1
`
`Mercedes-Benz USA, LLC, Petitioner - Ex. 1007
`
`

`
`s4-o-07
`Automated Vehicle/Highway System
`
`Norio Komoda, Keiji Aoki, Takaharu Saito,
`Takashi Shigematsu, Hidetoshi Ichikawa
`Toyota Motor Corporation
`
`Abstract
`
`This presents TOYOTA’s concept, experiments and
`future scope of AVHS (Automated Vehicle/Highway
`System) which could contribute to a possible solution to
`automobile traffic/transportation issues in the 21st
`century.
`Concept: This system enables smooth, automated
`cruising on highways by keeping the distance to the
`leading vehicle and avoiding obstacles. Compact, light-
`weight actuators are designed from a practical viewpoint.
`The system is intended to have broad benefits for
`vehicles with add-on devices as well as for automated
`vehicles.
`
`Findings: The prototype runs smoothly over 100 km/h
`satisfying the above require‘ments with simple control
`algorithm. CCD lane sensor with compensation to dis-
`turbances can detect
`the lane except under severe
`weather conditions. The improvement of road structure
`and lane would make the sensor more robust. To make
`
`the system more reliable, misperception of vague lane is
`corrected by the onboard memory of 3-D road curvature
`as a backup. Onboard laser radar is feasible for obstacle
`or distance sensing and obstacle avoidance control with
`assist of road side TV camera with computer image
`analyzer. which can detect smaller obstacles and is a key
`solution. 'l11is forms a cooperative intelligent vehicle]
`infrastructure. With some compensation laser radar can
`detect the leading vehicle except under severe conditions
`such as small road curvature, bad weather, etc.
`Scope: AVHS is expected to penetrate effectively
`because intelligent
`infrastructure can widely provide
`beneficial information for vehicles with telecommunica-
`
`tion receivers as well as sure backup for automated
`vehicles. Further studies and discussions are necessary to
`obtain system reliability and social consensus.
`
`Background
`In Japan as well as in American and European
`countries, automobile traffic/transportation issues in the
`21 C have been focused on in recent years in pursuit of
`effective and efficient ways to improve safety. conges-
`tion and environmental protection. In the following the
`related backgrounds are overviewed concerning Japanese
`traffic/transportation
`issues.
`the
`trend of AVCS
`(Advanced Vehicle Control System) and the historical
`overview of automated vehicle control systems. Our idea
`stands on the basis of this overview.
`
`Section 3: Technical sessions
`
`-
`Automobile Traf]'icITransportation in Japan
`The following is our future prospect for Japanese
`automobile traffic/transportation (Fig. 1-4): The con-
`struction of highways should be eagerly pursued because
`of their much lower accident rate than that of normal
`roads. However, the future construction plans in Japan
`will not provide enough capacity to absorb the predicted
`increase of VKT (Vehicle Kilometers of Travel) if the
`future highway remains in its traditional form.
`The accident statistics show that for the effective
`accident avoidance on normal roads. measures should be
`taken for rear end. head-on and side collision with
`vehicles, and collision with road side constructions. On
`highways, collision with road side construction and rear
`end collision are the major issues. The increase of aged
`drivers and pedestrians should not be neglected either.
`
`Total
`
`lenath of
`
`length of
`Total
`hi hway
`(P anning)
`
`' 9 0
`
`2 O 0 0
`
`2 O 1 0
`
`Figure 1. Total Length of Highway and VKTIday in Japan
`
`Number of’ fatal vehicle accidents
`in the each year and
`its ratio to '85
`8. 14x10" 1/VKT
`
`2.0
`
`'85'86'87'88'
`
`Figure 2. Number of Fetal Vehicle Accident: in Japan
`
`The future congestion issue should seriously he
`considered both for highways and normal roads.
`Thus the cooperative intelligent vehicle/infrastructure
`should possibly providea key solution for the trafficl
`
`2
`
`

`
`13th lntematlenal Technical conference an Experimental safety Vehicles
`
`ufmulti-vehilcwcidem:
`-
`
`orfizothm
`8 E :rIruevenlaIaaeclaama
`calm :veI|eleeulIlslenotltulnnIear-uttl
`ouEl :other|ol|'I1fl:ve1tlclnecldents
`or
`an
`
`margin for accident avoidance maneuvers. although that
`margin depends on human factors.
`Acccrelat lomm/-'
`performance
`tea
`:.I I
`[B In! til:
`rel‘!-In
`circle
`ve}:lcla.wI In-knrgle I-lvor
`
`Vehlcleatxitlennon
`hightvaysinhpen
`
`Vehlcleneeloentson
`hl¢ImyelnU5.A.
`
`[D] Normal vehicle
`with nonml drlvar
`
`V
`
`50.8
`
`El
`
` own
`
`Anxlnnnl ronsd
`on DI
`roads
`
`3' ‘mar-V.
`
`clonal rounds
`r
`Nala :
`
`nlflmavt
`
`on II! D‘
`roads
`
`-
`
`:D:::|aly Inttlgltorl ¢i:§1;'I‘crt“:hA dllirlct of Id 3%:
`r3":mn 4&8“? no-U‘-"u°=S more min 8888 I’ ‘’
`i":2n"c‘..
`Egpéd-C:?3?l:On'né|0 - (fly: c’v"gl'ig-o cl lha)rotd / thy)
`-
`road’ I width on (has: Itnklstics in over 6.511;
`- an : Mean val My at
`he t me a nu: mun tr: tic voluno
`
`Flgtle 4. mo congestion Statistics In Japan
`
`transportation issues in the 21 C. While it would be
`effective both for highways and normal roads, from the
`viewpoint of technical feasibility the first plan should be
`for highways.
`
`Overview on AVCS Technology
`The trend is best understood when it is divided into 3
`
`phases based on the typical evolutional features as shown
`
`Flgure 5. Overview of Trend on AVCS
`
`Phase 1: The current AVCS Technology is going to
`cover almost all kinds of vehicle control subsystems and
`integrate them for smooth and high vehicle dynamic per-
`formance to the maximum of the tire friction circle as
`
`shown in Fig 6. The main subsystems including ABS,
`TRC, 4WS, 4WD and Active Suspension are currently
`being developed amidst tough competition. They could
`be more effective if equipped with more advanced active
`actuators. They are considered as fundamental factors for
`the so-called active safety system that provides the safety
`
`460
`
`Decerelnuonrlt/3'
`
`Figure 6. Effects of AVCS by Integration of Vehicle
`htelllgent control subsystems for High Performance.
`Easy and Safety Drlve
`
`Phase 2: Phase 2 systems are now moving from the
`research phase to the development phase. The important
`technical evolution in this phase is to substitute human
`perception and reaction with sensors and advanced active
`actuators. This could bring about a revolutionary change
`to the future of automobile safety and mobility. Various
`types of AVCS products could be introduced such as the
`rear end collision warning or avoiding system, lateral
`warning or control system, etc.
`Phase 3: In addition to the AVCS in Phase 2, a more
`advanced and wide spread application of Info-Mobility
`System (intelligent
`traffic management system and
`vehicle-road telecommunication system) would make a
`great contribution to automobile traffic/transportation in
`the 21 C. This paper treats AVHS based on this back-
`ground.
`
`Historical Overview ofAutomated Vehicle Control
`System
`As shown in Fig. 7, over 20 years many papers have
`been contributed mainly from technical interests in the
`most advanced technology at that time. We studied on
`these previous contributions thoroughly and selected
`carefully compact, light weight. cost-effective and reli-
`able control devices to construct the best cooperative
`intelligent vehicle/infrastructure system available at this
`point.
`
`n.,5i5;a*g=._-1:°a>m..4 .. .....
`-are-....°.:'
`on
`us
`.
`‘m n : ‘_
`Aru'—'r.a uunu sun a-tun-i nl I-amnion -In «rule
`r— n tau». an vunn nut-can-n
`‘I
`tune-use of End?“
`-- -'..:-°-..;..
`
`~'---.....
`3_g;.__..;, 39...
`4-u—uur vlillslu u:|I earning. 1 running: an un-
`a:“.‘°..':.._..5{'.".’.'.i.'”:.Tnr:. our ea-new n-at lllom
`6‘-n. Igguu-i
`an
`Inn nrIeua.t.lA(
`.'
`am. a mm W 1: run an
`an run run
`me
`‘”.~..._...g""""".'.'..'.’.‘l'..‘‘u'.‘:'.‘.‘.-“ w“""'_‘.'.!".... .. um.
`XIKC VCIIGIZ fill Ij I‘! lIl|
`u
`ul unau-
`I wt e-—a7‘:u.:
`u'u’7'.'¥3'€-'-'u’5n"'-'33‘-i"3u"°
`03.3.. c an
` ‘IEII11 KMII lm
`a.x""""""v'H.'.l'a‘
`ll a-3519:: lane nne
`3:.-t
`nevus emu
`Into -um aunt-n or Inn lI—
`of." E
`aahunaaruvu-int.-nmmn--nu,
`kWi|. ICIGIO :10: Kill} III‘ Ill—
`‘I I? 1 IT
`08 Itdlu
`VIII-nib flu llufl unlril
`u-mu rut - aria:-/n - nu-run. an-n
`
`-mu
`. NI‘!-lfigla UQIGII I,?r|n"I n_
`u mu-you en -nu-unis
`1 :III A; II I’ hf IRIUIIIBBIIII
`
`Figure 7. Historical Overview of R&D on
`Automatedmutonomous Vehicle control System
`
`3
`
`

`
`System Concept
`The R&D project of AV!-IS gives challenging chances
`to R&D engineers in this field to accelerate the progress
`in the system development itself and to look for feasible
`technical byproducts in this time of emerging new tech-
`nology.
`This system is planned to enable the automated
`vehicle with lane and obstacle sensors to run auto-
`matically between ICs over 100 krn/h on the cooperative
`intelligent highway lanes. It runs on the 2 lanes with
`intelligent infrastructure of the 2.6 km 3- (partly 4-) lane
`circuit with the parking lot or the IC. The presence of
`any other normal or automated vehicles on these 2 lanes
`is allowed.
`
`The system provides smooth lane trace control. safe
`distance control, cruise control, obstacle avoidance by
`stopping or lane changing control and exit/entrance
`control using extremely simple control algorithms.
`The onboard system has the following:
`
`- Compact, lightweight and cost-effective actuators for
`the steering, brake and throttle systems.
`- Cost-effective lane sensor and obstacle sensor (that
`senses only four-wheel vehicles and motorcycles)
`that cooperatively work with the intelligent infra-
`structure system, which provides backup for both
`onboard sensors. Onboard 3-D road curvature
`
`memory is also provided for the backup of onboard
`lane sensor.
`
`- ECU and vehicleroad telecommunication system.
`
`The intelligent infrastructure system has the following:
`- White lane line for cruise and red lane line for
`exit/entrance, which are easy to see even under bad
`weather conditions.
`- Obstacle detecting system which serves as a re-
`dundant system for the onboard detector, but also as
`a more robust. precise detector of smaller obstacles
`under severe disturbances.
`- Traffic control center with the vehicle-road tele-
`
`communication system which provides (a) informa-
`tion to assist the automated vehicle to run smoothly
`and (b) information for traffic control.
`
`Cooperative Vehicle/Infrastructure Concept:
`- The investment should be reasonable and efficient
`
`compared to the broad benefit not only for auto-
`mated vehicles but also for nonnal vehicles equipped
`with only some subsystems and/or'the telecommuni-
`cation receivers that make effective warning systems
`and/or
`semi-automated
`systems
`for
`accident
`avoidance.
`
`- The investment would be relatively small compared
`to the much greater investment for highway con-
`struction in Japan even if the most advanced tech-
`nologies are deployed for the cooperative intelligent
`vehicle/infrastructure. However,
`it would not be
`
`Section 3: Technical Sessions
`
`reasonable to pursue perfect backup under very
`severe disturbances.
`
`.
`Plan and Design of AVHS
`The following are the special features of the proto-
`type. The basic model is 1990 Toyota Camry (Fig. 8 and
`9).
`
`Figure 9. Steering Actuator and color OCD Lane sensing
`system
`
`Onboard System
`Actuators. The steering actuator is a lightweight,
`compact and high powered brushless DC motor which is
`installed coaxially with the steering main shaft of the
`hydraulic power assisted steering, as shown in Fig. 9 and
`11. The specification of the motor is shown in Table l.
`The driver can take over the steering wheel at any time.
`
`Table 1. Actuator Specification
`""'°'"°
`Em"
`a
`t 111-!
`r W“
`IIIIV
`:?no'§‘.‘.'.?.§':'..g:" I-35$
`I
`::"u° 4'5?‘
`SA
`
`~
`
`V. "H"
`1
`‘a:‘°'
`: O .
`
`4
`
`

`
`130! International Technical Contetance an Experimental safety Vehicles
`
`Figure 12. Electronic Controlled Throttle
`
`The onboard traffic monitoring system is installed in
`the center instrument panel (Fig. 25).
`ECU and telecommunication systems are installed in
`the luggage compartment.
`
`Infrastructure System
`Intelligent proving ground (Fig. 13). (1) The proving
`ground is 2.6 km-long oval circuit. The central 2 lanes
`of the 3-(partly 4-)ianes are used for AVHS. with the
`parking lot assumed as the IC for exit/entrance control.
`The specially painted bright lane line with many small
`spherical asphalt spots is perceptible in bad weather. (2)
`Ten beacons and a TV camera are implemented on the
`course. (3) The traffic control center is located at the
`assumed IC.
`
`‘
`
`1i
`
`detection
`rafflclrnonl cor dl play
`
`Computer inane
`one
`control
`center --0-. Connunlcatlan
`E controller
`
`Figure 13. Intelligent Infrastructure System
`
`The vehicle-road telecommunication is done between
`
`the antenna on the roof of the prototype and the 10
`beacons located along the test track through the traffic
`control center. The information exchanged is used for
`traffic control and for smooth and safe drive control as
`
`shown in Fig. 14. The communication protocol is also
`shown in Fig. 14.
`Road side TV camera and computer image analyzer
`for obstacle detection: The TV camera is implemented
`on a pole of 8.8 m-height to detect obstacles on the road
`from 10 to 30 m ahead or from 100 to 500 tn ahead. The
`
`TV camera and the computer image analyzer in the con-
`trol center function as a redundant backup system as well
`as a reliable obstacle detector for smaller obstacles. The
`specification is shown in Table 3.
`The trafiic monitor display is in the control center as
`shown in Fig. 26.
`
`Findings
`Lane Sensing and Steering Control
`The prototype succeeds in running along the lane over
`100 km/h using a simple steering control algorithm to
`detect the lane of 10 to 20 m ahead (Fig. 15-17). in this
`
`The brake actuator is driven by a hydro-electronic
`valve powered by the ABS pump and a lightweight,
`high-response solenoid with moving core, which are so
`installed in parallel to the master cylinder of the-foot
`brake that the driver can actuate at any time. The specifi-
`cation is shown in Table 1.
`
`The throttle actuator is a direct drive pulse motor (Fig.
`12 and Table 1).
`
`Sensors. The lane sensor is a color CCD sensor (TV
`camera) mounted at the inside rear view mirror immedi-
`ately behind the top of the windshield glass as shown in
`Fig. 9. It watches for the lane line from 10 to 20 m
`ahead. The specification is shown in Table 2. The 3-D
`course curvature memory provides instantaneous backup
`in case of any failure of the lane sensor with assist from
`the vehicle position information from the roadside
`beacon.
`is a scanning laser radar
`The obstacle detector
`mounted at the front radiator grill, as shown in Fig. 10.
`It watches mainly for vehicles and motorcycles from S
`to 120 m ahead. The specification is shown in Table 2.
`The detection of smaller obstacles'depends on the
`intelligent infrastructure.
`
`Figure 10. onboard system
`
`Table 2. onboard Sensor specification
`rndn
`'«‘a=?"om-ole «see-cu-s
`woe 2 dimensional 0010-‘ Cm :
`l
`:
`5 '- 1
`can l
`1333
`"its :r::i=':°'.:.i-i°'.*.:.-!:°"* ='°*°=‘-° ° ".""°.
`2.
`Heal mile : :1.5 -
`25 ~ I/10000 sec
`aet|nn‘l’:; speed
`:
`120 ms
`olfgltronic shutter :
`tsroamn n‘
`G
`
`n.4ue~.-e»4<ne—- n...m-auraueawws
`
`944.
`
`.-
`
`:_......
`
`I
`
`Figure 11. Electronic controlled Steering Actuator
`
`462
`
`5
`
`

`
`test. the detection range is set to 10 to 20 m ahead in
`order to eliminate the influence of vehicle pitching
`phenomenon, the minimum road curvature of 50 m R on
`Japanese motorways, the distance to the leading vehicle,
`the reach of the head lamp at night. etc.
`
`infrastructure
`side
`
`Vehicle side
`
`T : Trteeer
`D : Deirnnd text :'l'rafflc control center to vehicle
`S : Service Text
`:Veritcle to traffic control center
`A : Acknowledgement reception
`Service text
`
`travel mode
`velocity
`distance between
`vehicles
`
`travel mode
`obstacle information
`velocl Iv
`distance between
`vehicles
`
`I!
`
`Figure 14. Telecommunication Protocol and Information
`
`Table 3. Reed Side Obstacle Deteetlon System
`R-‘-“-'e'i’5‘-«Ric--onnae
`electnanicl
`body Illa w I
`70 I 70
`
`5 - fI(l.v) 1: + l'.t(l.v) no 1- I II
`E : urge: oflaterul displacement
`V : velocity
`l'.. (1: functions of I and v
`I ICDHINQ
`miaduected point line to the
`moon, round from
`eompllllnn with the detected
`
`Lune line
`
`Aulomlled vehicle
`Lane line
`
`Ilhu
`I0
`
`l3ln l.‘im Ilim mu
`8
`6
`4
`2
`
`Figure 15. Lane Sensing Algorithm and steering Control
`Algorithm
`
`e
`Steerin a
`znsl
`deg
`l
`
`Yaw rate
`deysec
`1
`
`-l0
`
`0
`
`4
`
`s
`
`-1
`
`0
`
`4
`
`12
`sec
`
`8
`
`12
`SOC
`
`Steering angle
`Automated deg
`steering
`10
`control
`0
`
`Yaw 111°
`deysec
`I
`0 _e. au-
`
`O48l2 04812
`SCC
`SOC
`
`Figure 16. stability and control of the Prototype: steering
`Correction and Yaw Rate on straight Line
`
`Section 3: Technical Sessions
`
`The course lane detecting algorithm for the CCD
`sensor is quite sensitive to the change of the sun shine
`due to the influence of the shadows from road side trees
`or constructions and the brightness of the sun itself
`owing to the change in clouds. etc. (Fig. 18-20). The
`feedback control of the CCD sensor using illuminance
`meter output of the road surface brightness is an
`effective way to make the lane sensing system robust to
`changes in brightness.
`
`Root mean square
`of steering single
`
`$3
`-
`1' 0
`o_ 5
`
`u:Dr-iver’ s
`$1
`Note:
`steering control
`,
`J3
`I D I
`o:AutcI'm.ted
` ’m 1oo-irtngetan/ta“ 13s°3¢E.r°l
`lei, ,
`m2 60km/h
`sec;
`
`O. 1 0.2 0.3 deg/sec
`0
`Root mean square or yew rate
`
`Figure 17. Stability and control of the Prototype:
`Root Mean square of Steering Angle and Your Rate on
`Straight Line
`
`._. —.._
`
`Figuve 18. Pereeptibility of the Lane Under the shadow
`of Trees
`
`Pereeptibltlty of the Improved"
`Flgure
`Rainy Weather
`
`6
`
`

`
`
`
`
`
`
`
`
`
`13th international Technical Conference an Experimental Safety Vehicles
`
`§
`
`
`
`Figure 20. Perceptibility of the Improved Lane at Night
`
`
`
`
`
`
`
`
`
`It is also sensitive to the influence of the wet road that
`
`
`
`
`
`
`
`
`
`
`
`interferes with lane detection by the CCD sensor. Many
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`small convex spots on the pavement and/or bright paint
`
`
`
`
`
`
`
`
`
`on the lane line improve the detectability to a consider-
`
`
`
`
`
`
`
`
`
`
`able extent (Fig. 19) as well as the application of a
`
`
`
`
`
`polarizing filter on the lens.
`
`
`
`
`
`
`
`
`An effective way to eliminate the influence of dis-
`turbances to the lane sensor is to reproduce the lane
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`outline by the smoothing method using the past memory
`
`of the 10 points detected at a period of 1/30 see, as
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`shown in Fig. 15. If the automated highway has only one
`
`
`
`
`
`
`
`
`
`
`
`
`lane, it is most practical to place the lane marker on the
`side wall of the road side construction to avoid disturb-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`ances by bad weather conditions as mentioned above.
`
`
`
`
`
`
`Exceptionally difficult cases are during sunset or
`
`
`
`
`
`
`
`
`sunrise, etc., when the sunshine beams directly into the
`lens. For such exceptional cases the memory of the road
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`curvature and the present location of the vehicle in-
`formed from the telecommunication through the beacons
`
`
`
`
`
`
`
`are effective as a redundant backup to the lane sensor
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`system. However, since the current system does not
`
`
`
`
`
`
`
`
`
`provide information on the lateral position of the vehicle,
`
`
`
`
`
`
`
`
`
`
`this backup system is effective only for the short period
`
`
`
`
`
`
`necessary to stop the vehicle safely.
`
`
`
`
`
`
`
`
`
`Obstacle Detection and Obstacle Avoidance Control
`
`
`
`
`
`
`
`
`
`
`
`
`
`The onboard scanning laser radar can detect vehicles
`
`
`
`
`
`
`
`
`
`
`
`
`from 5 to 120 m ahead as shown in Table 2. Although it
`
`
`
`
`
`
`
`
`sometimes misdetects in the case of severe pitching or
`
`
`
`
`
`
`
`
`severe curvature of the course, errors can be compen-
`
`
`
`
`
`
`
`
`
`
`sated by the memory of the past detection to some
`
`
`
`
`
`
`
`
`
`
`
`extent. Even on a tight curvature or in the fog the
`
`
`
`
`
`
`
`
`leading vehicle can be detected at shorter distances.
`However, since the passengers might feel unsafe at very
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`close distances, the system would need the integration of
`
`
`
`
`
`
`
`
`
`
`2 kinds of sensing systems and would cost more unless
`the road side detector could form a reliable backup.
`
`
`
`
`
`
`
`
`
`The road side TV camera can detect obstacles of 0.3
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`m x 0.3 m x 0.05 m on the road at a distance of 10 to 30
`maheadorof2mx2mx0.3montheroadata
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`distance of 100 to 500 m ahead by means of image
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`analysis at every 1/30 sec as shown in Fig. 22 and Table
`
`
`
`
`
`
`
`
`
`
`
`3. This method would be one of the most effective ways
`
`
`
`
`
`
`
`
`
`
`to detect obstacles on the road and probably be a reliable
`
`
`
`
`
`
`
`backup system for the onboard obstacle detector.
`
`464
`
`
`Figure 21. Beacon on the Test Truck
`
`
`
`
`
`
`
`
`
`® Range of
`obstacle deteclion
`
`(D Range til"
`
`obst.-n tlor.
`
`i’
`
`‘
`
`TV Camera
`
`Figure 22. Range of Obstacle Detection on the Road
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`When an obstacle is detected either by the onboard or
`the road side detector, the vehicle can be controlled for
`
`
`
`
`
`
`
`
`
`
`straight stopping or lane change under instructions from
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`the traffic control center (Fig. 23 and 24).
`' J“
`
`Figure 23. Detection of a Standing Vehicle Ahead and
`
`
`
`
`
`
`
`Lane Change
`
`
`
`
`
`Traffic Control
`
`
`
`
`
`
`
`
`
`
`
`The traffic monitoring display is shown in Fig. 25 and
`26.
`
`Informations and instructions for exit/entrance control,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`emergency stop, lane change, speed limit and distance
`
`
`
`
`
`
`
`
`control can be exchanged through beacons by the
`
`
`
`
`
`
`
`vehicle-road telecommunication system (Fig. 14 and 21).
`
`
`
`7
`
`

`
`Section 3: Technical Sessions
`
`
`
`
`
`
`
`be on the side wall and less prone to the disturbances of
`
`
`
`
`
`
`
`
`
`
`
`
`severe weather. etc. Since the actuator system is more
`
`
`
`
`
`
`
`
`
`responsive than the human driver,
`the distance of
`
`
`
`
`
`
`
`
`detection could be shortened, reducing the influence of
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`thick fog, heavy rain, pitching phenomenon, etc. on the
`
`onboard lane sensor. It would be practical to apply the
`
`
`
`
`
`
`
`
`
`HOV lane at the 1st step. When the highway has 2 lanes
`
`
`
`
`
`
`
`
`
`
`
`
`or more, the lane line should be set up so that it can be
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`detected under bad weather conditions by the use of such
`
`
`
`
`
`
`
`
`
`
`means as convex reflecting markers on the lane line. The
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`backup system can be more complete if it is combined
`with another sensing system like a lateral position
`
`
`
`
`
`
`
`
`sensor, the vehicle position information and 3-D road
`
`
`
`
`
`
`
`
`curvature memory. Since the automated vehicle looks at
`
`
`
`
`
`
`
`
`20 m ahead at most, other leading vehicles do not disturb
`
`
`
`
`
`
`
`
`
`
`
`the lane sensor of the automated vehicle, allowing even
`
`
`
`
`
`
`
`
`
`
`
`normal vehicles to run on the same lane.
`
`
`
`
`
`
`
`
`The following are possible alternative or backup
`
`
`
`
`
`
`methods for CCD lane sensor. Each of them has advan-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`tages and disadvantages compared to the CCD lane
`sensor:
`
`The underground cable and the coil sensor system
`
`
`
`
`
`
`
`
`have the advantage that
`it
`is robust
`to the weather
`
`
`
`
`
`
`
`
`
`
`disturbances, but has the disadvantages of vulnerability
`
`
`
`
`
`
`
`to any magnetized material or the electric wires around
`
`
`
`
`
`
`
`
`
`the road structure, and the difficulty of its maintenance.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Even though the lane detection is done right under the
`vehicle it can control the prototype to run at a speed of
`
`
`
`
`
`
`
`
`
`
`
`
`100 km/h.
`
`
`The laser radar to sense the distance to the side wall
`
`
`
`
`
`
`
`
`
`
`
`would work as a backup system when combined with the
`
`
`
`
`
`
`
`
`
`
`onboard memory of the road curvature with the vehicle
`
`
`
`
`
`
`
`
`
`position information provided that
`the application is
`
`
`
`
`
`
`
`limited to the one lane system.
`
`
`
`
`
`
`The laser radar and the road lane marker with reflect-
`
`
`
`
`
`
`
`
`
`ors would be more robust to weather disturbances than
`
`
`
`
`
`
`
`
`
`the CCD sensor system, but would cost more.
`
`
`
`
`
`
`
`
`The above methods would be feasible to some extent.
`
`
`
`
`
`
`
`
`
`However the most important PԤint is that any attempt to
`
`
`
`
`
`
`
`
`
`
`build a reliable system with onboard sensors alone would
`
`
`
`
`
`
`
`
`
`probably cost too much and not be practical. From this
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`viewpoint
`the cooperative
`intelligent vehicle/infra-
`structure system is preferred, and the CCD sensor system
`
`
`
`
`
`
`
`
`
`would probably be the most practical and efficient way
`
`
`
`
`
`
`
`
`
`for lane sensing.
`
`
`
`
`
`
`Figure 24VDetection of an Obstacle by Fioad_§ide
`
`
`
`
`
`
`
`
`TV Camera
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Figure 26. Traffic Monitor and Computer Image Analyzer
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Concluding Remarks and Discussions
`Lane Sensing and Steering Control
`
`
`
`
`
`The CCD sensing system with the illuminance feed-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`back and the simple steering control algorithm can make
`successful
`test runs over 100 km/h. Hereafter further
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`efforts should be given to extend its sensing ability.
`The lane sensing ability would become almost com-
`
`
`
`
`
`
`
`plete if an improved road infrastructure is set up. Then
`
`
`
`
`
`
`
`
`
`the total cooperative system would be more reliable,
`
`
`
`
`
`
`
`cost-effective and redundant. For example if the auto-
`
`
`
`
`
`
`
`mated vehicle is limited to one lane, the lane line could
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Obstacle Detection and Obstacle Avoidance Control
`
`
`
`
`
`
`The onboard scanning laser radar has to overcome
`
`
`
`
`
`
`
`
`some difficulty to get a reliable detection performance
`
`
`
`
`
`
`
`
`under
`the influence of
`road curvature, etc.
`It
`is
`
`
`
`
`
`
`
`
`
`vulnerable to foggy weather, too. If the system is limited
`
`
`
`
`
`
`
`
`
`
`to one lane, on onboard scanning laser radar with the
`
`
`
`
`
`
`
`
`
`
`assistance of CCD lane sensor system would be more
`
`
`
`
`
`
`
`
`
`reliable. Among other onboard sensor systems now under
`
`
`
`
`
`
`
`
`
`R&D, a simple image analyzer combined with the CCD
`
`
`
`
`
`
`
`
`
`sensor might have the possibility of an effective solution
`
`
`
`
`
`
`
`
`
`in the future.
`
`
`
`
`465
`
`
`
`8
`
`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`13th Intematlonal Technical conference an Experimental Safety Vehicles
`
`
`
`
`
`
`
`
`
`
`
`Road side TV camera would be the most sturdy and
`reliable way to detect smaller obstacles on the road,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`since it detects them on a stationary background, elimi-
`nating the influence of weather. The investment for the
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`system per unit road length would be relatively small as
`
`
`
`
`
`
`
`
`compared to the huge investment for the highway con-
`
`
`
`
`
`
`
`
`
`struction even including the cost of maintenance for the
`system.
`
`
`Practical Approach
`
`
`
`
`
`
`
`
`
`
`An exclusive lane for automated vehicles might be
`
`
`
`
`
`
`
`
`
`
`possible in the distant future, but would not be practical
`
`
`
`
`
`
`
`
`
`
`during the process of the penetration of AVHS. The
`automated lane should provide broad benefits even for
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`normal vehicles that only have sensor systems and/or
`telecommunication receivers to get information from the
`
`
`
`
`
`
`
`traffic control center.
`
`
`
`The detection of other vehicles on the two or more
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`lanes is not easy from the vehicle side alone. It needs
`
`
`
`
`
`
`
`
`assistance from the road side detection system. Although
`
`
`
`
`
`
`
`
`
`
`
`
`its possibility was proved by the use of the road side TV
`camera and computer image analyzer, traffic control on
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`two or more lanes would be the 2nd step after the pene-
`
`
`
`
`
`
`
`tration of the single lane automated highway.
`
`Other Issues
`
`
`
`
`
`
`
`
`
`
`It is necessary to proceed to further studies to estab-
`
`
`
`
`
`
`
`lish sufficient reliability as the cooperative vehicle/
`
`
`
`
`
`
`
`
`
`highway system. Ample discussions and field tests to get
`
`
`
`
`
`
`
`
`social consensus on the institutional and legal issues
`
`
`
`
`
`
`
`
`concerning any possible failure are also required before
`their deployment.
`
`
`
`
`Acknowledgment
`
`
`
`
`
`
`
`This project has been proceeded by the engineering
`
`
`
`
`
`
`
`
`
`staffs of the special project team including the staffs in
`
`
`
`
`
`
`
`charge of the intelligent proving ground. The contribu-
`
`
`
`
`
`
`
`
`
`
`
`tions in this field over the past 20 years were quite
`instructive as well as advice and support from the related
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`supervisors and colleagues. The authors would like to
`thank all of them.
`
`
`
`
`
`
`
`
`
`References
`[1] Y. Ohshima, E. Kikuchi, M. Kimura, S. Matsu-
`
`
`
`
`
`
`
`
`
`
`
`
`
`moto, Control System for Automatic Automobile
`
`
`
`
`
`
`Driving, Proc. IFAC Tokyo Symposium on Sys-
`
`
`
`
`
`
`tems Engineering for Control System Design, ’65.
`K. Miki, T. Makino, M. Nagai, K. Hasegawa, Re-
`
`
`
`
`
`
`
`
`search on Automatic Vehicle Operation, Toyota
`
`
`
`
`
`
`
`
`
`
`Gijyutsu No. 20-1, '68, June.
`
`
`
`
`
`
`
`
`
`[3]
`
`
`[4]
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`H. M. Morrison, A. F. Welch. E. A. Hanysz, Auto-
`
`
`
`
`
`mated Highway and Driver Aid Developments,
`
`
`
`
`SAE preprint, ’60, March.
`T. Ito, M. Furumata, F. Harashima, H. Inaba, S.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Matsumoto, An Automatic Driving System of
`
`
`
`
`
`
`Automobiles by Guidance Cables, SAE Paper
`730127, ’73, Jan.
`
`
`
`
`
`
`
`
`
`
`
`T. Hirose: Study on Machine Vision of Intelligent
`
`
`
`
`
`
`Vehicle, Report of Mechanical Engineering Lab.
`No. 152 ’90, Nov.
`
`
`
`
`S. Takaba, N. Hashimoto, K. Sakai, N. Nakamura,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`K. Sugimoto: Development of the Traffic Flow
`
`
`
`
`
`
`
`
`Analysis System Using The VTR, The 16th Joint
`
`
`
`
`
`
`Conference on Image Technology, Tokyo,
`’85,
`Dec.
`
`In Tunnel Traffic Volume
`K. Sasaki et al:
`
`
`
`
`
`
`
`
`
`
`
`
`Measuring System Using
`an
`lTV Camera.
`Sumitomo Denki No. 134, ’89, March.
`
`
`
`
`
`
`
`
`
`
`
`
`Express Highway Research Foundation of Japan,
`
`
`
`
`
`
`
`SE-2000: Report on Basic Study of Automated
`
`
`
`
`Vehicle System, ’84, Feb.
`
`
`
`
`
`
`Ministry of Construction, RACS, ’90, May.
`
`
`
`
`
`
`
`1. Muramoto, S. Okabayashi, M. Sakata, M. Tsuji,
`
`
`
`
`Application Technology of Cooperative Laser
`
`
`
`
`
`Radar System for Automobiles on Expressways,
`
`
`
`
`
`
`EEE Tokyo Section, No. 26, '87.
`
`
`
`
`
`
`
`
`V. Graefe, K. D. Kuhnert. A High Speed Process-
`ing System Utilized in Autonomous Vehicle
`
`
`
`
`
`
`
`
`
`
`
`
`Guidance, IAPR Workshop CV, Tokyo, ’88, Oct.
`V. Graefe, Robot Vision and Autonomous Ve-
`
`
`
`
`
`
`
`
`
`
`
`hicles, Institute of Measurement Science, Faculty
`
`
`
`
`
`of Aeronautical and Space Engineering, Federal
`Armed Forces University Munich, ’9l.
`
`
`
`
`
`
`
`
`
`
`
`
`M. A. Turk, D. G. Morgenthaler, K. D. Gremban,
`M. Marra, VITS—A Vision System for Autono-
`
`
`
`
`
`
`
`
`
`
`
`
`mous Land Vehicle Navigation, IEEE Transactions
`
`
`
`
`
`
`
`on Pattern Analysis and Machine Intelligence, Vol.
`
`
`
`
`
`10, No. 3, ’88, May.
`C. Thorpe, M. H. Hebert, T. Kanade, S. A. Shafer,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Vision and Navigation for the Carnegie-Mellon
`Navlab, EEE Transactions on Pattern Analysis and
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Machine Intelligence, Vol. 10, No. 3, ’88, May.
`
`
`
`
`
`
`
`R. Stengel, A. Niehaus. Intelligent Guidance for
`
`
`
`
`
`Headway and Lane Control, Engineering Applica-
`
`
`
`
`
`
`
`
`tions of Artificial Intelligence, Vol. 2, No. 4, ’89,
`Dec.
`
`
`
`
`
`
`S.E. Shladover: Advanced Vehicle Control System
`
`
`
`
`
`(AVCS), SAE Paper 901129.
`
`
`
`
`
`
`
`R. D. Ervin: Proving Radical Functionality to
`
`
`
`
`
`
`
`Serve Highway Transportation: A 20 Year Vision
`
`
`
`
`
`
`for IVHS, SAE Paper 901125.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`9