The invention relates to a suspension system according to the preamble of patent claim 1 as well as a method for operating a suspension system according to the patent claim 5.

Such a suspension system for at least two axles of a vehicle, in particular a commercial vehicle and a method for operating such a suspension system are already known from GB 2495231 A. the suspension system comprises an adjusting device configured to variably distribute loads among the axles. In other words, in said method, loads are variably distributed among the axles by means of the adjusting device so that a variable load distribution is realized.

It is an objective of the present invention to provide a suspension system and a method of the previously mentioned kind, by means of which a particularly efficient operation of the vehicle can be realized.

This objective is solved by a suspension system having the features of patent claim 1 as well as a method having the features of patent claim 5. Advantageous embodiments with expedient developments of the invention are indicated in the other patent claims.

In order to provide a suspension system of the kind indicated in the preamble of patent claim 1 , by means of which a particularly efficient operation of the vehicle can be realized, according to the present invention, the suspension system comprises a prediction module and/or a detection module. The prediction module is configured to determine at least one upcoming change of at least one condition in the surroundings of the vehicle, wherein the adjusting device is configured to distribute the loads on the basis of the determined change. The detection module is configured to determine a consumption of a medium by means of which the loads are distributed, the adjusting device being capable of being operated on the basis of the determined consumption. For example, said change of at least one condition in the surroundings of the vehicle can be an upcoming change of a road the vehicle travels along. For example, such a change can be a change of a grade and/or a surface, in particular a structure of a surface, of the road. Moreover, such a change can be a change of geometry, i.e. a course of the road. In other words, the road can comprise at least two subsequent portions. At first, the vehicle travels along a first one of said portions of the road. Since the vehicle travels along the road, the vehicle will reach the second portion and leave the first portion in the near future so that the vehicle will then travel along the second portion abutting the first portion. For example, the second portion is different from the first portion with respect to the grade and/or the surface and/or the course. This means, for example, the second portion can comprise a left turn or a right turn whereas the first portion is straight. This change of the road is a change of at least one condition in the surroundings of the vehicle, wherein this change of the at least one condition can be determined, i.e. predicted by means of the prediction module. Thereby, the load distribution, i.e. the distribution of the loads among the axles can be changed or such a change of the load distribution can be prepared when the vehicle still travels along the first portion and has not reached the second portion yet. Thus, the load distribution can be changed immediately after the vehicle has actually reached the second portion, i.e. the at least one condition has actually changed.

Alternatively or additionally, the at least one upcoming change of at least one condition can be a change of the weather and/or a change of other conditions and/or other variables which can be predicted by the prediction module. For example, the prediction module can predict or determine the change of the at least one condition on the basis of a digital map of the road, the digital map being stored in, for example, a memory of a navigation system of the vehicle. Moreover, the change of the condition, in particular the road can be determined on the basis of a current position of the vehicle, the current position being determined by means of the navigation system. For example, the navigation system can use a satellite navigation system such as GPS (global positioning system) to determine the current position of the vehicle. Additionally or alternatively, the upcoming change can be predicted by using audio and/or visual inputs, learned behaviour and/or other data sources.

With respect to the detection module, the suspension system can be configured as, for example, an air suspension system. In such an air suspension system the medium used for distributing the loads among the axles is air which is compressed by at least one air compressor. If there is a leakage in the air suspension system, compressed air can flow out of the air suspension system. Thereby, for example, the air consumption exceeds a predeterminable threshold value which, in turn, indicates that the air consumption is higher than it should be. In such a case, the leakage can be determined and, for example, the air compressor, in particular the adjusting device can be switched off in order to protect the air compressor and, thus, the suspension system from further damages and excessive wear.

In order to provide a method indicated in the preamble of patent claim 5, by which a particularly efficient operation can be realized, according to the present invention the method comprises at least one of distributing the loads on the basis of at least one upcoming change of at least one condition in the surroundings of the vehicle, and operating the adjusting device on the basis of a determined consumption of a medium by means of which the loads are distributed. Advantages and advantageous embodiments of the suspension system according to the present invention are to be regarded as advantages and advantageous embodiments of the method according to the present invention and vice versa.

Further advantages, features, and details of the invention derive from the following description of a preferred embodiment as well as from the drawings. The features and feature combinations previously mentioned in the description as well as the features and feature combinations mentioned in the following description of the figure and/or shown in the figure alone can be employed not only in the respectively indicated combination but also in other combination or taken alone without leaving the scope of the invention.

The drawings show in:

Fig. 1 a diagram illustrating axle load distribution at different tandem axle loads with a first target drive axle load;

Fig. 2 a diagram illustrating axle load distribution at different tandem axle loads with a second target drive axle load;

Fig. 3 a diagram illustrating axle load distribution at different tandem axle loads with a third target drive axle load Fig. 4 a diagram illustrating the relationship between traction coefficient and longitudinal slip; and

Fig. 5 a diagram illustrating fuel consumption delta from equalized load based on rolling resistance and slip.

With respect to Figs. 1 to 5, a suspensions system for at least two axles of a vehicle, in particular a commercial vehicle, and a method for operating such a suspension system are described. For example, the axles form a tandem axle so that the axles are rear axles extending from a first side to a second side of the vehicle. The vehicle is, for instance, a heavy duty commercial truck, wherein the axles are rear axles. For example, a first one of said axles is a driven axles, in particular a driven rear axle, wherein the second axle is an undriven axle, in particular an undriven rear axle. Additionally, the vehicle can have an undriven front axle, wherein the first axle, the second axle and the front axle carry two wheel-ends respectively so that the vehicle has six wheel-ends in a so-called 6x2 configuration, i.e. the vehicle has six wheel-ends with two driven wheel-ends mounted on the driven axle. A wheel-end can have one or more wheels on it, typically either two wheels (dual wheels or a dually) or a single wheel.

The driven axle is the axle in the axle tandem that receives power from an engine of the vehicle, the engine being, for example, an internal combustion engine. The driven axle receives the power from the engine via transmission and drive line. For adequate traction, a minimum amount of load must be applied to this driven axle. The undriven axle is the axle in the axle tandem that does not receive power and freewheels. The undriven axle may also be referred to as a tag or pusher axle. A minimum amount of load must be applied to the undriven axle simply to keep it from bouncing off a road surface.

Preferably, the undriven axle is configured as a lift axle. Such a lift axle is an axle that can be completely lifted off the road surface by means of a separate pneumatic, hydraulic, or mechanical linkage from that of the primary suspension system. A tandem axle comprises two axles at the rear of a six-axled truck or tractor.

By means of the suspension system and the method described in the following an optimal traction and fuel economy of the vehicle can be achieved at any given speed or in any condition so that a particularly efficient low-wear operation can be realized. The suspension system comprises an adjusting device configured to variably distribute loads among the axles of the axle tandem. Therein, for example, the suspension system can be configured as an air suspension system. Such an air suspension system having an adjusting device configured to variably distribute loads among the axles is, for example, shown in GB 2495231 A which is fully incorporated by reference herein. In such an air suspension system, the adjusting device comprises for each axle at least one first air bellow on the first side of the vehicle and at least one second air bellow on the second side of the vehicle, a pressure tank for storing compressed air, a valve system connected to the air bellows, and a control unit for the valve system, the control unit being adapted to electronically control the pressure in the air bellows by operating valves of the valve system. For example, the load on one of the axles can be increased by increasing the pressure in the air bellows belonging to that axle. By decreasing the pressure in the air bellows, the load on that axle can be reduced. Thereby, the load distribution among the first axle and the second axle can be adjusted.

In other words, the adjusting device is a means for selectively affecting the effective ground load through any given axle on the vehicle. As an alternative to an air suspension, a mechanical load distribution can be realized. Usually, such an adjusting device for adjusting the load distribution comprises a height sensor, a load or pressure sensor, an air compressor, a computer controller, and a series of valves. The controller is linked to a speed sensor or a vehicle controller bus such as a CAN bus that can feed the current speed, brake position, engine brake position, etc.

The suspension system described in the following comprises a basic system consisting of two primary modes. A first one of said primary mode is a so-called eco mode, the second mode being a so-called traction mode. Additional possible modes include equalized mode, dump mode, and fault mode. The transition between eco mode and traction mode is primarily by means of a speed sensor and timer. The suspension system described in the following is an all-inclusive system that optimizes traction, tire wear and fuel economy regardless of the vehicle speed, road conditions, or terrain. Further additional possible modes can include a so-called power on reset mode and a failsafe mode. In traction mode, by changing the air suspension bellows pressure, the majority of the load is placed on the drive axle to provide enhanced traction at low speed. In eco mode, the majority of the load is transferred to the undriven axle, which, when equipped with low rolling resistance tires will decrease the fuel consumption of the vehicle. A method of rapidly equalizing the driven and undriven axle bellows pressure is provided by an equalizer valve, which would activate in a hard braking event. As will be further described in the following, the load distribution can be adjusted on the basis of vehicle speed, wheel torque, engine brake consideration, predictive and preemptive load distribution, excessive wheel slip detection and mitigation and/or excessive air consumption detection and mitigation. The following inputs are preprogramed:

LDmin = minimum allowable load on the driven axle

LDmax = maximum allowable load on the driven axle

LUmin = minimum allowable load on the undriven axle

LUmax = maximum allowable load on the undriven axle

LTmax = maximum allowable load on the axle tandem

Htar = target ride height

Veco = velocity above which the vehicle should be in eco mode

Vtrac = velocity below which the vehicle should be in traction mode

Teco = time to wait before transitioning to eco mode once vehicle velocity is greater than Veco

Ttrac = time to wait before transitioning to traction mode once vehicle velocity is less than Vtrac, not applicable on startup.

The pressure in the air suspension bellows or similar method of applying a varying force can be correlated to a corresponding ground load produced by the axle when loaded. This relationship can be easily established by varying the pressure or force in the system and using wheel scales to determine the ground load.

Startup into traction mode:

When the vehicle is turned on while stationary, the vehicle will be in traction mode. To determine the load distribution between the driven and undriven axle, the following inputs are read into the controller from sensors on the vehicle. The measured inputs are:

LT = measured total load on the axle tandem, which is equivalent load on the driven axle plus the load on the undriven axle

H = measured ride height. Based on these inputs, the controller determines the following outputs based on the bellow scenarios and corresponding algorithms, which are then used by the suspension system to properly distribute the loads. These outputs are:

LD = load on the driven axle

LU = load on the undriven axle.

A first scenario is an underload scenario. If LT < LUmin + LDmin, then the axle is unloaded. Thus, for example, the load is equalized by setting LD = LU = (LD+LU)/2. A second scenario is a minimum undriven axle limitation scenario. If LDmin + LUmax > LT > LUmin + LDmin, then the undriven axle is loaded to its minimum rating and the remainder is put on the driven axle by setting LU = LUmin, LD = LT - LUmin.

A third scenario is a maximum driven axle limitation scenario. If LTmax > LT > LDmin + LUmax, then the driven axle can be maximized and the remaining load is put on the undriven axle by setting LD = LDmax and LU = Lt - LDmax.

A fourth scenario is an overload scenario. If LT > LTmax, then the axle is overloaded. Then, for example, the load is equalized LD = LU. The driver is warned via an audible and/or visual cue that the axle is overloaded. The ride height should be maintained by pressurizing or depressurizing the air bellows for both axles equally to achieve the target height.

Transition from traction mode to eco mode:

Once the vehicle begins moving, at predetermined intervals (e.g. every second), the vehicle velocity, V, is read from CAN bus to the controller. If V > Veco, then the controller starts a timer. The timer is reset and stopped if V < Veco and the vehicle remains in traction mode until V > Veco and the timer is started again. If and when the timer exceeds Teco, the vehicle enters eco mode. To determine the target load distribution between the driven and undriven axle for eco mode, first, the following inputs are read into the controller from sensors on the vehicle:

Measured inputs: LT = measured total load on the axle tandem, which is equivalent load on the driven axle plus the load on the undriven axle

H = measured ride height.

Based on these inputs, the controller determines the following outputs based on the below scenarios and corresponding algorithms, which are then used by the suspension system to properly distribute the loads.

Outputs:

LD = load on the driven axle

LU = load on the undriven axle.

In the underload scenario if, LT < LUmin + LDmin, then the axle is underloaded so that the load is equalized by setting LD = LU.

Now, the second scenario is a minimum driven axle limitation, in which, if LUmin + LDmax > LT > LUmin + LDmin, then the driven axle is loaded to its minimum rating and the remainder is put on the undriven axle by setting LD = LDmin, LU = LT - LDmin.

The third scenario is a maximum undriven axle limitation scenario, in which, if LTmax > LT > LUmin + LDmax, then the undriven axle can be maximized and the remaining load is put on the driven axle by setting LU = LUmax and LD = LT - LUmax. Moreover, the fourth scenario is an overload scenario in which, if LT > LTmax, then the axle is overloaded and the load is equalized by setting LD = LU. Then the driver is warned via an audible and/or visual cue that the axle is overloaded. The ride height should be maintained by pressurizing or depressurizing the air bellows for both axles equally to achieve the target height.

Transition from eco mode to traction mode:

While the vehicle is in eco mode, at predetermined intervals the vehicle velocity, V, is read from the CAN to the controller. If V < Vtrac, then the controller starts a timer. The timer is reset and stopped if V > Vtrac and the vehicle remains in eco mode until V < Vtrac and the timer is started again. If and when the timer exceeds Ttrac, the vehicle reenters traction mode. To determine the axle load distribution for reentering traction mode, the same inputs and algorithm are followed as in the startup into traction mode section above until the truck is turned off.

Additional possible modes:

An additional possible mode can be a dump mode which allows the user to lower the suspension when the vehicle (truck) is moving slowly to couple and decouple from trailers. A further additional possible mode is a fault mode which occurs when there is an error in the system such as a leaking or punctured bellows. The driver is warned via audible and/or visual indicator. A further additional possible mode can be an equalized mode which can be made available to the driver when he would like the load on the axles to be equalized.

Variations and additions to the basic system:

Another type of system that can be realized is one that has more than two primary modes. This may be desired because there may be conditions when the optimal balance of fuel economy and traction lies between the two extremes of the previously mentioned traction mode and eco mode. Thus, there may be any number (including infinite) of modes between full eco and full traction mode as previously described. The goal of this system is to have enough traction to not excessively slip the tires, but still be achieving the best fuel economy possible. This is realized by attempting to have the least amount of load on the drive tires possible without slipping. The result of this weight distribution is that the greatest amount of load possible will then be on the low rolling resistance tires on the undriven axle, thus giving the lowest rolling resistance while still providing adequate traction.

Besides the basic two mode system, the next simplest system would have three steps - full eco, equalized, and full traction mode. Another example would be with five steps: full eco, semi-eco (half way between full eco and equalized), equalized, semi-traction, and traction. If full eco has a load distribution of 80% and 20% on the undriven and driven axle, full traction mode has a load distribution of 20% and 80%, and equalized mode has a load distribution of 50% and 50%, then semi-eco mode could be 65% and 35% and semi-traction mode could be 35% and 65%. The two primary and distinctly separate methods to determine where on the continuum the axle load distribution should be are speed based and wheel torque based. Speed based variable distribution:

A first method to determine which of the three or more load distributions the system should be in is by assigning an operating speed range to each mode such as the basic system is broken into two modes using speeds. For example, for three modes, there is a speed with a hysteresis between low speed and medium speed modes, a speed with a hysteresis between medium speed and high speed modes. Each mode has a set load distribution such as 80% and 20%, 50% and 50%, 25% and 75% for the driven and undriven axle, respectively, for example. Alternatively, each mode can be assigned a target drive axle load and can use an optimal axle load distribution from target driven axle load algorithm shown below.

Torque based variable distribution:

An alternative, second method to determine which load distribution the system should be in is based on the torque applied at the wheels. To maintain traction, the drive tires need a greater normal force when they are attempting to transmit a greater torque. Therefore as the torque through the drive wheels increases, as may be the case when decreasing speed (and thus downshifting the transmission), driving up an incline, or towing a heavy load, the downward force on the driven axle must be increased. This results in a decrease in load on the undriven axle which is equipped with low rolling resistance tires. Thus, the fuel economy is decreased when more torque is required. The optimal balance of adequate traction and fuel economy is to have the minimum amount of load on the drive axle for traction so that remainder of the load can be placed on the undriven axle.

A first step is to determine a target driven axle load, which is the minimum axle load for traction. This may be calculated in a number of ways. In this example, the target driven axle load is calculated from the equations for static friction, but there are other more complex tire traction equations and models that can be used instead.

The target driven axle load is calculated from the equation for static friction:

Ffriciton ^' = Ν _∞Ιβ μ& Φ (1) Where F is the force of friction, N is the normal force on the object, μ is the coefficient of friction, and Φ is the angle between the normal force and the surface the object is on.

For example, N is the load on the drive axle, which can be varied by changing the pressure in the air bellows.

Opposing the force of friction is the tractive force generated by the engine. This is calculated from the equation.

Where F _tractive is the tractive force generated by the engine, T _en gine is the output torque of the engine, Rtransmission and Rfinaidnve are the selected transmission gear ration and the final drive ratio, respectively, and r _tir e, effective is the effective radius of the drive tires.

The goal is to place as little load on the drive tires while still providing adequate traction, i e. Ftractive is the minimum force required and thus load to eliminate drive wheel slip. Setting Ff _rict ion to equal F _tra ctive will achieve this, but realistically there must be a safety factor, SF, that will ensure there is adequate traction because these equations cannot account for variability in parameters and conditions (e.g. SF = 1.1). SF is multiplied by Ftractive to account for any variations in system application: X F ^~ F

T x R x R

x £F = sin Φ (4)

Solving for N _ax i _e gives

T x R x R

N _axk = x SF (5)

X ^ ^ an equation from which the optimal drive axle load can be calculated. LDtar = target drive axle load, which equals _ax ie as calculated above.

Because the torque at the wheels is constantly changing, a best practice is to use a method of averaging out the torque signal using a running average over a preselected period of time with a hysteresis rather than the instantaneous torque signal.

Implementing the target drive axle load:

Achieving the target load on the driven axle is limited by the axle load limits mentioned above in the sections explaining eco and traction mode from the basic system. That is, depending on what the target driven axle load is, it can be limited by either underload, minimum driven axle limitation, maximum driven axle limitation, minimum undriven axle limitation, maximum undriven axle limitation, or overload. These limitations create an envelope in which the driven axle load can exist in. If the target drive axle load is above or below one of these limitations, then it will be set to the maximum or minimum distribution allowed by the envelope.

Figs .1 and 2 show plots, respectively, which were generated for various tandem axle loads using inputs and limitations, for target drive axle loads of 1,000 lbs. (Fig. 1) and 22,000 lbs. (Fig. 2) respectively.

The plot in Fig. 1 illustrates how the system will distribute load if the target driven axle load is below the minimum allowable driven axle load (LDmin) (LDtar of 1 ,000 is less than LDmin), so LDtar loading cannot be achieved and is limited by LDmin. The plot in Fig. 2 illustrates how the system will distribute load if the target driven axle load is above the maximum allowable driven axle load (LDmax), i.e. LDtar of 22,000 is greater than LDmax, so LDtar loading cannot be achieved and is limited by LDmax. In Figs. 1 and 2, the underload state is designated by U, the minimum driven axle limitation is designated by MLmin, the maximum driven axle limitation is designated by MLmax and the overload state is designated by O. Moreover, a respective graph 10 illustrates the driven axle, and a graph 12 illustrates the undriven axle. The respective abscissa 14 of the diagram shows the load as a sum of all axles, and the ordinate 16 shows the individual axle load. Said envelope is calculated using the algorithms for the basic system. For a high target drive axle loading, the envelope is generated using the algorithm for traction mode. For a low target drive axle loading, the envelope is generated using the algorithm for eco mode.

If the target drive axle load is within the bounds of the envelope, then the drive axle is set to that ideal loading or to a mode that corresponds to that loading. An algorithm such as a source code shown below can be used to determine the optimal driven (LD) and undriven (LU) axle loads given a target drive axle load from the above torque and/or speed calculations.

Optimal axle load distribution from target driven axle load algorithm: %lnput Examples

LT = 14000; %Current axle load (Driven Load + Undriven Load)

LDmin = 7000; %Minimum allowable load on the Driven axle LDmax = 20000; %Maximum allowable load on the Driven axle LUmin = 4000; %Minimum allowable load on the Undriven axle

LUmax = 20000; %Maximum allowable load on the Undriven axle

LTmax = 34000; %Maximum allowable load on the axle tandem

LDtar = 10000; %Target load as determined by calculations

%Calc Traction mode LD (Driven Axle Max)

%Test for Underload

if LT <= LUmin + LDmin

LDtrac = LT/2;

%Test for Minimum Undriven Axle limitation

elseif LT <= LUmin + LDmax

LDtrac = LT-LUmin;

%Test for Maximum Driven Axle limitation

elseif LT <= LTmax

LDtrac = LDmax;

%Test for Overload elseif LT > LTmax

LDtrac = LT/2;

end

%Calculate Economy mode LD (Driven Axle Min)

%Test for Underload

if LT <= LUmin + LDmin

LDeco = LT/2;

%Test for Minimum Driven Axle limitation

elseif LT <= LDmin+LUmax

LDeco = LDmin;

%Test for Maximum Undriven Axle limitation

elseif LT <= LTmax

LDeco = LT-LUmax;

%Test for Overload

elseif LT > LTmax

LDeco = LT/2; end

%Calc Optimal Driven Axle Load

%lf Target Axle load is less than Eco Mode Driven Axle Load (LDeco), %then set Driven Axle Load to Eco Mode Driven Axle Load

if LDtar <= LDeco

LD = LDeco;

%lf is Target Axle load is greater than Eco Mode Driven Axle Load, but %less than Traction Mode Driven Axle Load, then set Driven Axle Load to

%Target Axle Load

elseif LDtar <= LDtrac LD = LDtar;

%lf is Target Axle load is greater than Traction Mode Driven Axle Load, %then set Driven Axle Load to Traction Mode Driven Axle Load elseif LDtar > LDtrac

LD = LDtrac;

end

%Calc Optimal Undriven Axle Load, which is equal to LT-LD LU = LT-LD;

Fig. 3 shows a further diagram which was generated using the above code and using the inputs shown in the following table:

The diagram in Fig. 3 illustrates how the target drive axle load can be achieved with the proper total axle load and is limited for other total axle loads. In Fig. 3, the achieved target drive axle load of, for example, 10,000 lbs. is illustrated by AL.

Engine and service brake consideration:

When travelling downhill in eco mode or similar distribution of load on the undriven axle, it is possible that the braking torque of the engine brake on the driven axle could be less effective or could cause the axle to lock up, especially when used with the service brakes. To counteract this effect, the vehicle's tandem axle load distribution should be adjusted to have more load on the driven axle when the engine brakes and/or service brakes are used. When the driver activates the engine brake, a representative signal is sent over the CAN bus. The controller which is, for example, an ECU (electronic control unit) can detect this signal and will then assume that the engine brake is being used. When the engine brake is activated, the system should enter a mode with a load distribution biased more to the driven axle. For example, the system could change to equalized or traction mode or it could use the torque-based distribution algorithm by substituting the absolute value of the engine brake torque for the engine output toque (T _eng ine) in equation (2) an include an appropriate safety factor.

Steering input consideration:

Because the effective wheelbase of the vehicle changes when undergoing a load distribution change and this wheelbase change affects handling, it is best if the system does not transfer load during a steering event. On vehicles that are equipped with steering angle input sensors, or other means of determining if the vehicle is turning, these signals can be read by the suspension controller from the CAN bus. When there is a detected amount of steering above a preprogramed threshold, then the suspension controller will suspend operations until the steering manoeuvre is complete.

Predictive, preemptive, and grade-based load distribution:

Using GPS (global positioning system) and map data, audio and/or visual inputs, learned behaviour, and/or other data sources, an algorithm can determine when to change the suspension load distribution prior to it being needed for varying road grades, curves, conditions, or other variables. For example, patent application DE 10 2008 038 078 uses such technology for cruise control, transmission gear selection, and brake control, but not in relation to axle load distribution. One example would be using GPS and road grade data to predict that the truck will need more torque to drive up a steep grade. The truck would then distribute load to the drive axle just before it is needed so that there is adequate traction without slipping and losing momentum. The predictions could include but are not limited to: a) Predict inclines and transfer load to driven axle for additional torque transfer b) Predict declines and transfer weight to the driven axle for additional engine brake torque transfer c) Predict flat terrain and transfer weight to the undriven axle for additional fuel savings d) Predict icy/snowy conditions when ambient temperature is near freezing and transfer weight to driven axle

e) Predict continuing low traction situations when low traction is determined and

temperature is near freezing

f) Predict curves in road and suspend changes in load transfer

g) Predict upcoming speed increase or decrease and determine optimal time to transition into eco mode or traction mode.

h) Predict weigh stations and equalize weight

i) Predict bridges, changes in state and/or territory and/or country boundaries, or toll road that require certain weights on each axle. May also lift or lower an axle if the vehicle is equipped with a lift axle if above structure or location require and/or regulate fewer or greater axles.

For example, the traction of the driven axle can be increased in a need-based manner by lifting the lift axle off the road surface, when/while, for example, the vehicle has a vehicle speed which is below a predeterminable threshold value. Said threshold value for the speed can be predetermined by law or by the vehicles specification such as its admission.

To predictively change axle distribution based on the required engine torque, the target drive axle load can be calculated by the following.

The total running resistance of a vehicle is the sum of its aerodynamic drag, rolling resistance, and climbing resistance. Assuming that the aerodynamic drag and rolling resistance can be calculated from programmed vehicle parameters or from learned vehicle behaviour at a given speed, then only the climbing resistance must be calculated to determine the additional torque required to climb or descend a grade and thus what the target driven axle load should be.

Climbing resistance is given from the following equation:

F _eKmh = W$\n<[> (6)

Where F _C | _imb is the climbing resistance force, W is the vehicle weight, and Φ is the angle of the road grade from horizontal. The upcoming road grade angle can be calculated by comparing the vehicle's current GPS location, analyzing the upcoming road grades, and averaging them for a given distance. The upcoming average grade can be converted into an angle and plugged in as Φ. The vehicle weight, W, can be inputted to the suspension controller by the operator after weighing the vehicle, from an onboard mass sensor in the transmission, or from other sensors.

Next, the additional torque that will be required due to the grade can be back calculated using equation 2 and solving for torque:

T x R R ~ „ U

The new engine torque that will be required is the sum of the current engine torque and the climb engine torque.

T = T + T (7)

The new climb torque can be plugged into equation 5 for the engine torque and the rest of the calculation can be completed. Note that the vehicle will likely have to change transmission gears when approaching the hill. The climbing transmission gear can be predicted using technology like that in patent application DE 10 2008 038 078 and the predicted transmission ratio should be used instead of the current one when calculating the additional climb torque.

Compensation for other effects of load transfer:

There are multiple unintended consequences of redistributing load to the undriven axle that can actually increase fuel consumption, which is counter to the goal of the method. These are described below:

1. Wheel slip: The amount a tire slips is a function of the torque applied and the vertical load on it. Decreasing load on the drive tire increases slip, thus requiring more power and fuel to achieve the same speed. This is described more specifically below. 2. Changes in tire wear and tire type: If the tires wear in such a way that the undriven axle tires no longer have significantly lower rolling resistance than the driven axle tires, then the tire system will no longer reduce the vehicle's rolling resistance and may actually increase as load is transferred to the undriven axle. Similarly, if tires are used that do not have a significant difference in rolling resistance or a tire with higher rolling resistance is placed on the undriven axle, then the vehicle's rolling resistance may increase as load is transferred to the undriven axle.

3. Unequal bearing loading: If the wheel bearings have significantly greater friction at higher loads (i.e. relationship between load and friction is non-linear), then increasing the load on one of the axles can cause increased friction.

4. Driveshaft and axle misalignment due to frame flex: When load is unequally distributed on the axles, this will cause the frame to flex in a different way than when the axles are loaded evenly. This can cause an increase in driveline angle which can affect the driveline efficiency or it can misalign axles that use a track-rod type suspension.

5. Ride height change: If the ride height is not accurately controlled (best practice is to have a ride height sensor on each axle and average the values to determine the ride height), then the real ride height can change as the frame flexes. This can result in loss of fuel economy due to aerodynamic drag increase caused by changing the side skirt clearance, front trailer height, etc.

6. Increased onboard air loss: If there is a small leak in the system, distributing a higher pressure to one set of bellows will increase the rate of air loss. Also, if the system requests a load redistribution, but the ride height is not achievable, then the system may fight itself to achieve the ride height and waste air in the process. Wasting pressurized air increases the duty cycle of the air compressor on the vehicle, which consumes power and fuel.

The above considerations increase fuel consumption in opposition to the fuel

consumption reduction caused by distributing load onto the low rolling resistance tires on the undriven axle. Therefore, the purely theoretical optimized loading for the best fuel economy as described in the patent application WO 2013 119 252 is not completely accurate and may result in a diminishing return as the load is shifted further to the undriven/low rolling resistance tires.

The best and most definitive, if time consuming, method to account and compensate for these possible losses is to empirically test the trucks to determine which load distribution results in the optimal fuel economy.

To achieve this, the truck must be tested at multiple total axle loads and load distributions while recording the fuel consumption. Because the fuel economy improvements are likely to be less than 1%, performing this test on a controlled test track with minimal outside influence is ideal. From the fuel consumption data, a look-up table can be generated that will contain the optimal load distribution for fuel economy for a given total axle load, which may not necessarily be the loading with the greatest load on the undriven axle.

As it relates to the systems and variations, the look-up table of optimal loading for fuel economy would replace the upper bound for the undriven axle load. That is, there is no reason to distribute any more load to the undriven axle above what is listed in the look-up table. For the basic system, the ECU would use the look-up table value as the desired eco mode distribution and the traction mode distribution would be the same as previously described. For the variable distribution system, the look-up table values become the bounding envelope for maximum load on the undriven axle.

Wheel slip compensation:

When a torque is applied to a pneumatic tire, it is assumed that that torque causes the outer surface of the tires that are in contact with the road surface to have the same tangential speed as the forward velocity of the vehicle and the same as the tangential speed of the undriven tires. However, it has been shown empirically that there is always some percentage of slip that occurs in a driven tire. This slip is a function of the normal force (vertical load) and the traction force (torque/radius) on the tire and is essentially linear when the normal force is much higher than the traction force as is the case in most steady-state driving cases.

Slip is typically characterized by the slip ratio: Tangential velocity of tire edge - Vehicle velocity

Slip ratio = x 100%

Vehicle velocity

When the slip increases, the tires and engine must rotate faster. The increased engine speed with the same torque equates to increased required power and thus an increased in fuel consumption.

Therefore, to truly optimize the load distribution for traction and fuel economy, the effect of slip must be taken into account. One method besides the empirical method mentioned above is to use theoretical models to estimate how the effects of load distribution and torque relate to slip and thus to a fuel consumption increase. This can then be compared to the theoretically optimized load distribution as described and an optimization can be found from those two datasets.

Fig. 4 shows a diagram illustrating an approximate relationship between traction coefficient shown on the ordinate 18 and longitudinal slip shown on the abscissa 20. The power loss attributed to slip can be calculated using equation 8.

The power change can be compared to the total power consumption to find the fuel economy loss. The most accurate way to find the fuel loss is to compare the location of the operating point with and without a load redistribution on the brake specific fuel consumption (BSFC) map of the engine, but this is a more complicated operation. rad

Engine torque[W] x Engine speed x Slip[%]

APower[W] = - (8)

+ Slip[%]

Slip threshold limit:

Another way to approach the problem of tire slip is to determine a threshold slip beyond which it is undesirable to operate due to diminishing returns of the axle load distribution. Once this is determined by whatever methods (empirical study or mathematical modelling), it can be compared to the current slip ratio of the vehicle as described below. Because it is difficult to accurately measure the velocity of a vehicle, the wheel speed ratio between the driven and an undriven axle can be used as a measureable substitute for the slip ratio described above:

Driven axle rot. velocity - Undriven axle rot. velocity

Wheel speed ratio = (9)

Undriven axle rot. velocity

The wheel speeds can be recorded to the ECU using the wheel speed sensors that vehicles with Antilock Brake Systems (ABS) are equipped with. To obtain a baseline wheel speed ratio, the controller will equalize the load on the axles for a predetermined time, record the wheel speed ratio, and save it in the controller memory before

transitioning to the desired biased axle load distribution.

Once the baseline wheel speed ratio is obtained, the current wheel speed ratio is monitored. If a load distribution change causes the wheel speed ratio to change more than a predetermined percentage above the baseline (wheel speed ratio percentage increase threshold), then the distribution will be biased more towards the driven axle until the wheel speed ratio is below the wheel speed ratio percentage increase threshold.

Input parameter method:

The following is a description of how a system can use provided inputs (either manually entered, automatically entered, or measured) to optimize the fuel efficiency and traction of the vehicle at all times.

Instantaneous Tire Conditions:

Pressure: The tire pressure affects the contact patch size and stiffness of the tire. It can be measured manually and manually input into the ECU via a human interface or it can automatically be monitored using a tire pressure monitoring system (TPMS) and be electronically sent to the ECU.

Temperature: The tire temperature affects many characteristics of the tire because the tire materials behave differently at different temperatures. It can be measured manually and manually input into the ECU via a human interface, it could be derived from a pressure sensor in the tire using the ideal gas law or similar equation, or it can automatically be monitored using a thermocouple or infrared thermometer and electronically sent to the ECU.

Tire Wear State (Tread Depth): As the tire wears, tire tread depth decreases, resulting in a lower rolling resistance and other property changes. If the tires wear in such a way that the difference in rolling resistance between the driven and undriven tires becomes negligible or the driven tires becomes lower rolling resistance than the undriven tires it can negate the fuel economy improvement of the system unless it is accounted for. It can be measured manually using a tread depth gauge and manually input into the ECU via a human interface, measured using an imbedded onboard device, or it can automatically be monitored using a laser measurement device and electronically sent to the ECU.

Normal Force: The normal force on the tire, or load, can be calculated from the suspension bellows air pressure using the relationship between pressure, load, and height as given by the suspension manufacturer.

Longitudinal Force: The longitudinal force on the driven tires is caused by the engine torque. The inputs of current engine torque, current transmission gear selected, rear axle ratio, and tire revolutions per mile are required for this calculation.

Road Surface Conditions: The tire rolling resistance and slip are a function of the coefficient of friction and other factors of interaction between the road surface and the tire. Generally, the fewer asperities (smoother) the road, the more slip will occur, but the rolling resistance will be lower. The road surface is generally very hard to measure and can change often over the course of the vehicle's route. The surface can be characterized using laser scanners that can determine the relative roughness of the road, but cannot necessarily determine the hardness of the surface. It is possible for the road surface conditions to be manually input into the ECU via a human interface. If the road surfaces can be generalized by geographic location or individual road then a look-up table of road surfaces can be referenced by the ECU using GPS data equipped on the vehicle.

Besides the road surface, the presence of any substances on the road surface will affect how the tire interacts with it. Water, snow, ice, oil, dirt, or debris are examples of these substances. Water, snow, and ice can be predicted to be present using rain sensors in combination with ambient temperature sensors. This information can be sent to the ECU. Also, if the tires are slipping excessively without another explanation, it may be that there is a foreign substance on the roadway and the system should compensate for this.

Velocity: The tires characteristics change as the tire rotates at different velocities.

Presumably all the tires will be affected equally, but if not, then the differential rolling resistance and slip between the normal and low rolling resistance tires could change as the vehicle changes velocity.

Tire Properties:

Slip and Rolling Resistance Relationships: To accurately model the tire slip

characteristics, the relationship between slip and tire pressure, temperature, normal force, tangential force, tread depth, and surface condition should be known. Likewise, to accurately model the tire rolling resistance characteristics, the relationship between rolling resistance and tire pressure, temperature, normal force, tangential force, tread depth, and surface condition should be known. These relationships are different for every tire and are best established through empirical methods, although computer modelling can generate estimates of the relationships if the tire materials and geometry are well known.

Revolutions per mile: The number of revolutions the tire does in a mile is required to calculate the tangential force between the tire and the road surface from the engine torque.

Vehicle Parameters:

Current engine output torque: The engine output torque is broadcast over the CAN bus and is needed to calculate the tangential force on the tires.

Rear Axle Ratio: The rear axle ratio is required to calculate the tangential force between the tire and the road surface from the engine torque. This can be manually entered by an operator. Current transmission ratio: The current transmission ratio is required to calculate the tangential force between the tire and the road surface from the engine torque. This can be read from the CAN bus.

Use of Parameters: One way to estimate the instantaneous coefficient of rolling resistance given the above inputs and relationships is to create best fit relationship equations from the data and compare the data to a baseline scenario such that the equations are in the form of where P is a parameter. The normalized parameters can then be multiplied together to achieve an estimate of the real coefficient of rolling resistance.

For example, the equation relating to vertical load, F _z , is commonly expressed as:

Slip can be calculated similarly, where the equations are in the form:

The normalized parameters can then be added together to achieve the total slip.

Alternatively, an empirical study can be performed to determine the real interaction between each parameter and a more accurate calculation of C^ and slip can be established rather than multiplying or adding the C„ or slip factors together as this method may not be accurate enough. The calculation of real Crr and slip can then be used to calculate the change in fuel consumption that would result in different load distributions for a given vehicle weight. That is, the ECU can run an algorithm calculating the aggregate fuel consumption change from the rolling resistance change.

To calculate the fuel consumption change from the load distribution, the percentage of load on each tire is multiplied by the C _rr of that tire. This is compared to a baseline figure and gives the percentage change in rolling resistance. Generally, for heavy trucks, a 3% improvement in rolling resistance results in a 1 % improvement in fuel economy. This rule or other correlation can be used to find the percentage change in fuel consumption. Calculating the fuel consumption change due to slip was outlined previously.

To optimize fuel economy, the fuel consumption increase/decrease due to slip must be subtracted from the fuel consumption decrease/increase due to the improvement in rolling resistance to find the net fuel consumption change. From this new relationship between fuel consumption delta and percent load shift, an optimized solution can be found for eco mode in the simple system or using the torque-based traction optimization system. An example plot generated from slip and rolling resistance data is shown in Fig. 5, wherein the percent load shift (load on tag axle) is shown on the abscissa 22 and the fuel consumption delta in [%] is shown on the ordinate 24.

In Fig. 4, a graph 26 illustrates the fuel consumption delta from slip and a graph 28 illustrates the slip and a graph 28 illustrates the fuel consumption delta from C _rr .

Moreover, a graph 30 shows the fuel consumption delta in total as a sum of the graphs 26 and 28.