`Farrall
`
`I Ill lllllll Ill lllll lllll lllll lllll lllll lllll lllll lllll 111111111111111111
`US005656921A
`[111 Patent Number:
`[451 Date of Patent:
`
`5,656,921
`Aug. 12, 1997
`
`[54] CONTROL OF A VEHICLE POWERTRAIN
`
`[75]
`
`Inventor: Simon David Farrall, Coventry,
`England
`
`[73] Assignee: Rover Group Limited, Birmingham,
`England
`
`[21] Appl. No.: 443,583
`May 17, 1995
`
`[22] Filed:
`
`[30]
`
`Foreign Application Priority Data
`
`[GB] United Kingdom ................... 9410389
`
`May 24, 1994
`Int. Cl. 6
`........................................................ B02P 9/04
`[51]
`[52] U.S. CI . ................................... 322/40; 32'l/9; 32'l/38;
`180/65.2
`[58] Field of Search ................................ 32'l/9; 180/65.2,
`180/165; 290/16; 60/716, 718
`
`[56]
`
`References Cited
`
`U.S. PATENT DOCUMENTS
`
`8/1981 Kemper ................................... 180/165
`4,282,947
`6/1982 Kawakatsu .............................. 364/424
`4,335,429
`4,407,132 10/1983 Kawakatsu et al ....................... 60n16
`4,533,011
`8/1985 Heidemeyer et al .................. 180/65.2
`
`5,081,365
`5,343,970
`5,389,050
`5,390,117
`5,487,007
`
`1/1992 Field et al ................................. 290/45
`9/1994 Severinsky ............................. 180/65.2
`2/1995 Sakai et al ................................ 477n8
`2/1995 Graf et al ............................ 364/424.1
`1/1996 Suzuki et al ....................... 364/424.05
`
`FOREIGN PATENT DOCUMENTS
`
`0570242 11/1993 European Pat. Off ..
`2267364 12/1993 United Kingdom .
`
`Primary Examiner-Steven L. Stephan
`Assistant Examiner-Nicholas Ponomarenko
`Attorney, Agent, or Firm-Davis and Bujold
`ABSTRACT
`
`[57]
`
`A controller 20 for a vehicle powertrain, the powertrain
`comprising an electric motor 16 powered by a battery and an
`engine 14 powered by fossil fuel. The engine 14 is typically
`an internal combustion petrol engine. The controller 20 is
`operative to use maps stored in memory to control the torque
`contributions from the engine and the motor so as to
`maintain a de.sired relationship between the contributions of
`the electric motor 16 and the engine 14 towards the pow(cid:173)
`ering of the vehicle, and the controller is adaptive in that it
`is arranged to adapt said maps by altering one or more values
`stored therein to obtain a desired relationship.
`
`24 Claims, 4 Drawing Sheets
`
`15 FUEL FLOW RATE
`
`VOLTAGE FROM
`ACCELERATOR PEDAL 10
`12
`
`------
`
`POWERTRAIN
`CONTROLLER
`
`14
`
`16
`
`ENGINE
`
`ENGINE
`OUTPUT TORQUE
`DRIVES HAFT
`TORQUE
`
`18
`
`TRANSMISSION___.. ___
`
`ENGINE SPEED
`13
`
`MOTOR DEMAND
`
`MOTOR
`
`MOTOR
`OUTPUT TORQUE
`
`17 BATTERY CURRENT
`
`BMW1044
`Page 1 of 10
`
`
`
`U.S. Patent
`
`Aug. 12, 1997
`
`Sheet 1 of 4
`
`5,656,921
`
`15 FUEL FLOW RATE
`
`VOLTAGE FROM
`ACCELERATOR PEDAL 10
`12
`
`POWERTRAIN
`CONTROLLER
`
`ENGINE
`
`14
`
`16
`
`ENGINE
`OUTPUT TORQUE
`DRIVESHAF1
`TORQUE
`
`18
`
`TRANSMISSION...___.._
`
`ENGINE SPEED
`13
`
`FlG.1
`
`MOTOR DEMAND
`
`MOTOR
`
`MOTOR
`OUTPUT TORQUE
`
`17 BATTERY CURRENT
`
`y
`
`ENGINE SPEED CONSTANT
`
`DEMANDED
`ARMATURE
`CURRENT
`
`0
`
`DEMANDED THROTTLE ANGLE
`(DEGREES)
`
`FIG.3
`
`X
`
`90
`
`BMW1044
`Page 2 of 10
`
`
`
`U.S. Patent
`
`Aug. 12, 1997
`
`Sheet 2 of 4
`
`5,656,921
`
`DEMANDED
`THROTTLE 60
`ANGLE '-..
`(DEGREES) 30
`
`· · · · ·.. :
`:
`
`0
`7000
`Y 5000
`
`3000
`
`ENGINE
`)
`SPEED
`(R.P .M)
`
`1000
`F1G.2A
`
`....
`
`.. ···.
`
`90
`
`60 X
`' - - PEDAL VALUE
`(DEGREES)
`
`DEMANDED
`.... --:· · ·
`_40 ..... · · · · t ..
`MOTOR
`ARMATURE
`CURRENT'-..
`(AMPERES) - ZO
`
`...... .
`
`.. :- . -
`
`... : .
`.
`
`.
`
`. . . : . .. ·.
`
`. .
`
`.
`
`.
`
`..
`
`·.• .• .
`
`..
`...
`. ·. · ..
`··:- .
`
`. .. ·.
`
`90
`
`60 X
`' - - PEDAL VALUE
`(DEGREES)
`
`0
`7000
`Y 5000
`
`3000
`
`ENGINE
`)
`SPEED
`(R.P .M)
`
`1000
`FIG.28
`
`BMW1044
`Page 3 of 10
`
`
`
`PEDAL
`INPUT
`
`12
`
`r--------------L20
`13
`POWERTRAIN SPEED
`1
`PERFORMANCE --.-1~~~~-,
`MEASUREMENT
`BATIERY CURRENT
`
`.----~--- r 24 - - - -22--J FLciJE~ATEI
`I THROTILE I
`I 15
`
`RULE
`MODIFIER
`....__--.--r---' : DE~AND -.. ENGINE
`.----~'-----. I
`
`I
`
`1~17
`
`ENGINE
`OUTPUT TORQUE
`
`I
`FUZZY
`I CONTROLLER ....... I..___,,
`I
`I
`L _____ _J
`26
`ARMATURE
`CURRENT DEMAND
`
`MOTOR
`
`14
`
`16
`
`FIG.4
`
`TRANSMISSION L DRIVESHAFT
`
`TORQUE
`
`.--...i
`
`MOTOR
`OUTPUT TORQUE
`
`18
`
`e •
`00
`•
`~
`"""" a
`
`>
`~
`
`~
`"'N
`~
`
`~
`---1
`
`~
`m.
`w
`~
`.a:..
`
`Ot
`-..
`~
`Ot
`~ -..
`\0
`N
`~
`
`BMW1044
`Page 4 of 10
`
`
`
`O
`
`y
`
`INSTANTANEOUS
`PERFORMANCE
`MEASURE
`
`-0.2, .....
`
`'.' '""lTIIU I I 1mr1mr m 11111i1111rn 111 I WIIIM ~
`,.,.
`
`I
`
`U\I
`
`FIG.SA
`
`-0.4•.
`
`~
`
`• • • • •I• •I· • ' I ' · • •I• t, •
`
`'
`
`~
`
`• • • • •
`
`.-: '
`
`•
`
`'
`
`• • •
`
`• •
`
`• • 1 · 1:· • • • •
`
`t
`' ....... ·•· pt· .. -~.. .. . . . .. •· ·•· ..... .
`
`'
`
`0
`
`200
`
`400
`
`600
`TIME
`
`800 .
`
`1000
`
`1200
`
`1400 X
`
`Y o ~ - - - - - - - - - - - - - - - - - - - - - -
`
`.
`.
`.
`.
`·
`·
`-0.05 ............. : ............. ·:·. ·. ·. · · · · · ·. ·:· · · · · · · · · · · · · :· · · · · · · · · · · · · -~ · · · · · · · · · · · · · ·:· · · · · · · · · · · · · ·
`j
`\
`!
`!
`.
`.
`.
`.
`
`!
`.
`
`:
`.
`
`-0.2 L - . . - - - - - - - - - - - - - - - - - -~ - - - - - - - '
`0
`200
`400
`600
`800
`1000
`1200
`1400 X
`TIME
`
`MOVING
`AVERAGE OF
`INSTANTANEOUS
`PERFORMANCE
`MEASURE
`
`FIG.SB
`
`•
`
`~ • r:JJ.
`~ = [
`
`> = ~
`t,-;. J~
`
`t,-;.
`
`~
`
`~
`
`ga
`~
`~
`,I::..
`s,
`
`,I::..
`
`°' -..
`Q'\ °' Q'\
`
`-..
`\C
`N
`~
`
`BMW1044
`Page 5 of 10
`
`
`
`5,656,921
`
`1
`CONTROL OF A VEHICLE POWERTRAIN
`
`FIELD OF THE INVENfION
`
`The invention relates to the control of a vehicle power-
`train.
`
`5
`
`BACKGROUND OF THE INVENTION
`
`2
`means to adapt stored memory maps by replacing stored
`values therein with altered values in response to the mea(cid:173)
`surement of said performance of the powertrain with respect
`to the desired relationship between the contributions of each
`of the motor and engine.
`According to another aspect of the invention there is
`provided a method of controlling a powertrain using a
`controller as set out in the immediately preceding paragraph
`including the steps of:
`a) measuring the powertrain performance and
`b) adjusting the powertrain controller accordingly.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`Control of a vehicle powertrain in accordance with the
`invention will now be described by way of example with
`reference to the accompanying drawings in which:
`F1G. 1 is a schematic block diagram showing the form of
`the powertrain and the role of the powertrain in a hybrid
`20 vehicle;
`F1G. 2 shows an example of the maps used to control the
`powertrain;
`F1G. 3 is a graph of the relationship between throttle angle
`and demanded armature current as pedal value increases
`with engine speed held constant;
`F1G. 4 is a schematic diagram of an adaptive fuzzy hybrid
`powertrain controller; and,
`F1G. 5 shows a plot of instantaneous performance mea(cid:173)
`sure of the powertrain in terms of its ability to use the energy
`resources to the vehicle in a desired ratio over time with two
`corresponding plots of moving average estimates over time.
`
`30
`
`DEI'AII..ED DESCRIPTION OF THE
`PREFERRED EMBODIMENTS
`
`In F1G. 1 a powertrain controller 10 is provided which is
`used to interpret movements of an accelerator pedal 12 of an
`internal combustion engine 14 and, using engine speed,
`generate a throttle angle demand signal for the engine 14 and
`an armature current demand signal for an electric motor 16.
`Torque is delivered by the engine 14 and the motor 16 to a
`transmission of a vehicle.
`Other electric motor demand signals could be used depen(cid:173)
`dent upon the type of the electric motor and the way in which
`the electric motor operates. It should be noted, in passing,
`that a vehicle having this particular arrangement of the
`internal combustion engine and the electric motor is known
`as a parallel hybrid vehicle. This is because there are parallel
`paths for the fuel and electrical energy to reach the wheels
`of the vehicle. An alternative type of hybrid vehicle is
`known as a serial hybrid vehicle, in which the energy from
`the internal combustion engine passes through electrical
`machines to reach the wheels of the vehicle.
`A feature of the arrangement of F1G. 1 is that the
`powertrain controller 10 observes the engine fuel flow rate
`and the battery current via signals 15 and 17 respectively,
`whilst the output torques of the internal combustion engine
`14 and electric motor 16 cannot be measured.
`The powertrain controller 10 is implemented using fuzzy
`logic. A detailed understanding of fuzzy logic is not required
`in order to appreciate the manner in which the basic hybrid
`powertrain controller 10 works. As stated above, the con(cid:173)
`troller 10 has two pieces of input information derived from
`movement of the accelerator pedal 12 and engine speed from
`engine speed input 13 and generates two pieces of output
`information, the demanded throttle angle which is delivered
`
`Emissions regulations which must be met by vehicles are 10
`becoming increasingly strict. In order to meet these increas(cid:173)
`ingly strict regulations, it is felt that, in the future, something
`other than a conventional internal combustion engine
`vehicle will be required.
`One suggestion that might meet this requirement is a 15
`hybrid vehicle which is powered by a combination of an
`internal combustion engine and an electric motor. An addi(cid:173)
`tional factor motivating the use of hybrid vehicles is the
`suggested zero emission zone which might be used in city
`centres to reduce acute urban pollution. A hybrid vehicle
`could meet the requirements of the zero emission zone
`within the city and yet still have sufficient acceleration and
`range to allow its use outside urban areas.
`Having designed a hybrid vehicle, there is a need to
`establish the way in which the powertrain of the vehicle 25
`should be used. The term "powertrain" means the engine and
`drive transmission by means of which power is transmitted
`to drive the vehicle. Under some circumstances, one method
`of operating the powertrain will be obvious. For instance,
`when a hybrid vehicle is used for a long motorway journey,
`the electric part of the powertrain will be unable to make a
`useful contribution because of the amount of power and
`energy required. Conversely, in a zero emission zone, the
`internal combustion engine can make no contribution.
`However, there are potentially two other ways in which the
`powertrain of the vehicle could be used. The power for
`propelling the vehicle could be drawn partly from the
`internal combustion engine and partly from the electric
`motor. That would have the advantage of being less emissive
`and cheaper than using the internal combustion engine
`alone. A further method might be to use the electric motor
`to oppose the internal combustion engine of the vehicle, in
`order to increase the state of charge of the battery used to
`power the electric motor. That might be useful when
`approaching a zero emission zone with a depleted battery.
`The first two categories mentioned above will not require
`any large increase in complexity since the drive arrange(cid:173)
`ments will be similar to those of conventional internal
`combustion engine vehicles and electric vehicles respec- 50
`tively. The second two methods of operation will require the
`separate actions of the engine and the electric motor to be
`combined and one object of the present invention is to
`provide a suitable controller for that purpose so that the
`driver does not have to control the electric motor and the 55
`engine individually.
`
`35
`
`40
`
`45
`
`SUMMARY OF THE INVENTION
`
`According to one aspect of the invention there is provided
`a controller for a powertrain, the powertrain including an 60
`electric motor powered by an electrical power source and an
`engine powered by combustible fuel, the controller com(cid:173)
`prising means for measuring the performance of the pow(cid:173)
`ertrain with respect to a desired relationship between the
`contributions of each of the motor and engine, the controller 65
`being operative to maintain the desired relationship and
`further comprising means for providing the controller with
`
`BMW1044
`Page 6 of 10
`
`
`
`5,656,921
`
`4
`If those points have incorrect values then the powertrain will
`not generate the correct amount of torque leading to poten(cid:173)
`tial problems in the driveability of the vehicle. The tradi(cid:173)
`tional fuzzy logic design approach uses the knowledge and
`5 engineering intuition of a design engineer to place and shape
`the fuzzy sets. However. when that approach was used in the
`design of the hybrid powertrain controller of the present
`invention it proved difficult to design controllers that pro(cid:173)
`duced sensible, consistent output torques. Additionally, the
`10 traditional fuzzy design approach does not allow the relative
`use of the electrical energy stored in the batteries of the
`vehicle and the fossil fuel to be easily controlled. To improve
`the powertrain output torque generation and to control the
`relative use of the energy resources of the vehicle the
`15 following steps were taken in accordance with the present
`invention.
`As shown in FIG. 4, which shows the operation of the
`powertrain controller (indicated at 20) in block diagram
`form, the powertrain controller has a control unit 26 and two
`20 additional functional blocks 22,24 that are not present in the
`basic powertrain controller 10 shown in FIG. 1. The opera(cid:173)
`tion of these two blocks, ( a performance measurement block
`22 and a rule modifier block 24) will be considered sepa-
`rately below.
`The performance of the powertrain in terms of its ability
`to use the energetic resources of the vehicle in a desired ratio
`can be measured using a function such as
`
`3
`to the engine 14 and the demanded electric motor armature
`current or equivalent electric motor load demand signal
`which is delivered to the motor 16. Fuzzy logic can be
`thought of as a method of interpolating between fuzzy sets.
`A fuzzy set is a method of representing vaguely defined
`quantitative notions such as "a medium amount of throttle
`angle". A fuzzy controller can be constructed by combining
`a number of such sets using fuzzy rules. If the sets are of a
`particular form, and they are combined as rules in a par(cid:173)
`ticular way, the input-output relationship for the controller
`10 reduces to linear interpolation between the points that are
`used to define the fuzzy sets. In effect, this is equivalent to
`the use of maps in engine management systems. Such maps
`relate one dependent variable to one or more independent
`variables. An example might be the ignition timing map
`which relates the ignition timing (the dependent variable) to
`the engine load and speed the independent variables).
`In the controller 10 there are two output variables, which
`might be referred to as dependent variables. There are,
`therefore, two maps which relate the dependent variables of
`demanded throttle angle and demanded armature current to
`the independent variables of pedal value and engine speed.
`FIG. 2 shows an example of the two maps used in the
`powertrain controller 10. The left hand map (i) plots pedal
`value on the x-axis and engine speed in revolutions per 25
`minute on the y-axis against demanded throttle angle in
`degrees on the z-axis. The right hand map (ii) plots pedal
`value on the x-axis and engine speed in revolutions per
`minute on the y-axis against demanded motor armature
`current in Amperes on the z-axis. The maps would be used 30
`in the mode of operation in which the electric motor 16
`opposes the action of the internal combustion engine 14. It
`is seen that. where the electric motor 16 generates relatively
`large negative armature currents, the throttle angle is
`increased to maintain the powertrain torque output.
`The dimensions of the maps used would normally be in
`the order of 5 to 20 points, by 5 to 20 points and, for
`example, a 5 by 13 array of values can constitute such a map.
`The effectiveness of operation of controller 10 is deter(cid:173)
`mined by the relationship between the throttle angle and the
`demanded armature current. An example of this relationship
`is shown in FIG. 3. in which the engine speed is constant and
`the throttle angle is plotted along the x-axis against the
`demanded armature current on the y-axis as the pedal value 45
`varies indicated by arrow A. The crosses in FIG. 3 represent
`the points through which the linear interpolation passes and
`are, when the fuzzy logic is implemented in the manner
`referred to above (Page 6 Lines 15-26) the points which
`define the fuzzy sets in the hybrid powertrain controller. 50
`FIG. 3 indicates that the relative position of the fuzzy sets
`that define the engine throttle angle and the demanded
`armature current define the operation of the combined
`powertrain. For zero movement of pedal 12. the engine
`throttle angle should be zero and the electric motor currents 55
`should also be zero ( engine throttle angles are measured
`relative to the idle throttle angle since clearly, at idle, the
`throttle angle is not zero). When the pedal is fully depressed
`the throttle angle is 90° and, again, the electric motor
`currents are zero. This gives the vehicle the same maximum 60
`acceleration or top speed during hybrid operation as in
`internal combustion engine only operation. Clearly this is
`not the maximum acceleration or top speed of the vehicle
`since the electric motor 16 could be used to assist the
`internal combustion engine 14 if desired.
`The points defining the throttle angle and armature current
`fuzzy sets, define how the powertrain will generate torque.
`
`35
`
`in which Ibat is the battery current drawn by the electric
`motor, ~ is the internal combustion engine fuel flow rate
`and ke is a constant that defines the equivalence of fuel and
`electrical energy. In FIG. 5, the upper plot shows the
`instantaneous values of the performance measure for a
`controller 20 along the y-axis against time on the x-axis,
`which causes the electric motor 16 to work against the
`internal combustion engine 14 being used in a simulation of
`an urban route. As can be seen from the upper plot of FIG.
`40 5 the instantaneous values of this performance measure vary
`enormously and give little indication of the performance of
`the vehicle in general. For this reason a moving average
`estimate of the performance measure, F is used and is
`calculated as follows
`
`where F n is the moving average value of the instantaneous
`performance measure, F, at time index n. The parameter v
`can be used to adjust the rate at which the average is
`influenced by new information. The lower plot of FIG. 5
`shows the moving average value along the y-axis against the
`same x-axis as in the upper plot.
`By way of example, the solid line in FIG. 5 is derived for
`v=5000 in the above equation, whilst the dashed line is
`derived for v=2500.
`The fuzzy rules are repeatedly modified after a period of
`time called herein the adaptation interval, Ta which can be
`50 seconds for example. As mentioned above the powertrain
`controller can be reduced to the form of maps of throttle
`angle and demanded armature current over the independent
`variables of engine speed and pedal movement value. Vec(cid:173)
`tors of engine speed and pedal movement value are used to
`65 define a grid in which the controller output variables of the
`powertrain controller 20 are stored. The powertrain control(cid:173)
`ler 20 uses the pedal movement value and the engine speed
`
`BMW1044
`Page 7 of 10
`
`
`
`5,656,921
`
`5
`to calculate a new throttle angle and a new demanded
`armature current. This calculation can be carried out evecy
`0.1 second, which is called herein the controller sampling
`interval. Associated with the controller output maps is an
`array of counters the same size as the controller output maps.
`At each controller sampling interval the nearest point to the
`current operating point on the grid over which the controller
`output map is defined is identified. The corresponding
`element in the counter array is then incremented. These
`counters are reset after each adaptation interval. If, during an
`adaptation interval, the counter corresponding to a particular
`location on the grid is equal to or greater than one, then the
`fuzzy rule which corresponds to this counter is termed "a
`modification driving fuzzy rule" and if the counter is zero
`the fuzzy rule is termed "a modification following fuzzy
`rule." The two types of rules are modified differently. It will
`be appreciated that a threshold other than one could be used
`to distinguish between the two types of rules.
`When modifying the controller maps, the powertrain
`output torque should remain constant. This means that, if the 20
`throttle angle is reduced, the demanded electric motor arma(cid:173)
`ture current should increase. The relative amounts of change
`that should be made to the controller output variables is
`defined using information about the internal combustion
`engine 14 and the electric motor 16. This information 25
`defines a direction in FIG. 3 along which the points in the
`controller maps defined by the crosses in FIG. 3 should
`move. Regardless of whether the controller output sets are
`classified as modification drivers or followers, the points
`represented by the crosses in FIG. 3 are always moved in a 30
`direction which maintains the powertrain output torque.
`The information about the internal combustion engine and
`electric motor used to define the direction in which the
`points in FIG. 3 move, relates the rate of change of engine
`output torque to throttle angle over all engine speeds and 35
`throttle angles, and the rate of change of motor output torque
`with respect to demanded motor armature current or other
`motor demand signals over all motor speeds and motor
`demand signal values. This information is obtained by
`earlier measurements carried out on the engine and motor in 40
`the condition in which they are used in the vehicle.
`The crosses representing the driving rules are moved in
`the direction which maintains a constant powertrain output
`torque a distance proportional to the difference between a
`desired value F d of the performance measure F and the 45
`current estimate, F of the performance measure. This dis(cid:173)
`tance is modified where the shape formed by the crosses in
`FIG. 3 is not a smooth curve. By enforcing a smooth curve,
`the controller 20 will gradually vacy the use that is made of
`the two powertrain elements as the driver's pedal is further 50
`depressed and this is desirable from a practical point of view
`since it will tend to maintain the driveability of the vehicle.
`The fuzzy rules which have been identified as modifica(cid:173)
`tion followers during the last adaptation interval are moved
`such that the smoothness of the overall shape formed by the 55
`crosses in FIG. 3 is maintained.
`The driving rules strongly influence the basic nature of the
`shapes formed by the location of the fuzzy output sets,
`shown by the crosses in FIG. 3, the following rules are
`moved such that the smoothness of this shape is maintained. 60
`The performance of the controller is determined by the
`values stored in the controller output maps. Since the
`modification of the shapes formed by the maps, an example
`of which is shown in FIG. 3, is driven by the modification
`driving rules, the shape formed by the crosses in FIG. 3 is 65
`determined by the driving rules. The driving rules are
`identified by the use of the controller 20 over an adaptation
`
`6
`interval and, since this use will vacy depending upon the
`driver of the vehicle and the route being driven, the hybrid
`vehicle powertrain can adapt to use the energy resources of
`the vehicle in the correct ratio for different drivers over
`5 different routes.
`The operation of the adapted hybrid vehicle powertrain
`control system is greatly improved if the value of v used in
`the calculation of the moving average F is set at relatively
`low values when the value of F is a long way from the
`10 desired value F d• and relatively high values when the value
`of Fis closer to F d· This improves the responsiveness of the
`adaptation procedure when F is a long way from F d• and it
`improves the accuracy of the adaptation procedure when F
`is closer to F d· When a sampling interval of 0.1 sis used, a
`15 small value of v might be 2500 and a large value of v might
`be 40000.
`A further example of an operable system according to the
`invention, which specifically uses a fuzzy logic control
`system is given in the inventor's PhD thesis entitled "A
`Study in the use of Fuzzy Logic in the Management of an
`Automotive Heat Engine/Electric Hybrid Vehicle
`Powertrain", which is available from the University of
`Warwick, United Kingdom.
`Other published hybrid vehicle powertrain controllers
`have used a fixed control strategy for a powertrain, and the
`relative use of the energetic resources of the vehicle could
`not then be adjusted to match the requirements of individual
`users. Experiments have been performed using data obtained
`from different drivers over the same urban route. It was
`found that, due to differences in the way in which the drivers
`used the vehicle, the adapted controller 20 demanded 30%
`more armature current for one driver than for the other in
`order to use the energetic resources of the vehicle in the
`same ratio. The ability of the controller to adapt to the
`different driving characteristics between drivers in this man(cid:173)
`ner allows the vehicle performance to meet the requirements
`of different users.
`The adaptive hybrid vehicle powertrain controller 20
`could be used in two ways. Firstly it could be used as an
`engineering development tool for the purposes of develop(cid:173)
`ing the controller 20 on prototype vehicles. Secondly, the
`controller 20 could be used in an adaptive manner on a
`production hybrid vehicle in service. There are two modes in
`which both elements of the powertrain are used at the same
`time. In the mode in which the electric motor 16 acts against
`the internal combustion engine 16 to increase the battery
`state of charge, the adaptive hybrid powertrain controller 20
`is unlikely to be of a great deal of use in service. This is
`because this mode might only be used as an "emergency
`mode" and would not be used frequently enough for the
`adaptation process to be effective. The adaptation process
`can be used in developing a fixed controller used in this
`mode.
`Conversely, the adaptive hybrid powertrain controller is
`vecy useful in the mode where the electric motor 16 assists
`the internal combustion engine 14. Many vehicle users have
`a fixed pattern of vehicle usage which does not vacy greatly
`from day to day. For such users there is a considerable
`advantage in being able to deplete their battecy to a reason(cid:173)
`ably low, but safe, depth of discharge evecy day and recharge
`the battecy overnight. The signal F d can be provided by a
`potentiometer, or some other device that the user of a vehicle
`can adjust to bias the actions of the powertrain controller 20
`more towards the electric motor 14 or the internal combus(cid:173)
`tion engine 16 of the vehicle. Over a period of a few days
`the user of the vehicle can adjust the value of F d so that the
`desired depth of battecy discharge is obtained over daily
`
`BMW1044
`Page 8 of 10
`
`
`
`5.656,921
`
`10
`
`25
`
`7
`motoring. Where the user varies the customary usage of the
`vehicle. a different mode of powertrain operation could be
`selected.
`Regenerative braking by using the electric motor to
`provide a braking load can be used to generate electric 5
`current for charging batteries on the vehicle. Such regen(cid:173)
`erative braking is likely to be of use primarily in an urban
`environment where there are likely to be more instances of
`braking than when driving on non-urban roads.
`I claim:
`1. A method of controlling a powertrain of a hybrid
`electric vehicle, the powertrain including an electric motor
`and an internal combustion engine which supply torque in
`varying contributions to a transmission, the method com(cid:173)
`prising the steps of:
`a) providing a fuzzy controller utilizing fuzzy logic oper(cid:173)
`ating means comprising fuzzy rules;
`b) representing a fuel flow rate controlling output and a
`motor current controlling output as a function of a drive
`demand input signal and a powertrain speed input 20
`signal;
`c) using said fuzzy rules to relate the drive demand input
`signal and the powertrain speed input signal to said
`outputs;
`d) defining a set of input reference points for each input
`signal which varies over time, said sets together form(cid:173)
`ing an input reference grid;
`e) defining an output map over the input reference grid,
`said map relating each output value to the input values, 30
`interpolating within the output map using the input
`reference grid to obtain an actual output value corre(cid:173)
`sponding to a pair of actual input values;
`f) distinguishing driving points in the input reference grid
`which have a greater effect in powertrain performance, 35
`from following points in the input reference grid which
`have a lesser effect on powertrain performance;
`g) adjusting the values in the output map corresponding to
`the driving points in the input reference map;
`h) adjusting the values in the output map corresponding to
`the following points in the input reference map in order
`to obtain a desired relationship between said contribu(cid:173)
`tions;
`i) replacing the actual output value with an altered value
`corresponding to the adjusted values in steps g) and h).
`2. A method according to claim 1 including the steps of
`providing an array of counters of the same dimension as the
`input reference grid, and distinguishing the points in the
`input reference grid as either driving or following points by 50
`the following steps:
`a) setting each of the counters in the array of counters to
`zero;
`b) monitoring actual input values of the drive demand
`input signal and powertrain speed input signal;
`c) determining the nearest point in the input reference grid
`to the actual input values;
`d) incrementing the counter corresponding to the point
`determined in step c) above in the array of counters by
`one for each of the given sampling periods of time that
`this remains the nearest point to the actual input values;
`e) repeating the steps b) and d) for an adaptation interval
`period of time;
`f) defining as driving points those points whose corre- 65
`sponding counters have a value greater or equal to a
`threshold and all other points as following points;
`
`8
`g) adapting the points in the output maps corresponding to
`the driving points to modifying the powertrain perfor(cid:173)
`mance towards the desired powertrain performance and
`to ensure that the powertrain output torque correspond(cid:173)
`ing to any given value of drive demand and powertrain
`speed remains substantially constant throughout the
`process of the controller modification; and
`h) adapting all the points in the output maps to ensure that
`as the drive demand and powertrain input speed vary
`the actions of the engine and the motor vary in a
`manner that is substantially smooth.
`3. A controller for a powertrain where the powertrain
`comprises a transmission, an internal combustion engine and
`an electric motor, where the engine and the motor supply
`15 torque in a varying contribution to the transmission, a
`battery providing an energy resource for the electric motor
`and a fuel supply providing an energy resource for the
`internal combustion engine;
`the controller comprising a measuring means for measur(cid:173)
`ing the ratio in which the fuel supply and the battery are
`used; and
`a memory having a plurality of stored maps comprising a
`plurality of values stored therein with at least one value
`obtained from the measuring means;
`wherein the controller can adapt the maps by updating one
`or more of the plurality of stored values with one or
`more altered values to reach a desired relationship
`between the contributions and then replacing said one
`or more stored values with said altered values.
`4. A controller according to claim 3, wherein the control(cid:173)
`ler is a fuzzy controller operating by means of fuzzy logic to
`control said contributions.
`5. A controller according to claim 4, wherein the control(cid:173)
`ler controls the engine by means of throttle position.
`6. A controller according to claim 5, wherein the control(cid:173)
`ler controls the motor by means of the motor current.
`7. A controller according to claim 3, wherein the control