`
`::>
`ffl
`
`011~1~47160
`
`PATENT DATE
`
`OB/247,760
`
`05/2:3/94
`
`280
`
`SUBCLASS
`
`S . m; FED,
`
`fO!O()NTON TOWNSH I , N.J •
`•
`
`·~DA'li[:t
`u
`:1
`~
`**CONTI NU I hE DATA*****·'~*** i **:*'i' * * * ** ,,, *
`VEFHFIED
`
`' .
`
`**I'OREifii'I/FCT APPLICATIONS************ 'I d~
`1/UUFIED
`
`J'>111"3J'b't/ 'i ~
`
`FOHEIGN FILINt:> LICENSE GRANTED 07J25/9A
`
`***** SMAU_ ENliTY •****
`
`Fort~lg,, priority clalmetl
`35 US(:: 119 concUtii.Jnl nl'let
`
`m-1! FKOV I TZ
`:;-<;•61
`P. 0. BOl:
`AHLINGiTON, 1/A ;2:2202
`
`--------~~~~----------~--~-----------------
`Ii41''At:T fHRBAl:i SYSTEI•1 IIliTH ANTICIPATORY SENSOR
`
`PJ'II'ITS OF A PPl.IC:ATION
`FILED SEPARATELY
`
`L;tbel
`J1.1rea
`
`fo<m PT0-4361'
`(!lev 8/92)
`
`Assistant Examiner
`
`Total Claims
`
`Sheets Drwg. Fi!IS. Drwg.
`
`Print Fig.
`
`ISSUE
`BATCH
`Examiner NUMBER
`
`WARNING: Th6 information disclosed herein may be restricted. Unauthorlze,f dl:;closure may be prohibited
`by the Untted States Code lltle 35, Sections 122, 181 ancl 368. P<•ssession outside tile U.S .
`. Patent & Trademark Office is restricted to authorl~8d emplc1yees ~nd contractc•rs only.
`
`/
`
`1'. ~E)
`
`IPR2013-00424 - Ex. 1010
`Toyota Motor Corp., Petitioner
`1
`
`
`
`Dale
`Ente~red
`or
`C•lUnted
`
`[late
`Rec:eived
`or
`Mailed
`
`, __ 30.
`___ 31. ________________________ _
`
`~----·---32. ----~-----------------
`(FRONT)
`
`·------
`
`2
`
`
`
`s EARCHED
`
`Glass
`
`,, b
`,)U -·
`
`Date
`
`Exmr.
`
`z..Kt?
`
`J LJ()
`
`7 35-
`'J ]0~
`·:>r,
`'I
`'f 3~ y
`'
`,?3
`
`·~ ' 7Y
`
`( I;/
`/t ~1
`
`;;t?
`
`,, J J•j7'i y?.{j
`
`I
`
`1~'}'1.) ~.a-
`
`'
`
`I
`
`SEARCH NO" TES
`
`0:, AJV I .Jc-.,(
`,Do·nV\; <:. C ~sf"'"' -1
`-tu z.6t7
`
`'3 Wj-
`
`l'rPS
`c~v _/\ -S' (.::= --~
`J'-I.Y' I -11!1111--W
`
`[ >ale
`1--
`
`Exmr..
`
`lfj
`I
`
`"
`
`INTERFERENCE SEARCHED
`Date
`Exmr.
`Class
`
`I ____ .,~_
`•
`
`(RIGHT OUTSIDE)
`
`3
`
`
`
`Staple Issue Slip Here
`
`'
`
`IDNO.
`.,..-/
`~ l)L
`-::;;-;:;;,
`'
`
`DATE
`~-0'=?(
`fo -/in--'fl.
`-'"j
`"-,
`v /L;
`I
`
`' / I
`
`POSITION
`CLASSIFIER
`EXAMINER
`TYPIST
`VERIFIER
`CORPSCORR.
`SPEC. HAND
`FILE MAINT.
`DRAFTING
`
`Oat·!
`
`-------
`- -
`
`------I-- - -
`- - - 1----
`- -
`- - 1----
`-- 1 - - -
`---
`- - I-- -
`---- --r-- -
`----r-- --
`. l-r-- r--
`r-- f - -
`r--'- --
`----- -- -
`- f-- ----- -
`- r- ------- -
`--~ -- f--
`-+--- -
`- -I --- -
`r- --r-- '-- --
`--r-- -- --
`---r-- r-- -
`-- r--r-- --
`~-r-- ~- --
`--
`
`--
`
`-
`r-- 1--- --
`-r-- 1--- --
`--r- --r-- 1--- --
`---
`--r- --· r-- '-- · -
`--1- -- r---- f -
`-
`f -
`--r- -- 1----- 1 -
`-- '-- -- 1 -
`r-- --1--- 1---
`r-
`1-'
`- f -
`f -
`. --- -- 1-
`-- -
`-- 1---
`1 -
`-~ -- 1 -
`-
`
`-- I-
`-- I-
`
`--
`
`-- I -----
`-'-
`-- 1------
`- -
`-
`'-- 1------
`- -
`
`INDEX OF CLAIMS
`
`,-----,--------------,
`Date
`
`3
`__ t~ . ~-+-+t--+-+-+-t---1f---l
`3
`
`SYMBOLS
`. ... Rejecllld
`../ ............
`.. ................................. Allowed
`• (Through numbenl) Canceled
`+ ....
`. ..... Restricted
`N ....
`. ..... Non·elecled
`I ........
`. ....... Interference
`A .........
`. ..... Appeal
`0 ......
`. ..... Objected
`
`4
`--,4' -+H-+-+-+-+-t---1'--1
`--14 --H-HI--1--t-t-t-+---1
`-181 ,.
`r-ilr~~'-~+4'+-t~~+--+~--1
`'~U-~~~--1---1-~~~--1
`47
`48 --+--+--+--+-l-t--t--1
`-
`-- 49 --+---t--+-l-+--t--1---1
`--- 50 --+--+-+-+~-+-+-!
`- - ___ __j__J_L-L-L-l-_.J___j
`
`(LEFT INSIDE)
`
`Claim
`"'
`"' "
`" ·c;,
`u:
`·c
`0
`51
`52
`53
`54
`55
`56
`57
`58
`59
`60
`61
`62
`63
`64
`65
`66
`67
`68
`69
`70
`71
`72
`73
`74
`75
`76
`77
`78
`79
`80
`81
`82
`83
`84
`85
`86
`87
`88
`89
`90
`91
`92
`93
`94
`95
`96
`97
`98
`99
`100
`
`4
`
`
`
`PATENT APPLICATION SERIAL NJl~ 2417760
`
`U.S. DEPARTMENT OF COMMERCE
`PATENT AND TRADEMARK OFFICE
`FEE RECORD SHEET
`
`·160 AA (16/06194 08247760
`
`1 101
`
`-786.00 CK All 78 ·
`
`PT0-15.56
`(5/8 7)
`
`5
`
`
`
`EiAR CODE LiW'i::---------~--~------------------,··---------------~
`
`1111111111111
`
`11111111
`
`U.S. PATENT APPLICATION
`
`SERIAL NUMBEFI
`
`08/247,760
`
`FILING DATE
`
`CLASS
`
`05/23/94
`
`280
`
`31)6
`
`DAVID S. BREED, BOONTON TOWNSHI, NJ.
`
`**CONTINUING DATA*********************
`VERIFIED
`
`**FOREIGN/PCT APPLICATIONS************
`. VERIFIED
`
`FOro~IGN FILING LICENSE GRANTED 07/25/94
`~:':":"=:
`HATE OR
`COUNTRY
`
`SHEETS
`DRAWING
`
`TOTAL
`CLAIMS
`
`INDEPENDENT
`CLAIMS
`
`***** SMALL EN'l'ITY *"'***
`ATTOI INEY DC CKET NO.
`FILING FEE
`RECEIVED
`
`NJ
`
`18
`
`46
`
`9
`
`$786.00
`
`ATl 78
`
`S~~UEL SHIPKOVITZ
`P.O. BOX 2961
`AR~INGTON, VA 22202
`
`--------------------------------------------------~1
`SIDE IMPACT AIRBAG SYSTEM WITH ANTICIPATORY SENSOR
`
`,, ,,
`.u '" '" ~
`
`'u .J
`1-
`i=
`
`This is to certify that annexed hereto is a true copy from the records of the United Staten
`Patent and Tmdemark Office of the application wh1ch is identified above.
`By authority of the
`COMMISSIONI:R OF PATENTS AND TRADEMARKS
`
`Dete
`
`Certifying Officer
`~------------~~~------------------·-----------~
`
`6
`
`
`
`f)_~~ (l Co 0
`SIDE IMPACT AIRBAG SYSTEM WITH ANTICIPATORY SENSOR
`
`ABSTRACT
`
`This invention is a system to identify and monitor objects, such as 1ellicles, 1rees,
`
`motorcycles or walls, which could impact with a motor vehicle, such as an auto nc bile or truck,
`
`by processing the signal received from the objects using one or more techniques, inGiuding
`
`neural networks or other pattern recognition systems, and technologies includino ultrasonic and
`
`electromagnetic radiation. The received signal may be a reflection of a transmitiE·d signal, the
`
`reflection of some natural signal from the object, or may be some signal ·emittecl naturally by
`
`the object.
`
`Information obtained by the identification and monitoring system s then used to
`
`deploy an airbag usually prior to the impact of the object with the vehicle.
`
`In the case of side
`
`impacts. the anticipatory sensor permits the_ occupant to be moved away from the impact and the
`
`deployment of a large alrbag to offer a level of protection heretofore not availabl '·
`
`28
`
`7
`
`
`
`r··
`
`SIDE IMPACT AIRBAG SYSTEM WITH ANTICIPATORY SENSOR
`----··-
`
`BACKGROUND OF THE INVENTION
`
`.A
`[18 2'47760
`
`Frontal impacts are the number one killer of vehicle occupants in auto motile accidents
`
`with about 16,000 fatalities each year. Side impacts are the second cause of au omobile related
`
`deatl1s with about 8,000 fatalities each year. The number of fatalities in frontal irr<pacts is now
`
`decmasing due to the introduction of airbags and mandatory seatbelt use laws.
`
`I i; natural now
`
`that a considerable effort be applied to saving lives in side impacts.
`
`SrlVeral automobile manufacturers are now considering the use of side inpact airbags to
`
`attempt to reduce the number of people killed or injured in side impacts. ThE side impact
`
`9roblem is considerably more difficult to solve in this way than the frontal imp<tct problem due
`
`to the lack of space between the occupant and the side door and to the significanr ir trusion which
`
`typically accompanies a side impact.
`
`Some understanding of the severity of the side impact problem can bu c btained by a
`
`comparison with frontal impacts.
`
`In the Federal Motor Vehicle Safety Standard FMVSS) 20B
`
`49 k.ph crash test which applies to frontal impacts, the driver, if unrestrained, will impact the
`
`steering wheel at about 30 kph. With an airbag and a typical energy absorbing ;tearing column,
`
`them is about 40 to 50 em of combined deflection of the airbag and steering column to absorb
`
`this 30 kph difference in relative velocity between the driver and vehicle inter lor. Also there
`
`is usuall)• little intrusion into the passenger compartment to reduce this available 1:pace.
`
`In the FMVSS 214 standard crash for side impacts, the occupant, whether restrai:ned or
`
`not, is impacted by the int(uding vehicle door also at about 30 kph.
`
`In this case there is only
`
`about 1 0 to 15 em of space available for an airbag to absorb the relative velodty between the
`
`occupant and the vehicle interior.
`
`In addition, the human body is more vulnerable to side
`
`1
`
`8
`
`
`
`impaets than frontal impacts and there is usually significant intrusion into tho passenger
`
`compartment. A more detailed discussion of side impacts can be found in a paper by Breed et al,
`
`"Sensin(t Side Impacts", Society of Automotive Engineers No. 940651, 1994, wl1ic1 is included
`
`herein by reference.
`
`Ideally an airbag for side impact protection would displace the occupant away from the
`
`intrudin(l vehicle door in an accident and create the required space for a S'Jfficiently large
`
`airbag. Sensors now being used for side impact airbags, however, begin sensing thE> crash at the
`
`beginning of the impact at which time there is insufficient time remaining to mo\ e the occupant
`
`before he is impacted by the intruding door. Even if the airbag were inflated lnstEntaneousty it
`
`is not possible to move the occupant to create the desired space without cau:;ing seriously injury.
`
`The problem is that the sensor which starts sensing the crash when the impact has betJUn, is
`
`already too late.
`
`There has been discussion over the years in the safety community about the use of
`
`anticipatory sensors so that the side impact accident could be sensed bofore it occurs.
`
`Heretofore this has not been practical due· to the inability to predict the severity ot the accident
`
`prior to the Impact. A heavy truck, for example, or a tree is a much more se11ere accident at
`
`low veloc:ity than a light vehicle or motorcycle at high velocity. Until now it ~as not been
`
`possiblE> to differentiate between these different accidents with a high degree of certainty.
`
`Once a sufficiently large airbag is deployed in a side impact and the dri•ter displaced
`
`awa'l frorn the door and the steering wheel, he will no longer be able to cont·,ol tl1e vehicle which
`
`could in itself cause a serious accident.
`
`It is critically important, therefore, that such an airbag
`
`not be deployed unless there is great certainty that the driver would otherwise be seriously
`
`injured or killed by the side impact. Anticipatory sensors have heretofore nc t been used
`
`becaus•3 of their inability to predict the severity of the accident. The pres•3nt nvention solves
`
`this problem and therefore makes anticipatory sensing practical. This permits side impact
`
`2
`
`9
`
`
`
`airbag systems which can save a significant percentage of the people who would otherwise be
`
`killed as well as significantly reducing
`
`the number and severity of injures. This is
`
`accomplished through the use of pattern recognition technologies such as neural networks such
`,...J/1-v 61 as discussed in co-pending patent application attorney docket number ATI·7~~iled May 9th,
`13i
`
`1994.
`
`Nr9ural Networks are capable of pattern recognition with a speed, accuracy ard efficiency
`
`heretofore not possible.
`
`It is now possible, for example, to recognize that the front of a truck or
`
`anothor Gar is about to impact the side of a vehicle when it is one to three meters :)r more away.
`
`This totally changes the side impact strategy since there is now time to inflate a la,-ge airba~J and
`
`push the occupant out of the way of the soon to be intruding vehicle. Natural'y not all side
`
`impacts are of sufficient severity to warrant this action and therefore there wil usually be a
`
`dual inflation system as described in more detail below.
`
`Although the main application for anticipatory sensors is in side impacts, Ire ntal impact
`
`anticipatory sensors can also be used to identify the impacting object before the cr3sh occurs.
`
`Prior to going to a full frontal impact anticipatory sensor system, neural network! c3n be used
`
`to detHct many frontal impacts using data in addition to the output of the normal crash sensing
`
`accelerometer. Simple radar or acoustic imaging, for example, can be add9d to current
`
`accelerometer based systems to give substantially more information about the crash and the
`
`impacting object than possible from the acceleration signal alone.
`
`The side impact anticipatory sensor of this invention can use any of a variety of
`
`technologies including optical, radar, acoustical, infrared or a combination of these. The sensor
`
`system typically contains a neural network processor to make the discrimination however a
`
`simulated neural network, a fuzzy logic or other algorithm operating on a microprocessor can
`
`also be usr3d.
`
`3
`
`10
`
`
`
`SUMMAFIY AND OBJECTS OF THE INVENTION
`
`This invention comprises an anticipatory sensor system which uses li) a source of
`
`radiant emergy either originating from or reflected off of an object or vehiclo wt ich is about to
`
`impaet tho side of a target vehicle, plus (ii) pattern recognition means to analrze the radiant
`
`energy c:oming from the impacting object or vehicle to (iii) assess the prohabl•l !'everit~· of a
`
`pending accident and (iv) if appropriate, inflate an airbag prior to the impact so a!; to displace
`
`the oGcupant away from the path of the impacting object or vehicle to create' sp.lCH required to
`
`cushion the occupant from an impact with the vehicle interior. Although the primary area of
`
`application of this invention is for protection in side impacts, the invention also provides added
`
`proteetion in frontal impacts by reducing the incidence of injury to out-ot-posi·ion occupants
`
`by permitting a slower inflation of the airbag and displacing the occupant away f·orl the airbag
`
`prior to the impact.
`
`The principal objects and advantages of this invention are:
`
`1 .
`
`To provide tor the enhanced protection of occupants in side impacts by clet.Jrmining the
`
`pmbable severity of a pending accident and inflating an airbag prior to the impact to
`
`displace the occupant away from the vehicle door.
`
`2.
`
`To provide for a method of identifying and classifying an object which is about to impact
`
`a vehicle.
`
`3.
`
`To adapt pattern recognition techniques, and particularly neural networks to permit the
`
`identification of objects external to an automotive vehicle and the determ nation of their
`
`approach speed and angle of potential collision.
`
`4.
`
`To provide a method for assessing the probable severity of a pending acci je 1t base1d on
`
`the identification of the class of an object which is about to impact thn vehicle plus
`
`stmed information about the class of such objects such as its mass, strength and
`
`att€tchment
`
`to the earth.
`
`4
`
`11
`
`
`
`5.
`
`To provide a method using an ultrasonic system for use in illuminating an •Jbject which
`
`is about to impact a vehicle and using the reflection of the ultrasonic illumination in
`
`combination with a pattern recognition system to identify the object.
`
`To determine the approach velocity of an object which is about to impact a VE•hicle.
`
`To identify that a truck is about to impact a vehicle.
`
`To identify that an automobile is about to impact a vehicle.
`
`To identify that a vehicle is about to impact with a tree.
`
`6.
`
`7.
`
`8.
`
`9.
`
`1 0.
`
`To provide a method using an electromagnetic wave system for use in illuminating an
`
`object which is about to impact a vehicle and using the reflection of the E lectroma!lnetic
`
`wave illumination in combination with a pattern recognition syste•m lo identify the
`
`object.
`
`11 .
`
`To provide a method using an the passive infrared electromagnetic waves r<,diatin~l from
`
`an object such as a motor vehicle in combination with a pattern recognition sys!llm to
`
`identify the object.
`
`1 2.
`
`To provide a system for identifying an object which is about to impact a vehicle• in a
`
`substantially side impact.
`
`13.
`
`To provide a system for identifying an object which is about to impact a vehicle• in a
`
`substantially frontal impact.
`
`1 4.
`
`To provide a system comprising a variable inflation airbag system where the control of
`
`the inflation of the airbag is determined by an prediction of the probable severity of an
`
`accident prior to the accident occurring.
`
`1 5.
`
`T<) provide apparatus for inducing slack into a seatbelt in the event of a sicle impact to
`
`pllrmit the occupant to be displaced sideways in the vehicle.
`
`1 6.
`
`To provide for a single airbag module for protection of the head and torso Jf 1n occupant
`
`in side impacts.
`
`5
`
`. -....
`
`12
`
`
`
`1 7.
`
`To provide a single airbag module for mounting in the seat back of a ve~icle for the
`
`protection of the head and torso of an occupant in side impacts.
`
`18.
`
`To provide a structure and method for moving the occupant and his seat in the event of a
`
`side impact accident to increase the space between the occupant and the intruding object.
`
`1 9.
`
`To provide for an airbag to be deployed external to the vehicle in conjL nction with an
`
`anticipatory sensor in side impacts.
`
`20.
`
`To provide a method using an ultrasonic wave system for use in illuminating an object
`
`which is about to impact a vehicle and using the reflection of the ultrasonic wave
`
`illumination in combination with a pattern recognition system to idemify lhe object
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`FIG. v·f~·an overhead view of a vehicle about to be impacted in the side by an approaching
`
`vehicle showing a wave transmitter part of the anticipatory sensor system.
`
`FIG. 1p.·1s a perspective view of a vehicle about to impact the side of itncther vtshicle
`
`/
`showing the location of the various parts of the anticipatory sensor system of thi~ invention.
`FIG. ,2·· is an overhead view of a vehicle about to be impacted in the front by an
`
`approaching vehicle showing a wave transmitter part of the anticipatory sensor s•1stsm.
`
`FIG. 3A a plane frontal view of the front of a car showing the headlight~, radiator grill,
`l
`bump1sr, lenders, windshield, roof and hood.
`,,
`FIG. ~Er' a plane frontal view of the front of a truck showing the headligttts, radiator
`
`grill, bumper, fenders, windshield, roof and hood.
`
`FIG. ~·Is an overhead view of a vehicle about to be impacted in the side by itn 1pproaching
`
`vehicl's showing an infrared radiation emanating from the front of the striking ~ehicle and an
`
`infraretd re-ceiver part of the anticipatory sensor system.
`
`'
`FIG. ~Vis a side plane view with portions cutaway and removed of a dual inlator airbag
`
`6
`
`13
`
`
`
`system with two airbags with one airbag lying inside the other.
`
`FIG. 6/(s a perspective view of a seatbelt mechanism illustrating a device to release a
`
`/
`
`controlled amount of slack into seatbelt allowing an occupant to be displaced.
`/
`FIG. 7' is a frontal view of an occupant being restrained by a soatbelt having two
`
`anchora(i€' points on the driver's right side where the one is released allowing the occupant to be
`
`laterally displaced during the crash.
`/
`1=1G. ?rA is an expanded view of the release mechanism within the circle 7A of FIG. 7.
`
`FIG. 78 is a view of the apparatus of FIG 7A within the circle 78 and rota:ec 90 degrees
`
`showing the release mechanism.
`
`FIG. ~'is a frontal view of an occupant being restrained by a seatbelt Integ ·al with seat so
`
`that whE>n seat moves during a crash with the occupant, the belt also moves .IIIowing the
`
`occupant to be laterally displaced during the crash.
`
`FIG. ~I( is a frontal view of an occupant being restrained by a seatbelt and a linear airbag
`
`module (930) attached to seat back to protect entire occupant from his pelvis to his head.
`
`FIG. ~B/is a view of the system of FIG. 9A showing the airbag in the inllatej condition.
`
`FIG. 1.0K'is a frontal view of an oc"cupant being restrained by a seatbelt and where the
`
`seat is displaced toward vehicle center by deploying airbag in conjunction with ot'1er apparatus.
`
`FIG. 1.0H is a frontal view of an occupant being restrained by a seatbelt and where the
`
`seat is rotated about vertical axis· in conjunction with other apparatus.
`
`FIG. 1 oc··is a frontal view of an occupant being restrained by a seatiJelt and where the
`
`seat is rotated about longitudinal axis in conjunction with other apparatus.
`
`FIG. UA is a perspective view with portions cutaway and removed of a V•lhicle about to
`
`impact the side of another vehicle showing an airbag stored within the side doc r of the target
`
`vehicle prior to being released to cushion the Impact of the two vehicles.
`
`FIG. 11 B is a view of the apparatus of FIG. 11 A after the airbag has deploy£ d.
`
`7
`
`14
`
`
`
`DESCRIPTION OF THE PREFERRED EMBODIMENTS
`
`In FIG. 1 an overhead view of a vehicle 110 about to be impacted in lhe side by an
`
`approaching vehicle 120 where vehicle 110 is equipped with an anticipatory se11sor system
`
`showing a transmitter 130 transmitting waves 132 toward vehicle 120. A per spec live view of
`
`vehicle 110 is shown in FIG. 1A and illustrates the transmitter 130 connocted to electronic
`
`module 140. Module 140 contains circuitry 142 to drive transmitter 130 and circuitrl' 144
`
`to process the returned signals from receivers 134 and 136. Circuitry 144 cor,tams a neural
`
`computer 145 which performs the pattern recognition determination based on ! ignals from
`
`receivers 134 and 136.
`
`Throughout the description herein, the term "approaching" when usetd as an objoct or
`
`vehicle approaching another will mean the relative motion of the object toward the vehicle
`
`havlnq thet anticipatory sensor system. Thus, in a side impact with a tre<3, the tree will be
`
`considemd as approaching the side of the vehicle and impacting the vehicle.
`
`In other words, the
`
`coordinate system used in general will be a coordinate system residing in the !argot l'ehicle. The
`
`"target" vethicle is the vehicle which is being impacted. This convention perrnitl a gE>neral
`
`description to cover all of the cases such as where {i) a moving vehicle impacts into I he side of a
`
`stationary vehicle, {ii) where both vehicles are moving when they impac1, 01 {iii) where a
`
`vehicle is moving sideways into a stationary vehicle, tree or wall.
`
`In a preferred implementation, transmitter 130 is an ultrasonic transmitter operating
`
`at a frequency of approximately 40 khz, although other frequencies could be Ul,ec. Similarly,
`
`receivers 134 and 136 are ultrasonic transducers and receive the reflected ult·a!onic waves
`
`from vehicle 120.
`
`A detailed discussion of pattern recognition technology as applied to the monitorin(J and
`
`8
`
`15
`
`
`
`identification of occupants and objects within a vehicle is discussed in detail in 3wed et al co-
`
`__.t,._,...11~z... pending US Patent application attorney docket number ATI-7~iled May 9, 199~ and included
`l>'l-
`herein by reference. Although the application herein is for the identification of •)bjects exterior
`
`to the• vehicle, many of the same technologies, principles and techniques are appli~al)le.
`
`Jl,n example of such a pattern recognition system using neural networks u~ing sonar is
`
`discussed in two papers by Gorman, R,. P. and Sejnowski, T. J. " Analysis of HiddE•n Units in a
`
`Layered Network Trained to Classify Sonar Targets", Neural Networks, Vol.1. pp ''5-89, 1988,
`
`and "Learned Classification of Sonar Targets Using a Massively Parall•~l f\eN1ork", IEEE
`
`Transactions on Acoustics, Speech, and Signal Processing, Vol. 36, No. 7, July H88.
`
`"Pattern recognition" as used herein will mean any system which process£"s 3 signal that
`
`is generated by an object, or is modified by interacting with an object, in ordEr to determine
`
`which one of a set of classes that the object belongs to. Such a system might determine only that
`
`the object is or is not a member of one specified class, or it might attempt to assi{ n ·he object to
`
`one of a larger set of specified classes, or find that it is not a member of any of thE· classes in the
`
`set. The• signals processed are generally electrical signals coming from transdu.;er s which are
`
`sensitive to either acoustic or electromagnetic radiation and if electromagneti•: they can be
`
`either visible light, infrared, ultraviolet or radar.
`
`"To identify" as used herein will mean to determine that the object bolongs to a
`
`particular set or class. The class may be one containing all trucks of a certain ;iz-3 or weight,
`
`one containing all trees, or all walls.
`
`In the case where a particular vehicle type is to be
`
`recogniz!ld, the set or class will contain only a single element, the particular vehicl•l type to be
`
`recognized.
`
`Some examples follow:
`
`In a passive infrared system, a detector receives infrared radiation from f.n object in its
`
`field of view, in this case the approaching object is most likely another vehicle, and processes
`
`9
`
`16
`
`
`
`the rec11ived infrared radiation radiating from the vehicle's engine compartnent. The
`
`anticipatory sensor system then processes the received radiation pattern to determine the class
`
`of vehicle, and, along with velocity information from another source, makes an assessm.ent of
`
`the probable severity of the pending accident and determines if deployment nf 1n airbag is
`
`required. This technology can provide input data to a pattern recognition system but it has
`
`limitations related to temperature. The sensing of a non-vehicle object such a:; a tree, for
`
`example, poses a particular problem. The technology may also fail to detect a ve 1icle which has
`
`just been started especially if the ambient temperature is high. Nevertheloss, fo· use in the
`
`identification of approaching vehicles the technology can provide importa 1t information
`
`espec:iallv if it is used to confirm the results from another sensor system.
`
`In a laser optical system an infrared laser beam 132 is used to morn en ar ly illuminate
`
`an object .as illustrated in FIG. 1 where transducer 132 is a laser transmitter.
`
`In scme cases a
`
`charge coupled device (a type of TV camera) is used to receive the reflected ligh a 1d would be
`
`used as one or both of the receivers 132 and 134. The laser can either be usee! ir• a scanning
`
`mode. or, through the use of a lens, a cone of light can be created which covers a ar!)e portion of
`
`the object.
`
`In each case a pattern recognition system, as defined above, is used to identify and
`
`classify the illuminated object and its constituent parts. This system provides the most
`
`information about the object and at a rapid data rate.
`
`Its main drawback is cc·st which is
`
`considerably above that of ultrasonic or passive infrared systems and the atlfm ation which
`
`results in bad weather conditions such as heavy rain, fog or snow storms. As the cost of lasers
`
`comes down in the future, this system will become more competitive. The attem1ation problem
`
`is not as severe as might be expected since the primary distance of concern br anticipatory
`
`sensors as described here is usually less than three meters.
`
`Radar systems have similar properties to the laser system discussE!d abo>~e with the
`
`advanta~tl! that there is less attenuation in bad weather. The wave length ot a r·ar icular radar
`
`1 0
`
`17
`
`
`
`system can limit the ability of the pattern recognition system to detect objeGt features smaller
`
`than a certain size. This can have an effect in the ability of the system to identify different
`
`objects ilnd particularly to differentiate between different truck and automobile models.
`
`Tho ultrasonic system is the least expensive and potentially provide·s less information
`
`than tl'le laser or radar systems due to the delays resulting from the speed of sound and due to the
`
`wave length which is considerably longer than the laser systems. The wave leng!h limits the
`
`detail which can be seen by the system.
`
`In spite of these limitations, as shown in the above
`
`referencE!d Breed et al patent application attorney docket ATI-77~ultrasonic> ,;an provide
`.6._3
`sufficient timely information to permit the position and velocity of an approachin::J object to be
`
`accurately known and, when used with an appropriate pattern recognition system, it is capable
`
`of positively determining the class of the approachillg object. One such pattErn recognition
`
`system usos neural networks and is similar to that described in the papers by Gorman et al and
`
`in the re1ar facing child seat recognition system referenced and described in the Breed et al
`
`patent application referenced above.
`
`A focusing system, such as used on some camera systems, could be usod tc d1Jtermine the
`
`position of an approaching vehicle when it is at a significant distance away but is too slow to
`
`monitor this position just prior to a crash. This is a result of the mechanical m(ltic ns required
`
`to opE!rate the lens focusing system. By itself it cannot determine the class of t~e 1pproaching
`
`object but when used with a charge coupled device plus infrared illumination fer night vision,
`
`and an appropriate pattern recognition system, this becomes possible.
`
`From the above discussion, it can be seen that the addition of sophistic~ ted pattern
`
`recognition means to any of the standard illumination and/or reception technologir s 'or use in a
`
`motor vehiGie permits the development of an anticipatory sensor system which cE n dentif~· and
`
`classify an object prior to the actual impact with the vehicle.
`
`The, application of anticipatory sensors to frontal impact protection sy:;terr s s shown in
`
`1 1
`
`18
`
`
`
`FIG. 2 which is an overhead view of a vehicle 110 about to be impacted in the front by an
`
`approaching vehicle 120.
`
`In a similar manner as in FIG. 1, a transmitter H>O liansmits waves
`
`162 toward vehicle 120. These waves are reflected off of vehicle 120 and receivEld by
`
`receiving transducers 164 and 166.
`
`FIG. 3A illustrates the front of an automobile 310 and Shows prefrmed IJcations for
`
`transmitting transducer 160 and receiving transducers 164 and 166. FIG. 3A also illustrates
`
`the distinctive features of the vehicle which cause a distinct pattern of reflocted v1aves which
`
`will differ from that of a truck 320, for example, as shown in FIG. 38.
`
`In some pattern
`
`recognition technologies, the researcher must determine the distinctive features of each object
`
`to be recognized and form rules which permit the system to recognize one object !rom another of
`
`a diffment class. An alternative method is to use artificial neural network t•3ChrrolrJgy wherein
`
`the identification system is trained to recognize different classes of objects.
`
`n this case a
`
`training session is conducted where the network is presented with a variety of objects and told
`
`to which class each object belongs. The network then learns from the trainin(l session and,
`
`providing a sufficient number and diversity of training examples are available, he network is
`
`able tr) categorize other objects which have some differences from those making up the training
`
`set of objects. The system Is quite robust in that it can still recognize objects as belongin~l to a
`
`particular c:lass even when there are significant differences between the object to be recognized
`
`and the objects on which the system was trained.
`
`Once a neural network has been sufficiently trained, it is possible to analy ~e the network
`
`and determine the "rules" which the network evolved. These rules can then so netimos be
`
`simplil'ied or generalized and programmed as a fuzzy logic algorithm. AlternEtely, a neural
`
`computer can be programmed and the system implemented on a semiconductor chip 1s available
`
`from IIIIo to rola.
`
`The• anticipatory sensor system must also be able to determine the distarlcE•, approach
`
`1 2
`
`19
`
`
`
`velocity and trajectory of the impacting object in addition to the class of object! to which it
`
`belongs.. This is easily done with acoustic systems since the time requirE1d for 'he acoustic
`
`waves to travel to the object and back determined its distance based on the sp13ed Jf sound. With
`
`radar and laser systems, the waves usually· need to be modulated and the phasE cl1ange of the
`
`modulation determined in order to determine the distance to the object as discus >ed in more
`
`detail in US co-pending patent application 08/040,978 filed 3/31/93 to BrE•ecl et al and
`
`included herein by reference. Since the same distance measurement techniques are used hE!re as
`
`in the two above referenced patent applications, they will not be repeated hem.
`
`There is a radar chip on the market which permits the distance determinati<1n based on
`
`the time mquired for the radar waves to travel to the object and back. T~.is tnctmology was
`
`developEld by Amerigon Inc. of Burbank California and is being considered fer ot1e1 automotive
`
`applications such as constant distance cruise control systems and backing-up warnir g systems.
`
`FIG. 3A is a plane frontal view of the front of a car showing the tleaolig~ts, radiator
`
`grill, bumper, fenders, windshield, roof and hood and other objects which ··eflect a particular
`
`pattern of waves whether acoustic or electromagnetic. Similarly, FIG. 3B is " ~lane frontal
`
`view of th11 front of a truck showing the headlights, radiator grill, bumper, fenders, windshield,
`
`roof a.nd hood illustrating a significantly different pattern. Neural