G9 is the first unmodified car to be qualified for autonomous robotaxi testing.Mercedes SMercedes S-C
53、lassClassWaymo 4Waymo 4thth GenGenXPeng XPeng G9G9BYD BYD YangWang U8YangWang U8Waymo 5Waymo 5thth GenGen924243939Number of sensors comparison(camera,radar,LiDAR)Number of sensors comparison(camera,radar,LiDAR)22Automotive LiDAR 2023|Conference|2023TECHNOLOGY TRENDSPlayer segmentationMech
54、anicalMechanicalHybrid SolidHybrid Solid-StateStateFully SolidFully Solid-StateState360 mechanical spinning1D Scan2D ScanOptical(some of these systems may need a mechanical scanning system)FlashIndividually addressablePolygon mirrorOtherMEMSPolygon+GalvoPrismsBaraja has an advanced development agree
55、ment with an OEMOEMsLiDAR ManufacturersAnd moreAnd moreThe InnovizTwo may abandon MEMS technology and use a polygon mirror instead.This should be related to vibration issues,production and calibration challenges.In the current development roadmap,LiDAR manufacturers are mainly focusing on reaching t
56、he longest range required by OEMs,around 200-250m for highway driving.The first generations of LiDAR have a stable price as more performant components are embedded.Once the range and resolution needs are achieved,LiDAR manufacturers will be able to focus on price reduction of the sensor.TECHNOLOGY T
57、RENDS23Automotive LiDAR 2023|Conference|2023Automotive LiDAR development roadmapOnce a sufficiently long range is achieved,LiDAR manufacturers will focus on sensor price reduction.Range&resolutionPriceGen 1Gen 2Gen 3Gen 4250mGen 5$1,000Focus on range Focus on range and resolutionand resolutionFocus
58、on Focus on price and price and volumevolumeScala 1Scala 2Scala 3Some new players like Hesai or Innovusion,are directly entering the market with a Gen 2 LiDAR.With an aggressive product development,their next generation could jump one generation and see a strong price reduction.2024202420
59、02Automotive LiDAR 2023|Conference|2023TECHNOLOGY TRENDSPassenger LiDAR comparison chart Short-rangeModelHFL110vFT120RS-LiDAR-E1Channels#FlashFlashFlashRange(m)(reflectivity)50m30m(10%)30m(10%)Points frequency(pts/sec)102,400192,000276,480(10Hz)FOV():H x V120 x 30100 x 75120 x 90Resolutio
60、n():H x V 0.9 x 0.90.6 x 0.60.6 x 0.6Power(W)15-15Dimensions(W-D-H)120 x100 x65-95x46x70AvailabilityAvailableQ3-2023Q3-2023Source:RoboSenseSource:HesaiThe typical position of short-range LiDAR will be on the side of the car,and the fender could be used by many OEMs.This is already the case for Lexus
61、 with the Continental Flash LiDAR.The goal is to detect small objects close to the car.In terms of driving functionality,city driving is a clear target with applications like overtaking,parking,and unprotected left turns,among others.But such a use case might be too early to implement for OEMs,and t
62、he penetration rate of this LiDAR is expected to remain low for the next few years.The typical architecture of these LiDARs will be the combination of a VCSEL array with a SPAD array.The recent release of the Sony SPAD(IMX459),with a resolution of around 100,000 pixels,is undoubtedly one of the enab
63、lers of the current flash LiDAR released by Hesai and RoboSense.In combination with that,VCSEL arrays using multi-junction technology,where a higher output power within the same surface can be obtained,is another enabler of flash LiDAR.NIR84%SWIR16%Wavelength choice-2023In the next ten years,the mai
64、n changes could be related to imaging technology.Since 2018,the vast majority of LiDARs have been mechanical-based due to their relatively low cost and because they provide good-enough performance.Flash imaging technology could rapidly increase to enable fully solid-state LiDAR with a high semicondu
65、ctor integration and contribute to a fast decrease of MEMS technology.The wavelength choice is not expected to change drastically.This is due to significant progress made on the receiver side.TECHNOLOGY TRENDS25Automotive LiDAR 2023|Conference|2023Conclusions Imaging technology and wavelengthIn the
66、next ten years,one of the main technology evolutions is expected on the imaging side.In the long term,technological choices will have to be made to be aligned with mass production,low cost,and high performance.In this context,we believe that flash and NIR wavelength could be dominant technologies.Hy
67、brid Solid-State68%MEMS30%Flash2%Optical0%Imaging technology-2023Hybrid Solid-State56%MEMS7%Flash33%Optical4%Imaging technology-2033NIR86%SWIR14%Wavelength choice-2033APD27%SPAD6%SiPM67%Photodetector type-2023Since 2018,the vast majority of LiDARs have been based on EELs,though in 2023,we see a stro
68、ng penetration of VCSELs and fiber lasers.Fiber laser penetration is linked to one LiDAR player,but we do not expect it to last due to its high cost.On the contrary,we expect VCSELs to continue their penetration due to their ever-improving performance.The wavelength choice is not expected to change
69、drastically.Significant progress has been made at the receiver side,and SiPMs are progressively replacing APDs,enabling high sensitivity and,therefore,longer range.TECHNOLOGY TRENDS26Automotive LiDAR 2023|Conference|2023Conclusions Light source and photodetector typeVCSELs and SiPMs are progressivel
70、y replacing EELs and APDs.In the long term,technological choices will have to be made to be aligned with mass production,low cost,and high performance.In this context,we believe that flash and NIR wavelengths could be dominant technologies.EEL45%VCSEL39%Fiber Laser16%Light Source-2023EEL43%VCSEL45%F
71、iber Laser12%Light source-2033APD18%SPAD22%SiPM60%Photodetector type-203327Automotive LiDAR 2023|Conference|2023LiDARs are being integrated in more affordable cars.This is related to the strong dynamics of Chinese OEMs:There is a clear evolution from the F-segment to the D-segment.More affordable ca
72、rs means higher volumes for LiDAR manufacturers.Between 2021 and 2022,the LiDAR market revenue has been multiplied by a factor x3.8(4).New players are emerging fast:Innovusion,Hesai.Valeo is no longer the leading player but is still a leading player.The ranking of players will change again in 2023 a
73、s the volume is expected to be multiplied by 3.The LiDAR market is at crossroads with revenue for passenger cars surpassing those from robotic vehicles for the first time.In total the LiDAR market is expected to grow from$317M in 2022 to$4.5B in 2028 at a 55%CAGR.KEY TAKEAWAYS28Automotive LiDAR 2023|Conference|2023If you have any questions,feel free to contact me by email:com THANK YOU FOR YOUR ATTENTION