A CONTROL UNIT FOR A VEHICLE

The invention relates to a control unit and method for a vehicle having a plurality of independently-driven wheels.

Background of the Present Invention

European Patent Specification No. 1 305 180A in the name Pro-Drive 2000 Limited discloses a system which controls the torque applied to individual wheels by using a conventional drive train with differential locking.

Summary of the Invention

According to a first aspect of the invention, there is provided a control unit for a vehicle having a plurality of independently-driven wheels, the control unit comprising an input unit for receiving a steering command, a speed command and outputs from acceleration sensing units being locatable at respective wheels of the vehicle; and a processor in communication with the input unit and being operable to calculate for each wheel a desired wheel rate including (i) a main component based on the speed command; and (ii) a manoeuvre response component based on the steering command and on accelerations induced by the steering command and/or terrain conditions.

By the term "independently-driven wheels", it is evident to the skilled person in the art there is understood to include a plurality of sets of independently-operated traction and/or motion mechanisms which may for example be wheels, or "caterpillar" tracks, or similar or equivalent treads.

During a manoeuvre being executed in response to a steering command, the processor may be operable to calculate a manoeuvre response component for a wheel on an outer turning track which is greater than that for a wheel on an inner turning track.

According to a second aspect of the invention, there is provided a vehicle having a plurality of independently-driven wheels comprising the control unit of the first aspect of the invention.

According to the first aspect of the invention, there is also provided a method of controlling a vehicle having a plurality of independently-driven wheels, the method comprising receiving a steering command, a speed command and outputs from acceleration sensing units being locatable at respective wheels of the vehicle; and calculating for each wheel a desired wheel rate including (i) a main component based on the speed command; and (ii) a manoeuvre response component based on the steering command and on accelerations induced by the steering command and/or terrain conditions.

The invention provides an electronic differential drive control system for an electrically powered multi- wheeled vehicle in which differential drive concerns a differential wheel rate comprising a component due to steering commands and an additional component due to a resulting lateral induced acceleration initiated by the steering command and involving any combination of wheels on the left and right side of the vehicle which may or may not include the steered wheels.

The control system provides optimum distribution of power to each independent wheel in either 2x2, 4x4, multiple wheel drive utilizing multi-pancake direct wheel drive or other electric drives utilizing power from either internal power storage, on board hybrid generator or an external power supply.

The control system is able to achieve coupled differential control of drive wheels to directional steering commands to maintain straight line or turning directional control and stability under motion, and to achieve coupled differential control of drive wheels to directional steering commands to maintain straight line or turning directional control while compensating for wheel slip resulting from changes in wheel-road surface frictional coefficient.

The control system may provide for recharging of power supply under regenerative braking.

The control system may concern a differential wheel rate comprising a component due to steering commands and an additional component due to a resulting lateral induced acceleration initiated by the steering command and accounting for all combinations of wheel contact or non contact with the road surface.

The vehicle may be a tracked vehicle in which differential drive concerns a differential track velocity comprising a component due to steering commands and an additional component due to a resulting lateral induced acceleration caused by the steering command.

The electronic differential drive control system comprises a routine that functions within an electric vehicle control unit as an Electric Vehicle Independent Drive- controller System henceforth to be known as the EVIDS routine.

Any manoeuvre results from an applied lateral g response resulting from the steering action and accomplished through frictional forces at the terrain surface interface with all tyres (assuming all tyres are in contact with the terrain). The EVIDS control routine in this same simple case invokes an increased wheel rate associated with following inner circle and outer circle tracks for inner and outer wheels when following the instantaneous circular path induced by the manoeuvre and initiated by the steering demand.

The EVIDS routine induces a wheel rate change under manoeuvre to assist traction in the manoeuvre and reduce wheel scrub while improving stability by reducing tyre slip angle (the angle between the wheel and the direction in which the tyre wants to go). Parasitic induced accelerations resulting from road surface conditions always exist but in particular when manoeuvring if slip angle limits are sensed being exceeded then the EVIDS routine (via the acceleration feedback signals and combined control routines) enhances stability by inducing an

appropriate change in wheel rate over and above that to maintain speed which biases the wheel to run inside the slip angle limit while maintaining the manoeuvre. This ensures a much safer driving experience with improved manoeuvre responsiveness and energy efficient road handling.

The control routines within EVIDS utilise feedback values in all three orthogonal axes from acceleration sensing units at each wheel station. Accelerations (excluding noise, structural or otherwise) include components due to vehicle manoeuvre in response to a steering wheel command and additional components attributable to road surface friction conditions and surface irregularities resulting in wheel bounce or incidental impulsive loads. As a result, wheel dynamics and local acceleration responses, whether induced by a steering command or not, are a cocktail of components that are largely non-deterministic. Thus, their effect is monitored through the acceleration sensing units.

Essentially a zero steering wheel demand results in a zero manoeuvre demand. Any non-zero steering demand at speed induces a lateral acceleration. The accelerations associated with these manoeuvres in the e-vehicle body fixed axes determine the primary differential wheel rate response. However, the cocktail of parasitic accelerations in all three wheel local body axes feeds into the EVIDS routine and therefore induces a parasitic wheel rate response which is a secondary nuisance. Structural filtering minimizes signal noise in the acceleration sensing unit signals processed at each wheel unit. The purpose of the EVIDS routine is to monitor the overall acceleration feedback from the wheel acceleration sensors and despite parasitic effects maintain control of the vehicle. Different cars with different weight and moments of inertia and suspension system dynamics respond differently to different wheel induced disturbances. Different control parameters are therefore required in each case for each vehicle type although these are customisable values within the EVIDS routine generic framework controller design. Final track testing is required to validate the parameter sets assigned to each EVIDS routine implementation associated with each vehicle design.

When wheel slip is experienced under traction and slip angle limits are exceeded under manoeuvre, EVIDS routine is able to distinguish between accelerations resulting from the steering command and those resulting from terrain conditions, as there are detectable changes in the feedback from the acceleration sensing units. This includes wheel bounce conditions where the wheel momentarily loses traction. Wheel slip under traction is monitored by checking the wheel rate/rate acceleration induced velocity/longitudinal acceleration at the wheel-tyre contact patch with the forward speed/body acceleration of the vehicle associated with the same wheel station (an assumed wheel radius is used here). Any significant computed error in velocity signifies slip which is confirmed by an unexpected rate acceleration outside expected values. In the case of slip angle being exceeded under manoeuvre, the lateral accelerometer at the wheel station records a leap in value as lateral inertial load exceeds the frictional force required to maintain wheel-tyre grip, hi this case, the EVTDS routine control loop is designed to improve on current traction control systems associated with this event through the idea of inducing increased wheel rate to reduce slip angle and force the e- vehicle along a more ideal manoeuvre track. In the event that differential wheel rate via the EVIDS routine fails to stabilise the vehicle within control limits, and the vehicle yaw rate increases while skidding, the change in induced lateral acceleration invokes a differential wheel response via the EVIDS routine that tends to assist the regaining of stability of the e-vehicle and aid bringing the e- vehicle back under control along the intended drive path.

Wheel bounce results in an unexpected increase in wheel rate acceleration akin to loss of friction grip at the road surface. Load sensors at each wheel in addition to vertical acceleration feedback aids confirmation of this transient response.

Detection of wheel bounce and loss of traction is used in the control logic circuit to redirect traction control to those wheels that are able to deliver the necessary road torque to maintain the desired speed and direction.

Corrective action comes from one of three principle events. These are a) wheel bounce, b) slip angle limits exceeded by a wheel or wheels under manoeuvre, c)

traction control lost due to loss of rolling contact friction at wheel or wheels in manoeuvre or straight-line motion. AU of these events are associated with a change in acceleration response at each of the wheel stations. In the EVEDS control system, wheel bounce is monitored from load sensors at each wheel location in addition to vertical acceleration sensor feedback. Traction control logic then reverts to transient use of all remaining wheels in contact with the road surface to deliver any instantaneously demanded manoeuvre and forward speed. Traction control always reverts to using those wheels that are able to deliver the required drive torque to the road surface. Slip angle limits exceeded results in a local wheel spike in lateral acceleration that couples through to all other wheel acceleration sensors. The EVDDS controller is designed to be robust enough to compensate for these events through design of the inner loop controller but is able to identify the lateral accelerator response as attributable to this event. Similar arguments apply to the event where a wheel or wheels lose rolling contact friction or traction control. In this instant, slip is monitored by comparing wheel rate with that associated with the vehicle speed at each wheel station in question, along with associated rate accelerations. If the computed velocity at the contact patch is other than zero, wheel slip is assumed and the controller adjusts all wheel demanded rates to bring all wheels in rolling contact, whether offering traction or not. This ensures no wheel scrub. Ih the rare event that no wheel combination is available to deliver manoeuvre or demanded forward speed, wheel rate demand reduces until wheel grip is achieved at a lower speed than that demanded. For very icy conditions where no traction control is possible at speed, the EVDDS routine defaults under this control logic to reducing wheel rate to zero, thereby bringing the vehicle to rest as a safety feature, hi this extreme case there can be no safeguard against maintained momentum on ice - the e-vehicle goes where it wants to even if wheel rate is reduced to zero. In this regard, the EVDDS routine is no different to any other car.

The acceleration sensing units are located at the four corners of a rectangle within the vehicle sub-frame assembly and in line with the virtual axle through each wheel hub. This is an analytical ideal assumed in the EVDDS control loop design.

Applications of the Present Invention

The present invention is applicable to a control unit for the operation of any form of vehicle, whether manned or un-manned, including a vehicle which is remotely- operated.

Advantages of the Present Invention

An advantage of the present invention may be that it may be more energy efficient because it can eliminate a conventional drive train and limited slip differential units, thereby reducing/avoiding friction losses.

Another advantage may be that the present invention may utilise multiple individual electric hub or electric drive shaft motors to distribute the torque to each driven wheel.

Moreover, a further advantage may be that the present invention may distribute the appropriate amount of torque to each wheel for a given manoeuvre by directing the electrical power supply to each wheel motor as required.

Furthermore, the present invention can utilise regenerative braking during the process to improve energy efficiency.

The present invention includes means and elements to implement each of the features referred to in the advantages disclosed herein.

The present invention is applicable to the control system and unit in a vehicle having a plurality of independently-operated traction and/or motion mechanisms which may for example be wheels, or "caterpillar" tracks, or similar or equivalent treads.

Brief Description of the Drawings

In order that the invention may more readily be understood, a description is now given, by way of example only, reference being made to the accompanying drawings, in which:- Figures Ia and Ib are schematic diagrams of an e-vehicle according to the invention;

Figure 2a is a schematic diagram of a control unit for the e-vehicle;

Figure 2b is a second schematic diagram of the control unit;

Figure 3 is a schematic diagram of the e-vehicle of Figure 1 showing a mathematical representation of wheel tracks for a turning circle;

Figure 4 is a schematic diagram of the e-vehicle of Figure 1 showing the general dynamics of the e-vehicle when executing a turning manoeuvre;

Figure 5 is a schematic diagram showing a mathematical representation of a wheel-ground friction interface for the e-vehicle of Figure 1.

Detailed Description of the Invention

The electronic differential drive control system comprises a routine that functions within an electric vehicle control unit as an Electric Vehicle Independent Drive- controller System (EVIDS) routine.

Figures Ia and Ib are schematic diagrams of an e-vehicle 10 according to the invention for the specific application of 4-wheel drive.

The e-vehicle 10 includes an EVIDS control unit 12 which outputs signals to four power controllers 14a-d which respectively control four electric drive wheel motors 16a-d, which in this embodiment are pancake motors.

The EVIDS control unit 12 receives an accelerator pedal demand 20, a braking demand 22, a steering wheel demand 24, a 2- or 4-wheel drive command 26 and a traction control demand 28, as well as outputs of four acceleration sensing units

18a-d positioned at respective wheel' stations (not shown) of the e-vehicle 10. In

this embodiment, each acceleration sensing unit 18a-d detects acceleration in three orthogonal axes.

A power supply 30 comprising a battery and/or a fuel cell provides power to the EVIDS control unit 12, the four power controllers 14a-d and the four wheel motors 16a-d.

Thus, each wheel of the e- vehicle 10 is driven by its own wheel motor 16a-d and is controlled by its own power controller 18a-d. Drive voltage to each motor 16a-d is demanded from the EVIDS control unit 12 and is drawn from the power supply 30. The power controllers 14a-d ensure that each wheel motor 16a-d is not overloaded and is maintained within safe operational power limits and comprises its own built in fail-safe monitoring system. The EVIDS control unit 12 manages the accelerator pedal demand 20, the brake demand 22 and the steering wheel demand 24 in conjunction with signals from the acceleration sensing units 18a-d at each of the wheel stations together with wheel positional data from position feedback sensors, as part of a closed loop controller system in which the driver of the vehicle is an integral part. Because of the direct wheel drive of the e-vehicle 10 with pancake wheel motors 16a-d and the EVIDS control unit 12, the invention provides a more responsive and sophisticated traction control system than is possible with a conventional car with mechanical transmission drive powered by an internal combustion engine.

During a manoeuvre involving the inducing of a lateral acceleration on a vehicle, the ability of the vehicle to maintain stability in the turn is significantly influenced by the limit slip angle of the tyres at the road interface and the prevailing road surface/tyre friction coefficient.

According to the invention, by imposing a wheel rate differential between left and right sides of the e-vehicle 10 in order to initiate the turn manoeuvre (which in the case of steered wheels is additive to that which supports the steering command), the slip angle of the tyre at the road/tyre interface is reduced compared to a

conventional vehicle where such extra wheel rate is not present. The result is reduced side slip and improved inherent stability for a given turn manoeuvre initiated by a steering command. As an independent wheel drive controller, EVIDS computes each individual wheel rate during a given turn manoeuvre allowing for those wheels that are in contact with the road surface at any instant in time and the nature of the tyre/road interface friction conditions experienced by each wheel to include the case of wheel bounce where wheel(s) may temporarily leave the road surface. Where wheel slip is encountered by any wheel in the cluster, that wheel defaults to a command wheel rate in agreement with a computed value consistent with that needed to maintain the manoeuvre and for the given location on the chassis of the e-vehicle 10. The quick responsive nature of the pancake wheel motors 16a-d under EVIDS control to achieve a given torque" (which in turn achieves a demand wheel rate) ensures that, when a wheel tyre that has left the road surface once again regains contact (for example in a bounce condition), the ability to regain rate under increased wheel load prevents instability induced by over- or under-steer.

By adopting a direct drive control system, EVIDS minimizes power consumption and thereby extends range capability by optimizing energy usage. Drive efficiency is also increased since there are no intermediate transmission losses through mechanical drive links as in conventional vehicles.

hi this embodiment, regenerative battery energy capability resulting from energy capture under braking further improves the efficiency of the e-vehicle 10.

Applications include: wheeled vehicles including cars (on/off road), light vehicles, tricycles, quadcycles, buses, lorries, support vehicles, trams, trolley buses and trains; civil and military tracked, track-band or multiple drive wheeled vehicles; tractors (to include garden tractors); planetary remotely controlled robotic rovers; and articulated vehicles. There is also application to hydraulic drive hub vehicles.

Figure Ib is a schematic diagram of the e- vehicle 10, showing the control unit 12 receiving the accelerator demand 20, the braking demand 22, and the steering demand 24.

The control unit 12 also receives feedback from each power controller 14a-d in maintaining energy efficient control of all wheels in achieving the required traction control.

The control unit 12 outputs to each power controller 14a-d a wheel rate demand 32a-d, the wheel rate for each wheel having been calculated according to the control method described herein.

As indicated by the dotted lines, each power controller 14a-d receives power from the power supply 30. Also indicated is the flow of power from each power controller 14a-d back to the power supply 30, which occurs under regenerative braking.

Figure 2a is a more detailed schematic diagram of the control unit 12.

The control unit 12 includes four wheel sections 34a-d, one for each wheel of the e-vehicle 10, and a control routine section 36 in communication with each wheel section 34a-d.

Each wheel section 34a-d receives the output of a respective acceleration sensing unit 18a-d and outputs to the control routine section 36 data representing accelerations in axes co-incident with vehicle fixed body axes, ύ.. , V.. , W ₁ . a roll rate acceleration p and a yaw rate acceleration r , each of which is derived from lateral accelerations from each of the wheel accelerometer units.

As shown, each wheel section 34a-d communicates with each other wheel section 34a-d. in order to confirm which wheels are in traction to maintain a required

manoeuvre and to enable distribution of drive torque appropriate to maintaining that manoeuvre under variable reactive wheel load with the road surface.

Each wheel section 34a-d receives from the control routine section 36 a said wheel rate demand 32a-d which it outputs to the respective power controller 14a-d.

The control routine block 36, as well as receiving the data output from each wheel section 34a-d, receives the accelerator demand 20 and the braking demand 22. The control routine block 36 implements the control routine as described herein.

In this embodiment, the wheel sections 34b and 34d are assigned to the front two wheels (not shown) of the e- vehicle 10, which are steerable wheels. The wheel sections 34b and 34d receive the steering demand 24 directly and the resulting wheel positional feedback is passed to control block 36 in demanding an incremental wheel rate response as part of the EVTDS control routine.

Figure 2b provides an overview schematic diagram of the combined EVIDS control block diagram structure incorporating the wheel logic/traction controller in combination with the individual wheel control units that essentially control the e- vehicle. Each of the four wheel blocks fL_Wheel ...rRJWheel correspond to each of the four blocks 34a-d in Figure 2a. The wheel logic/traction controller receives and exchanges information between each wheel in establishing which wheel combination provides the required traction force to achieve the required forward speed and manoeuvre demand. U_demand is the speed demanded by the driver with Uo the achieved response. Steering demands η_steer applied to the two front wheels fL_Wheel and fR_Wheel produce an induced acceleration at each of the wheel sensors and in each of the three orthogonal axes. The combination of these inputs applied within the EVIDS control routine induces an incremental wheel rate δω which added to the wheel rate ω associated with the vehicle velocity demand identifies a demand wheel rate at each of the wheel stations. The achieved wheel rate ω ₀ under the wheel controller is then fed back to achieve an error over the demanded wheel rate. In a steady state turn when the demanded velocity is ^•

achieved, all four wheels will experience a wheel rate consistent with the demanded velocity and a distributed wheel rate increment demanded by the EVIDS controller to maintain the manoeuvre which in turn ensures stable dynamic control of the e- vehicle within controllable limits.

The ensuing analysis establishes how directional control of the e-vehicle 10 can be achieved by employing direct drive units at each wheel and the use of the acceleration sensing units 18a-d located at each wheel station to provide dynamic feedback.

Figure 3 shows the mathematical representation of individual wheel tracks for a given turning circle.

The notation used in Figure 3 and the following mathematical expressions are defined as follows :- Velocity vector in vehicle fixed body axes component Inertial frame of reference accelerations in axes co-incident with vehicle fixed body axes at vehicle centre-of-gravity components

Body rate vector in vehicle fixed body axes components rads/s ² ] Wheel positional vector from vehicle centre of gravity (CofG)

Velocity components at wheel station ij in vehicle body fixed axes [m/s]. Inertial frame of reference accelerations in axes co-incident with vehicle fixed body axes (accelerations measured by acceleration sensing units) components e.g. front(f), right(R) whee Velocity vector co-incident with instantaneous centre of rotation [Set = 0 in non-slip condition] [m/s]

r _cg Position vector of vehicle CofG from instantaneous centre of rotation (defined in vehicle body axes [ O ₅ -R, 0] [m].

Position co-ordinates of front right wheel from CofG of vehicle (assumed on centreline of vehicle. Modulus | | used in analysis takes the absolute value of the position co-ordinate with appropriate sign. Thus for front right wheel the position co-ordinates ar

R See definition of r _cg above [m].

ω Vehicle angular rate (in turning circle) - se above [rads/s] . η Steering angle [rads]

7 Wheel slip angle [rads]

Front right wheel angular rate [rads/s] Front right wheel radius [m]

ncremental Front right wheel angular rate [rads/s]

hi formulating a foundation argument to this problem, several assumptions have been adopted:- i) all vectors are defined in body axes of the e- vehicle 10; ii) the suspension of the e-vehicle 10 is infinitely stiff; iii) the e-vehicle 10 is executing an instantaneous turning circle radius r _cg at angular rate ω _o; iv) all wheel based acceleration sensing units are located in the chassis of the e-vehicle 10 at the wheel centre of rotation at a constant height above the road surface.

Let i => e {f,r} and j => e {L,R} then in general from Figure 3, the directional velocity of each wheel is given by:

The acceleration of each wheel in an inertial frame of reference (denoted by suffix V) is then given by:

Note this acceleration vector is that identified by a three-axis acceleration sensing unit set located at the wheel station (fixed in the chassis of the e-vehicle 10 - note assumption ii) above). The three sensor axes are assumed to be co-linear with the orthogonal axes of the e-vehicle 10.

From Figure 3 we assume a non-slip condition for all four wheels in contact with the road surface with the e-vehicle 10 steering in an instantaneous arc of a circle radius r _cg . It then follows that Vo = Wo = 0 Uo ≠O and the velocity vector at the centre of gravity of the e-vehicle 10 is given by:

But we note from Figure 3 that,

And therefore,

From equations 1 and 5, it follows from the no-slip condition that,

We also note from assumption iii) and equations 3 and 6 that after differentiation,

and,

(8)

Substituting into equation 2 therefore yields,

(9) Which from equations 3 and 6 yields

(10)

We now introduce the expanded form of the above vector expressions. To demonstrate a control approach, we concentrate our attention on one of the front wheels under steering to maintain the instantaneous circle assumed in the analysis.

Substituting into equation 10,

(12)

In equation 12, the components o are the accelerations measured in body axes of the e- vehicle 10 at the wheel station. If we rearrange these components we can identify the effective radius of turn associated with the motion of the e-vehicle 10. Thus,

(13)

Also from equations 5 and 6 we find that,

(14)

But the wheel plane is at a slip angle of to the directional path of the tyre therefore in the wheel plane the velocity components are given b here,

(15)

From equations 13, 14 and 15,

(16)

In the wheel plane therefore the velocity at the hub axis ere,

(17)

But we note here that from equation 12 when we set |xf|=|yfj=O that we derive the acceleration vector at the centre of gravity of the vehicle 10. Thus,

Therefore since the acceleration at the CofG can be measured or determined by calculation from the accelerations detected by all four acceleration sensing units 18a-d, it follows from equations 13, 17 and 18 that in the wheel plane for the front right wheel used in the example above,

Assuming a wheel radius of γJR, it follows that the wheel angular rate is then given by ω_ _fR (rad/sec) where,

Similar analyses exist for the left front wheel and aft wheels. But for aft wheels that are non-steering, η = 0.

The conclusion of the above analysis is that wheel rate and directional control of the e-vehicle 10 for a non-slip condition is readily monitored by application of the

translational acceleration sensing units 18a-d located at all four wheel stations within the chassis of the e- vehicle 10. This lends itself to a means of directional control under cornering.

It is essential that any directional control of the e-vehicle 10 at speed under cornering is associated with small changes in angular rate of each drive wheel from that associated with straight line speed. For wheels involved in steering, sharp changes in wheel angular rate (about roll axis) can lead to dramatic under- steer/over-steer and associated directional instability. Equations 19 and 20 enable these small changes to be identified.

To see how this incremental wheel velocity and associated wheel rate may be identified during steering we need to look at the limiting condition of equations 19 and 20 when the turning radius extends to infinity. In this instant equation 19 should yield a wheel speed equal to the speed of the e-vehicle 10 in straight line motion.

From equation 12,

(21) and from equation 18,

= — • -

(22) Substituting into 19 yields;

Therefore,

But if the e- vehicle 10 manoeuvres from a straight line, there is no initial wheel steering angle and therefore η=0. Thus in the limit as R→ oc^

where U is the forward speed of the e-vehicle 10.

It therefore follows that the incremental wheel rate associated with cornering from straight line motion at speed U is given by

If we now set the angular rate of the wheel in straight line motion to be

Wo_ _fR [ - U/r_ _fR ] then the total wheel angular rate during cornering is given by ω_ _fR where,

Directional control is maintained by control of the wheel rate (about hub axis) using the principle derived from equations 1-20 and 21-27 above. The use of incremental wheel rate change under cornering (equation 26) derived from feedback of the acceleration sensing units 18a-d and vehicle speed lends itself to a method of controlling the e- vehicle 10 to maintain stability during turning. Although the above analysis assumes an infinitely stiff suspension in order to simplify the analysis and to help identify the principles involved, it is the basis of the invention that this method of control can be adopted in a vehicle with Ml suspension.

Jn order to maintain wheel rate in accordance with this method of control, the e- vehicle 10 utilises a combination of pancake motor drive with regenerative braking. An important feature of the drive mechanism is the ability to maintain the appropriate level of torque to achieve the correct wheel rate for directional stability. This ensures that during temporary lifting of a wheel during cornering in which the wheel may rise off the road surface, the controller maintains the correct torque and wheel rate (via a system design feature) so that when the tyre regains contact with the road surface, there is no wheel slip (resulting in tyre wear). The

invention also ensures that there is no over-steer/uiider-steer which is outside control manageable response times to maintain stability under steering.

To understand the benefit of this control routine in the general case requires consideration of the general dynamic behaviour of the e-vehicle 10 under steering induced manoeuvre. In doing so, we consider the e-vehicle 10 to manoeuvre on level ground with the earth's gravity vector acting normal to the ground plane. In order to represent the dynamic response of a suspension system, dynamic rate accelerations about all three body axes and translational acceleration in the body z axis are assumed along with those accelerations associated with generation of forward speed change and centripetal acceleration which complete all three acceleration components in the inertial frame of reference acting in co-incident body fixed axes. As is shown below, pitch rate acceleration is assumed to be of minor significance in dynamic coupling induced translational accelerations and is therefore ignored.

The ensuing analysis shows how the use of three axis acceleration sensing units located at each wheel station may be used to determine all control acceleration inputs required (as part of a feedback control system) to control the e-vehicle 10. Initially the origin of these acceleration sensing units are assumed to be located in a rectangular frame within the chassis of the e-vehicle 10 and at each corner which in turn is co-located with each wheel hub. The plane is arbitrarily assumed to be at a perpendicular distance 'z' from the e-vehicle 10 centre of gravity although in determining wheel dynamic load distribution this is later ignored as a realistic assumption that makes evaluation of the wheel loads readily solvable.

Figure 4 shows the general dynamics of the e-vehicle 10 executing a turning manoeuvre.

We assume transient dynamic behaviour of the e-vehicle 10 under manoeuvre and small angle approximation between Earth and Body fixed axes. Let G _D be the drive torque generated by the wheels on the e-vehicle 10 and F _D the associated

traction force (Figure 4). It then follows in the general case that if F, is the force vector at the wheel interface with the road assumed level with gravity vector normal to plane containing all four wheels, then;

and,

Assuming small Euler angle approximation in pitch and roll at any instant in time so that e-vehicle 10 body X ₃ Y plane and Earth axis X ₃ Y planes can be considered co-incident, it follows that the drive torque G _D and traction force F _D can be defined by,

Note at this instant we are not considering steering forces, we assume the drivers intent is to instantaneously execute a yaw manoeuvre by putting down on the road a given drive and differential torque through the wheels which translates into an overall torque N _D and traction force along the axis of symmetry of XD with Y _D forming the centripetal force required to maintain the e-vehicle 10 moving in an instantaneous arc of a circle.

Figure 5 shows a wheel — ground friction interface.

Referring to Figure 5, it is assumed that the slip angle of a tyre under manoeuvre is defined by a scalar multiple ( /X _x , μ. _y ) of the normal loading (Z) of the tyre on the road surface and acting in an orthogonal axis set in the plane of the road surface. The resulting components of force acting on the tyre are thus μ _x .Z acting along the x axis, and μyZ acting along the y axis.

Based on this representation, and allowing for the fact that in general for any given tyre-road interfac ary with tyre load (Z) it follows that for the four tyres of the e- vehicle 10 at the road surface interface,

The inertial frame accelerations recorded by acceleration sensing units located at each of the four wheel locations and in co-incident body fixed orthogonal axes is given by,

where,

Here we assume that y/ y _a = y , i.e. front and rear wheels are in line (wheel hubs at corner locations of a rectangular configuration), and front and aft wheels are at the same height above the ground (z) as the centre of gravity of the e- vehicle 10 but at different for and aft distances from the centre of gravity.

Also by adding we get,

Therefore rearranging,

(35)

But from above,

(36)

Therefore following substitution,

(37)

Also we can now obtain the vehicle body rate accelerations and coupled dynamic terms using the acceleration sensors by differencing and solving the accelerator sensor feedback values. In doing so we make the important assumption that the pitch rate of the e-vehicle 10 and pitch rate acceleration in body fixed axes is of low significance by comparison to the yaw and roll rate when manoeuvring i.e. q « p and q « r and q « p and g « r .

Thus we write ω and ω as follows,

(38)

Applying the differencing of sensor accelerations,

Similarly,

(40)

Hence we can derive termsβ ,f ,{p ² +r ² ), pr, p ² ,r ² in terms of sensor accelerations.

Before doing so we introduce the following conventions;

(41) and,

whereupon,

These can then be used in equation 37 to get e.,

(49)

Therefore following substitution for p ² , r ² , pr, p and r in general terms we may write,

After final substitution of the above in equations 1 and 2 we arrive at,

where,

(53)

Expanding and simplifying these equations take the form,

From these equations we can now generate the following,

(60)

If we assume non-slip conditions along the longitudinal and lateral vehicle body inertial axes the therefore,

and,

(62) and,

(63)

Substituting from above,

We now introduce the notion of wheel (radius Ry,) drive torque D,j where,

(65)

and,

or,

and,

A

and,

or,

From the above notation we see that straight line motion associated with equation 61 yields

It therefore follows that since comprises acceleration terms from all four wheels that may be associated with encountering uneven road

surface conditions then from equations 61 to 64 the wheel inter-related associated normal load Z, _j for each wheel in turn determines the weighted drive torque demand to achieve the required wheel rate that maintains e-vehicle 10 stability in forward motion. Incremental wheel drive torques derived from equation 68 and attributed to lateral induced accelerations resulting from a steering command- induced-manoeuvre establish a lateral incremental wheel drive torque necessary to effect the manoeuvre. The vector sum of drive torque components establishes the overall drive torque required to maintain manoeuvre at each wheel station.

To understand the method of control we make use of equations 61 to 64 inclusive. Clearly tyre slip and traction coefficients of friction are not determinable and vary with reactive load at each tyre station and associated road interface conditions. To facilitate a method of control we therefore now consider how the acceleration sensing unit sensor feedback values may be employed.

hi equation 64 we assume that the acceleration sensing unit are co-planar with the centre of gravity of the e-vehicle 10 or as a good approximation z is small and can be set zero. With this assumption the matrix on the LHS of equation 64 has coefficients that are purely dependent upon the geometric format of the acceleration sensing unit locations — see equation 72.

(72)

By taking each wheel in turn and transferring load values associated with it to the RHS of equation 64 we arrive at three equations with three unknowns (normal loads for other three wheels). The column vector on the RHS of the matrix equation then comprises terms that involve the selected wheel reactive load and acceleration sensing unit values at each wheel station together with inertial data for the overall e-vehicle 10. Solving this equation enables us to identify the

reactive load relationship between each successive wheel and those remaining in contact with the road surface in terms of all four wheel sensor acceleration pack values. We note in particular that the third equation in this matrix accounts for the weight of the e- vehicle 10 allowing for vertical acceleration transients that may be attributed to a suspension system here represented in acceleration response terms only (suspension system is not modelled). Once the weighted loading between wheels is identified equations 61 and 62 provide detail of each wheel μ and μ associated with this wheel weighting. After substituting for μ _x.. and μ _y> in equation 63 actual wheel loading values are determined in terms of three axis acceleration sensing unit readings. It therefore follows that the instantaneous wheel loading for all four wheels in contact with the road surface may be determined from processing the accelerations from each of four three axis acceleration sensing units located at each of the four wheel stations.

In the event of wheel bounce in which for example one or two wheels momentarily bounce clear of the road surface and therefore provide no reactive load into the e-vehicle 10, the LHS of equation 64 maybe column and row vector reduced to represent the equations that are associated with wheels remaining in contact with the road surface. We then note in this case that the third equation effectively redistributes the weight of the e-vehicle 10 over the wheels that remain in contact with the road surface and involve a vertical acceleration component that maybe attributed in the real e-vehicle 10 to the response of the suspension unit to an impulsive load that caused the wheel(s) to bounce. Following the analysis through in the same manner as that for which all wheels remain in contact with the road surface shows that again the reactive loads experienced by all wheels may be determined by the acceleration sensing unit located at each wheel station.

Thus far, dynamic loading between tyres under manoeuvre has assumed that all tyres in contact with the road surface do not slip either in providing forward drive torque or manoeuvre (turning) torque onto the road surface. In the event that a wheel 'ij' slips by exceeding the slip angle i.e. μ ^■ y , _v .. >μ ^y« or loses traction i.e.

then this induces an acceleration sensing unit response at the associated wheel station which in turn reacts through the other wheels. This in turn causes a redistribution of load between wheels as terms in the RHS column vector of equation 64 change accordingly.

In the general dynamic analysis we demonstrated how at any instant in time the drive wheel reactive loads may be determined by a three axis acceleration sensing unit located at each wheel hub location. In equation 26 we identified the incremental wheel rate needed to maintain rolling contact motion of a wheel executing a turning manoeuvre at constant speed 'U'. We did so by taking the example where the e- vehicle 10 executed motion in an instantaneous arc of a circle. Although we looked at one wheel, the analysis is applicable to all four wheels of the e- vehicle 10 by simply changing the acceleration sensing unit readings and geometric parameters associated with its location on the e-vehicle 10 chassis. In the more general case defined above, if we assume the e-vehicle 10 to be travelling at an instantaneous speed 'U' while executing a manoeuvre in an instantaneous arc of a circle with centripetal acceleration then the incremental wheel rate for the same wheel as before (front right - fR.) is assumed represented by the same form but in this case the acceleration terms are modified by additional terms associated with the increased level of coupled dynamic behaviour attributable to additional coupled body rates and rate accelerations,

(73)

Equation 73 describes the more generalised form of equation 26 and unlike the former simplistic solution accounts for the coupled body rates and translational velocities and associated accelerations attributed to the dynamic response of a suspension system as opposed to assuming the suspension system to be infinitely stiff. The generalised expression forms the foundation from which the EVIDS wheel control routine is derived.

Although pitch rate (q) and pitch rate acceleration ( q ) is included in equation 73 it is arguably of negligible effect compared with roll rate (p), roll rate acceleration (P ) and yaw rate (r), yaw rate acceleration (r) but is included here for completeness.

This embodiment of the invention differs from the prior art in that each wheel experiences an incremental demanded wheel rate attributed to a steering demand induced manoeuvre. In implementing such an increase in wheel rate, the e-vehicle 10 is forced to use more traction rather than slip angle in negotiating a corner and in so doing reduce the instantaneous slip angle of each wheel. Manoeuvring therefore occurs with an increased contingency to experiencing slip angle limits and as such provides for an in-built level of improved dynamic stability compared with the prior art.