SE542141C2 - Method and system for positioning an air deflector - Google Patents

Abstract

The present invention relates to a method (100) and system (300) for providing a reference data set (150) for determination of an optimum position of an air deflector (3, 4) for guiding an airflow around a vehicle (1, 2) during driving. The invention further relates to a method (200) and system (300) for positioning an air deflector (3, 4) on a vehicle using a reference data set.

Description

METHOD AND SYSTEM FOR POSITIONING AN AIR DEFLECTOR Technical field The disclosure is related to a method for optimizing a position of an air deflector guiding an airflow around a vehicle, more specifically a ground automotive vehicle, during driving in order to minimize air drag and fuel consumption. The disclosure is also related to an air deflector system of a vehicle.

Background Vehicles, such as heavy trucks carrying a load and trucks towing a trailer, are often equipped with air guiding devices, air deflectors, located on the roof of the truck cab and/or on respective lateral sides of the truck cab, for reducing air drag during journey and thereby reducing fuel consumption of the vehicle.

When travelling, vehicles displace air, and for a vehicle carrying load or towing a trailer which extends upwards and/or sideways beyond the vehicle, the front surface of extended parts of the load or trailer will cause significant air drag. The air guiding device, if correctly aligned, redirects the air which is moving over the moving vehicle in such a way that the air does not hit the front edge of the trailer or the load carried by the vehicle but instead the air is moving over the roof or along the sides of the carried load or the trailer.

The dimensions of trailers and of load carried by the vehicle often vary, why for optimal reduction of the air drag, the air guiding devices should be adjustable and adjusted for each truck-load or truck-trailer combination.

GB-2465393 discloses an air deflector system for use on a commercial vehicle. The system comprises air guiding devices and a plurality of pressure sensors provided on the air guiding devices. The pressure sensors sense the pressure on the air guiding devices during the travel of the vehicle. The sensed pressures form output signals transmitted to a control unit, which is configured to control the position of each air guiding device so as to alter the aerodynamic profile of the vehicle in response to the outputs of the pressure sensors, a speed signal and data relating to the geometry of the vehicle.

EP2626281 A1 discloses an air guiding device for automotive ground vehicles in which the air guiding device can be automatically adjusted by means of an adjuster in the form of an electric motor. A parameter corresponding to the adjustment force required by the adjuster during movement of the air guiding device from a first position to a second position is registered by a control unit. The control unit then adjusts the air guiding device into a position corresponding to a minimum value of the registered parameter, which position is a position corresponding to the lowest air resistance for the vehicle. For some circumstances it has been shown that the position corresponding to the minimum value does not correspond to the lowest air resistance.

EP2915726 A1 is building on the principles of EP2626281 A1. If the conditions during the adjustment process of EP2626281 A1 have been anomalous, for example, by the air guiding device being influenced by wind gusts, displaced air from oncoming vehicles or similar, the parameter that has been determined corresponding to the optimal condition for the lowest possible air resistance may be erroneous, such that the air guiding device is adjusted to a non-optimal condition, which results in an increased fuel consumption. In the method disclosed in EP2915726 A1 also the pressure of oncoming air against the vehicle in its direction of travel is continuously determined (e.g. by the air quantity gauge of the vehicle). The time course of the registered parameter is determined. Based on the determined pressure of the oncoming air and the time course of the registered parameter, the reliability of the time course of the parameter is determined. A further method is disclosed in SE539431 in which signal peak amplitudes or pulse center frequencies are used to identify a desired position of the air deflector.

The known techniques disclosed above either need a plurality of sensors, or need to match an absolute value of a physical quantity of a subsystem with that of another adjacent system. The disadvantage of the former relates to the cost of the plurality of sensors as well as the challenge of accuracy when sensing parameters in discontinuous positions. The latter has the disadvantage that no matter how simple the construction, i.e. the support structure of the air guiding device, may be, the construction will need to be kept in a good condition to not alter the balance in the physical quantity one rely on. The harsh environment on the exterior of a heavy truck will expose any mechanical construction to mechanical wear, dirt, rust, etc., which may alter the variation of the physical quantity needed to move the air deflector from a first to a second position.

Consequently, there is a need for a technique of determining an optimum position of an air deflector of a vehicle and to position the air deflector in such position, in a cost-efficient and reliable way.

Summary The object of the present invention is to provide a method for, by means of an electric actuator, positioning an air guiding device in an optimal position resulting in minimum air drag and fuel consumption, which method overcomes at least some of the disadvantages with known techniques.

The invention is defined by the appended independent claims.

Embodiments are set forth in the appended dependent claims and in the following description and drawings.

According to a first aspect, there is provided a method of positioning an air deflector for guiding an airflow around a vehicle during driving, the method comprising moving the air deflector by means of an electric actuator from a first position to a second position, while the vehicle is moving in a forward direction, recording a data series representing a signal related to the electric actuator, during said moving of the air deflector, sampling the data series by dividing it into a plurality of windows, each containing a plurality of signal values related to a range of positions, thereby forming a sampled actual data series, extracting at least one feature from each window of the sampled actual data series, thereby forming a feature actual data series, receiving a reference data set comprising at least one reference data series formed of at least one feature representing a corresponding sample window representing a plurality of positions of the air deflector, and identifying a feature or a position within a window as representing an optimum position of the air deflector, comparing the feature actual data series with the reference data series in the reference data set, determining which reference data series provides the best fit with the feature actual data series, and controlling the actuator to set the position of the air deflector to correspond to the optimum position identified in the determined reference data series to provide the best fit.

This method according to the present invention may provide an efficient and reliable positioning of an air deflector to an optimum position, without the need of complex hardware installations. By making the air deflector sweep to record the data series of the actuator signal, resulting in the feature actual data series, a reference data set may be used to find the optimum position to which the air deflector may be set. The minimum air resistance for the present vehicle setup (specific trailer/load height or width, or truck-trailer/load combination) may thereby be provided. Compared to air deflector systems using solely the electric actuator current as input for finding the optimum position, or using additional detectors and sensors, the present invention may provide a more accurate and cost-efficient method. The method of the present invention may provide a precise positioning of air deflectors to optimum positions, without the need of complex and costly hardware installations.

Making the sweep of the air deflector from the first position to the second position, the feature actual data series may be provided. The signal that is measured and recorded may for instance be the current consumed by the electric actuator. The registered signal may alternatively be the voltage of the electric actuator, rotations per minute of a rotor of a motor, motor switching frequency, position of a piston of a reciprocating engine, motor resistance, motor inductance or motor inertia. The feature actual data series, constituted by one or more extracted features, may provide a plot comprising one or more values. These values may be compared to corresponding values in the reference data series of the reference data set. The feature of the reference data series may preferably be the same feature as extracted to the feature actual data series. By recording it may be meant the signal of the electric actuator is measured and stored continuously during the recording.

The reference data set preferably comprises a plurality of reference data series. Each reference data series in the reference data set may represent a specific vehicle setup. The comparison between the feature actual data series and the reference data set may be performed to find the reference data series, and thereby the optimum air deflector position, for the present vehicle setup. The finding of the reference data series for the vehicle setup may thereby be based on the characteristics of the recorded electric actuator signal on the present vehicle.

In one embodiment, the recorded data series may be sampled into at least three windows, preferably between 3-15 windows, or more preferably between 3-10 windows.

The reference data set may in one embodiment be provided as a lookup table comprising data of the reference data series.

In one embodiment, the method may further comprise normalizing the recorded data series. By normalizing the recorded data series, a data series may be provided that comprises values at a normalized level. Thereby, the resulting feature actual data series is not affected by conditions not related to the air resistance causing an increase or decrease of the signal values. Such conditions may be an increased friction in the mechanical construction of the actuator and/or the air deflector, e.g. due to rust or certain weather conditions. When comparing the feature actual data series with the reference data set for positioning an air deflector, it may thereby not be necessary that it is the identical air deflector system used under the identical conditions as when the reference data series of the reference data set were formed. By normalizing it may be meant that the signal values are adjusted to a predetermined scale. It may be that the values are adjusted such that the mean of the signal values of the recorded range is a predetermined value, for instance zero, or that a peak value in the range is set to a predetermined value and the other signal values are adjusted correspondingly.

In another embodiment, the step of comparing may comprise comparing the values of each of the features in the feature actual data series with the values of the corresponding features in the reference data series in the reference data set. For each of the windows, the values of the features in the two data series may be compared in order to identify any differences. The differences between the two values in all window may together affect the determination in finding the reference data series for best fit. All reference data series in the reference data set may be compared to the feature actual data series before determining which may provide the best fit. All the feature value comparisons for a feature actual data series may be used in combination for determining which reference data series provides the best fit.

In one embodiment, the feature actual data series and the reference data series may comprise at least two windows, said at least two windows of the reference data series may be given individual weights, and said weights may be used in the step of determining which reference data series provides the best fit with the feature actual data series. The sampled data series comprises a plurality of windows, and for each window a feature may be extracted to form the feature actual data series. Among the plurality of windows, there may be windows more important in the step of determining which reference data series may provide the best fit. The correlation between the feature values in the feature actual data series and the reference data series in such more important windows may thereby have a larger impact on the determination of best fit reference data series than for other windows. This may further improve the accuracy in the determination of best fit reference data series and thereby the optimum position of the air deflector. The weights of the windows may be a part of the reference data series in the reference data set. Conditions when the reference data series were formed may have affected the importance and thereby the weights of the windows. Using weights as discussed above may provide a relaxation operation to be used in the determination of the best fit reference data series.

In a further embodiment, said comparing may comprise using a classifier algorithm to determine which reference data series provides the best fit with the feature actual data series. Using a classifier algorithm to determine which of the reference data series in the reference data set that provides the best fit with the feature actual data series may be a way of efficiently and accurately finding the most suitable reference data series, and thereby the most correct optimum position of the air deflector. Classification algorithms are used in machine learning and statistics for identifying to which of a set of sub-populations a new observation belongs. In this case, the reference data set may form a so called training set comprising observations in the form of the reference data series. The known optimum positions for the reference data series may then be transferred to the new observation, the feature actual data series. There are many known classifier algorithms that may be used, for instance Bayesian based classifiers (e.g. naive Bayes), boosted trees, random forest or support vector machines.

In one embodiment, the extracted feature in the sampled actual data series and the feature in the reference data series may comprise at least one of a mean value of the signal values, a standard deviation of the signal values, a maximum of the signal values, a minimum of the signal values, a mean of down sampled signal values, a standard deviation of down sampled signal values, a mean of a first derivative of the signal values, a standard deviation of a first derivative of the signal values, a mean of a second derivative of signal values, a standard deviation of a second derivative of the signal values, an oscillation of the signal values, and a peak value of the signal values. The features, at least one for each sample window, may be extracted by being calculated based on the signal values in the interval of each window. The selection of feature to be extracted may depend on predetermined conditions.

In another embodiment, the reference data series in the reference data set may each identify an interval of signal values or a plurality of signal values representing a set of possible optimum positions for the air deflector, and the reference data series may further provide criteria for determining one optimum position out of the set of possible optimum positions to be used based on a value of an additional feature for each window or a combination of windows, the method may further comprise a step of extracting an additional feature for each window of the sampled actual data series, and a step of, when a reference data series providing the best fit has been determined, determine which of the possible optimum positions in the set of possible optimum positions for the reference data series to be used based on the extracted additional feature, and the step of controlling the actuator to set the position of the air deflector may be performed with the determined optimum position based on the additional feature. After the reference data series providing the best fit has been determined, a specific optimum position for the air deflector may be determined in an additional step out of the set of possible optimum positions provided by the determined reference data series. To make this additional determination, an additional feature may be used. The additional feature may be extracted from the sample windows in the sample actual data series. The additional feature may not be used for the initial determination of finding the reference data series providing the best fit. The additional feature may be a feature different from any of the features extracted and used for the initial determination of best fit reference data series. The additional feature may be selected from the same list of possible features as described above, although other features may be used as well.

The criteria to be used for determining the optimum position to be used out of the set of possible optimum positions may be a threshold value for the additional feature, or intervals each pointing to different optimum positions. When the criteria have been used to determine which of the possible optimum positions to be used, the air deflector may be controlled to that determined optimum position.

Further, the method may comprise yet further steps of comparing a second additional extracted feature with a corresponding second additional feature of a reference data series. The determination of which optimum position to use out of a set of possible optimum positions may thereby be performed in a plurality of steps using the second additional feature, and optionally even further additional features, using different criteria for the different additional features. Each of the additional features may be a different feature from the other additional features and the originally extracted feature. The determination of optimum position may thereby be performed in two or more layers of determinations. When using such second or further additional features, the reference data series of the reference data set may comprise corresponding additional features and respective criteria for determination of optimum position.

According to a second aspect, there is provided a method of providing a data set for estimating an optimum position of an air deflector for guiding an airflow around a vehicle during driving, the method comprising moving the air deflector by means of an electric actuator from a first position to a second position, while the vehicle is moving in a forward direction, recording a data series representing a signal related to the electric actuator, during said moving the air deflector, sampling the data series by dividing it into at least one window, each containing a plurality of signal values related to a range of positions, thereby forming a sampled data series, extracting at least one feature from each window of the sampled data series, forming the reference data set to comprise a reference data series formed of said features for said windows, and identifying one of the at least one window, or a signal value in a window of the recorded data series, as representing an optimum position of the air deflector.

The method may provide data to be used for optimizing positions of air deflectors on vehicles. The reference data set may be stored and distributed to be used in the future when positioning air deflectors. Compared to air deflector systems using solely the electric actuator current as input for finding the optimum position, or using additional detectors and sensors, the present invention may provide a more accurate and cost-efficient method. An identified minimum current may occur for different reasons, not necessarily for an air deflector position providing the lowest air resistance. The method of the present invention may enable a precise positioning of air deflectors to optimum positions, without the need of complex and costly hardware installations.

When identifying the window or signal value representing the optimum position, the optimum position may be determined independently from the recorded data series of the actuator signal. Instead, during such test conditions when the method may be performed, the optimum position of the air deflector at a specific vehicle setup may be determined from e.g. visual inspection, that may be possible at test conditions, airflow simulations of which deflector position provides the minimum air resistance for a vehicle setup, or empirical data typically found in handbooks of the equipment (truck, trailer, deflector), may provide a determination of the optimum position of the air deflector providing a minimum air resistance.

Making the sweep of the air deflector from the first position to the second position, the recorded data series may be provided. The signal that is measured and recorded may for instance be the current consumed by the electric actuator. The registered signal may alternatively be the voltage of the electric actuator, rotations per minute of a rotor of a motor, motor switching frequency, position of a piston of a reciprocating engine, motor resistance, motor inductance and motor inertia. The reference data series, constituted by one or more extracted features, may provide a plot comprising one or more values to which a future data series may be compared.

In one embodiment, the recorded data series may be sampled into at least three windows, preferably between 3-15 windows, or more preferably between 3-10 windows.

In one embodiment, the method may further comprise normalizing the recorded data series. By normalizing the recorded data series, a data series may be provided that comprises values at a normalized level. Thereby, the resulting reference data series is not affected by conditions not related to the air resistance causing an increase or decrease of the signal values. Such conditions may be an increased friction in the mechanical construction of the actuator and/or the air deflector, e.g. due to rust or certain weather conditions. When using the reference data set for positioning an air deflector, it may thereby not be necessary that it is the identical air deflector system used under the identical conditions as when the reference data series of the reference data set were formed.

In one embodiment, the sampled data series may comprise at least two windows for which at least one feature may be extracted for each window, and each of the windows may be given an individual weight. When using the reference data set for finding an optimum position of an air deflector, the features in the reference data series may be differently important when comparing them to a recorded data series of the air deflector to be positioned. When giving each feature or window an individual weight, the windows’ importance may be considered, thereby enabling a more accurate determination of the optimum position.

In one embodiment, the method may further comprise repeating recording at same speed and same vehicle setup a plurality of times and updating the reference data series. By repeating the recording of data series for the same speed and vehicle setup, the resulting reference data series may be updated to provide a more accurate representation of the signal. Making several recording may further provide enhanced data in the reference data series in that it may more accurately cover the possible signal outcome for such vehicle setup. The purpose of a reference data series is to represent the basic behavior for a certain vehicle setup at a certain speed. When recording data, temporary fluctuations may occur due to for instance side winds and road bumps. By repeating the recording and updating the reference data series, the basic behavior may be easier to identify. If using weights of the sample windows, the repeated recordings may affect the weight value for a window. If the feature value of a window varies in a large extent during the repeating recordings, the weight of that window may be lowered to provide less impact when later using the reference data set for positioning an air deflector. By vehicle setup it may be meant a certain trailer/load height or width, and/or truck-trailer/load combination.

In another embodiment, the reference data set may be formed as a lookup table. The resulting recorded data set, comprising two or more reference data series, may be provided as a lookup table. Such lookup table may then later be used as input when positioning an air deflector.

In a further embodiment, the extracted feature may comprise at least one of: a mean value of the signal values, a standard deviation of the signal values, a maximum value of the signal values, a minimum value of the signal values, a mean of down sampled signal values, a standard deviation of down sampled signal values, a mean of a first derivative of the signal values, a standard deviation of a first derivative of the signal values, a mean of a second derivative of the signal values, and a standard deviation of a second derivative of the signal values. The features, at least one for each sample window, may be extracted by being calculated based on the signal values in the interval of each window. The selection of feature to be extracted may depend on predetermined conditions.

In one embodiment, the method may further comprise forming a plurality of reference data series for different vehicle setups, each reference data series comprising at least one feature for at least one window of the sampled data series and an identified window or signal value representing an optimum position for that reference data series, and the reference data set may comprise the plurality of reference data series. The characteristics of a reference series may depend on the vehicle setup for which it was determined. The reference data set comprising a plurality of reference data series may thereby comprise optimum air deflector positions for a plurality of vehicle setups. The feature values in the reference data series may be used for finding a suitable reference data series to select for use of its identified optimum position.

In one embodiment, the step of identifying an optimum position may comprise identifying an interval of signal values or a plurality of signal values representing a set of possible optimum positions for the air deflector, and the reference data series may further be provided with criteria for determining one optimum position out of the set of possible optimum positions based on a value of an additional feature for each window (a-h) or a combination of windows. The additional feature may be a feature different from the previous at least one extracted feature. The additional feature may be selected from the list of possible features as presented above, although other features may be used as well. The criteria to be used for determining the optimum position to be used out of the set of possible optimum positions may be a threshold value for the additional feature, or intervals each pointing to different optimum positions. The additional feature may be extracted for each window in the sampled data series.

In further embodiments, the criteria for each reference data series may be criteria for determining one optimum position out of the set of possible optimum positions based on the additional feature and at least one further additional feature in subsequent steps of determination. The criteria for stepwise determination may thereby provide layers of determination for using a plurality of features to in steps reach a determined optimum position. The second additional feature and other optional further additional features may each be different features from the other additional features and the original extracted feature.

According to a third aspect of the invention, there is provided an air deflector system of a vehicle for guiding an airflow around the vehicle during driving, the air deflector system comprising an air deflector which is moveably attachable to the vehicle, an electrically drivable actuator configured to adjust the position of the air deflector relative to the vehicle, a controller configured to perform any of the methods according to any of the embodiments above. The air deflector system may be configured to perform either of the method providing the reference data set and the method of positioning the air deflector, or it may be able to perform both of said methods.

Brief Description of the Drawings Embodiments of the present solution will now be described, by way of example, with reference to the accompanying schematic drawings in which: Figs. 1 a-1 c are schematic side-views of a truck-trailer combination provided with an air deflector located on the roof of the truck cab.

Fig. 2a is a schematic side-view of a truck cab with an air deflector arranged on the truck cab roof.

Fig. 2b is a schematic top-view of a truck cab with a roof-mounted air deflector and air deflectors arranged on both lateral sides of the truck cab.

Fig. 3 illustrates a flowchart of a method according to an embodiment of the present invention.

Fig. 4 illustrates a current-actuator extension diagram of examples of recorded data.

Fig. 5 illustrates a current-actuator extension diagram with a sampled data series.

Fig. 6 illustrates diagrams of reference data series according to an embodiment of the present invention.

Fig. 7 illustrates a flowchart of a method according to an embodiment of the present invention.

Fig. 8 illustrates a diagram of a feature actual data series according to an embodiment of the present invention.

Fig. 9 illustrates a diagram of a reference data series including weights according to an embodiment of the present invention.

Description of Embodiments The present invention will be described more fully hereinafter with reference to the accompanying drawings, in which preferred embodiments of the invention are shown. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. In the drawings, like numbers refer to like elements.

For an automotive ground vehicle such as a truck 1 carrying a load or towing a trailer 2, see Figs. 1a-1c, the front surface of extended parts of the load or trailer 2 will cause significant air drag during journey. Such air drag also occurs for example when a car tows a caravan.

For reducing the air resistance of the vehicle and to reduce fuel consumption the truck cab 1 may be provided with a roof-mounted air deflector 3 (see Figs 1a-1c and Fig. 2a) and/or side air deflectors 4 mounted on each lateral side of the truck cab 1 (Fig. 2b) for hiding the frontal face area of the trailer/load 2 from an oncoming airflow and to get a more streamlined complete vehicle.

Factors that may affect the air resistance are mainly wind direction and wind speed in relation to the vehicle and geometrical factors, namely height and width of the driver's cab 1 relative to the height and width of the trailer 2/load as well as the distance between truck 1 and trailer 2.

The air deflector 3, 4, comprises a major essentially planar airdeflecting surface 20, 20', which has a first edge 10, 10' and a second edge 11, 11' essentially opposite the first edge 10, 10' across the major airdeflecting surface 20, 20'. When arranged on the truck cab 1, see Figs 1a-1c and Figs 2a and 2b, the air deflector 3, 4 is arranged to be pivotable around its first edge 10, 10' to form an angle of inclination ?, ?' of the air-deflecting surface 10, 10' to a direction X parallel with the normal forward direction of the moving vehicle. When the angle of inclination ?, ?' is changed the second edge 11, 11' is correspondingly translated from one position to another position.

The air deflector 3, 4 may have a first extreme position in which the angle of inclination ?, ?' is low, e.g. about 0°-15°, and a second extreme position in which the angle of inclination ?, ?' is high, e.g. about 20°-60°.

For a side-mounted air deflector 4, see Fig. 2b, the second edge 11' is pivoted around the first edge 10' forwards and inwards relative to the side of the vehicle forming an angle ?' relative to the longitudinal axis X.

For a roof-mounted air deflector 3, see Fig. 2a, the second edge 11 is pivoted around the first edge 10 upwards and downwards relative to the roof of the truck cab forming an angle ? relative to the longitudinal axis X.

The air deflector 3, 4 may be mounted by means of a frame (not shown) to the truck cab 1 and the air deflector 3, 4 may be moveable with respect to the frame.

The vehicle, when travelling, will be subject to load from the air, where the load from the air depends on the angle of inclination ?, ?' of the airdeflecting surface 20, 20' and on the condition of the air deflector 3, 4 relative to the load/trailer. Load due to air that flows against the air-deflecting surface 20, 20' of the air deflector 3, 4 increases as the angle relative to the direction X increases.

In Fig. 1a the roof-mounted air deflector 3 is adjusted to a condition such that the second end 11 of the roof-mounted air deflector 3 is positioned significantly above the height of the trailer 2. The air flowing 5 off the second edge 11 of the air deflector 3 may create a separate flow in an opposite direction in a cavity between the air deflector 3 and the truck cab 1, and thus increase the pressure in the cavity and the air drag on the vehicle as compared to an optimally positioned air deflector (see Fig. 1c).

In Fig. 1b the second edge 11 of the air deflector 3 is instead in a position lower than the height of the forward end of the trailer. This part of the trailer will produce a stagnation of a part of the flow 5. The stagnation of the flow may force an increased air quantity into the cavity and thus create an increased pressure in the cavity and the air drag on the vehicle as compared to an optimally positioned air deflector (see Fig. 1c).

In Fig. 1c the second end 11 of the air deflector 3 is positioned at the same or approximately the same level as the height of the trailer 2. The air then may flow smoothly over the air guide and further above the roof of the trailer 2. In this optimal position the airflow 5 is guided around the vehicle during driving, i.e. the air deflector 3 redirects the air which is moving over the moving vehicle in such a way that the air does not hit the front edge of the trailer 2 but instead the air is moving over the roof of the trailer 3.

The same effects are seen for a truck carrying a load having a height extending over the height of the truck cab. If the load or towed trailer 2 extends laterally outwards of the truck cab 1 the optimal position of the air deflectors is when the second end 11' of the air deflector 4 is positioned at the same or approximately the same level as the width of the trailer 2 or load.

Even small angular variations in the settings of the deflectors 3, 4 may have significant effect on the total vehicle air resistance.

The dimensions of trailers and of load carried by a vehicle often vary, why for optimal reduction of the air drag, the air deflector(s) 3, 4 should be adjustable.

Traditionally, air deflectors 3, 4 have been manually adjustable. It is, however, desirable to, automatically, during driving of the vehicle, be able to identify and position the air deflector 3, 4 for any truck-trailer or truck-load combination in an optimum position or close to optimum position, in which position the air drag is the lowest.

Fig. 2a further illustrates a controller 30 and an actuator 31, which is in communication with the controller 30.

The actuator 31 may be any type of actuator that is capable of setting the position/orientation of the air deflector 3, 4, such as a linearly operating actuator or an angularly operating actuator.

For example, the actuator 31 may be an electrically driven actuator, and in particular one that is drivable by direct current, as that is what is in many cases most readily available in vehicles.

The controller 30 may comprise a processing unit and a memory.

Moreover, the controller 30 may be adapted to measure the current and/or voltage fed to the actuator 31.

For example, where the motor is a constant speed motor (e.g. a DC shunt motor), it may be sufficient to measure the current, provided that the voltage can be assumed to be constant.

To set the air deflector 3, 4 to the optimum position, a method is performed to find the optimum position and to position the air deflector 3, 4 at said position.

The method may be automatically run, for example when a vehicle after a certain time period of stand-still starts again, when the load or trailer 2 has been changed and/or when the vehicle reaches a certain speed (such as for example 20km/h). Alternatively, a driver of a truck could start the optimizing method manually by pushing a button.

The air deflector 3, 4 is moved, “swept”, from a first p1, p1' to a second position p2, p2' when the vehicle is moving in a forward direction 100. Here, the first position is the position with the smallest angle of inclination ?, ?' of the air-deflecting surface 20, 20'of the air deflector 3, 4 to a direction X parallel with the normal forward direction and the second position a position of largest angle of inclination ?, ?'.

It is to be understood that the first position equally well may be the position with the largest angle of inclination and the second position the position with the smallest angle of inclination and the air deflector 3, 4 being swept from the point with the highest angle of inclination to the point with the lowest angle of inclination.

The air deflector 3, 4 may be moved from the first p1, p1' to the second position p2, p2' by means of an electric actuator 30. The electric actuator 30 may comprise a reciprocating engine, an electro-mechanical actuator, an alternating current motor selected from a group comprising an induction motor, a synchronous motor, a sliding rotor motor and a repulsion motor, or a direct current motor selected from a group comprising a brushed motor and a brushless motor.

The controller 30 may be arranged for sensing a parameter of the electric actuator 31 for moving the deflector 3, 4 from the first to the second position and for registering a signal from the electric actuator when the air deflector is moved from the first to the second position. The registered signal represents the sensed parameter that may be current of the electric actuator, voltage of the electric actuator, rotations per minute of a rotor of a motor, motor switching frequency, position of a piston of a reciprocating engine, motor resistance, motor inductance and motor inertia.

Before starting the method, it is possible to make a pre-sweep with the air deflector 3, 4 between the two outermost positions of the air deflector identifying a portion of the positions in which it is likely that the most interesting information in the signal corresponding to the sensed parameter is to be found. The first and second positions p1, p2 are then set as the end points of this portion.

The first position p1 may be obtained by moving the air deflector 3, 4 in a direction from a first outermost position towards a second outermost position an angle of 2-5% of a total angle between the first and second outermost positions of the air deflector 3, 4 and the second position may be obtained by moving the air deflector 3, 4 from the first outermost position towards the second outermost position an angle of 95-98% of the total angle between the first and second outermost positions.

The angle between the outermost positions may be 0-80°.

Alternatively, the first position p1 may be obtained by moving the air deflector 3, 4 in a direction from a first outermost position towards a second outermost position for a time period of 2-5% of a total time for moving the air deflector between the first and second outermost positions, and the second position p2 may be obtained by moving the air deflector 3, 4 from the first outermost position towards the second outermost position for a time period of 95-98% of the total time for moving the air deflector 3, 4 between the first and second outermost positions.

The total time for moving the air deflector 3, 4 between the outermost positions may be 5-300 seconds.

Fig. 3 illustrates a flowchart of a method 100 of providing a data set for determining an optimum position of an air deflector 3, 4. The data set is provided to be used as a reference data set when later controlling the position of an air deflector. The method 100 comprises a first step of moving 102 the air deflector 3, 4 from a first position p1 to a second position p2. The air deflector 3, 4 is moved by means of the electric actuator 31. The air deflector 3, 4 is moved between these positions during movement of the vehicle in the forward direction X. During the movement 102 of the air deflector 3, 4, a data series is recorded 104 which represents a signal related to the electric actuator 31. The signal may for instance be the electric current used by the electric actuator 31 during the movement of the air deflector 3, 4. The data series is thereby a signal function of current over time or current over actuator extension. By actuator extension it is meant a position value for the actuator, and thereby indirectly also for the air deflector. The data series provides current values for each actuator extension value, or alternatively for each time instant or air deflector position, during the recording period.

Fig. 4 illustrates a diagram of current over actuator extension for an exemplified recorded data series. In the illustrated embodiment, the current values of the recorded data series are normalized in order to facilitate the use of the recorded values in the method. The recorded values providing the shape of the curve is important for the method rather than the position of the curve on the current y axis. The signal values being normalized facilitates the further use of the recorded data series. In the fig. 4 examples, the electric actuator 31 is moved from a 0 mm extension to a 400 mm extension, representing the first p1 and second p2 air deflector positions.

In one embodiment, the recorded signal values for a first portion of the actuator extension, and a last portion of the actuator extension may be excluded from the recording. The reason is that the signal values close to the end positions may not correctly represent the air resistance, but may to a large extent be affected of other properties. For such signals not to negatively affect the further calculations, they are excluded. In one embodiment, the signal values for the first 10 mm and the last 10 mm of the actuator 31 extension are excluded.

In the next step of the method 100, the recorded data series is sampled 106 by dividing it into at least one window, at least three windows, preferably between 3-15 windows, or more preferably between 3-10 windows. Each window contains a plurality of signal values related to a range of positions. This forms a sampled data series. Fig. 5 illustrates a diagram with sample windows a-h of the data series in fig. 4. The sample windows a-h each comprises an interval of signal values of the recorded data series.

Further, the method 100 comprises a step of extracting 108 a feature from each sample window a-h. The feature is selected from the group of a mean value of the signal values, a standard deviation of the signal values, a maximum value of the signal values, a minimum value of the signal values, a mean of down sampled signal values, a standard deviation of down sampled signal values, a mean of a first derivative of the signal values, a standard deviation of a first derivative of the signal values, a mean of a second derivative of the signal values, a standard deviation of a second derivative of the signal values, an oscillation of the signal values, and a peak value of the signal values. Other features calculated from the signal values may also be possible. In the illustrated example of fig. 5, each of the eight sample windows a-h would thereby provide a single value feature.

Fig. 6 illustrates diagrams of extracted features from the eight sample windows a-h. The extracted features are illustrated in the form of standard deviations (Std) of the recorded signal values in each sample window.

In the next step of the method 100, the extracted features from a set of sample windows a-h together forms a reference data series. In fig. 6, seven reference data series are illustrated provided from seven separate data series recordings.

In the final step of the method 100, an optimum position is identified for the reference data series. The optimum position is identified as the one of the sample windows a-h which represents the position of the air deflector 3, 4 as the optimum position, or as a position value (x-axis value) within one of the sample windows which position value provides the optimum position of the air deflector 3, 4. Fig. 6 illustrates, for each of the seven reference data series, an optimum position window (a, b, c, d, e, f, g) indicating the optimum position of the air deflector 3, 4. The actual optimum position may be determined based on, e.g. airflow simulations of which deflector position provides the minimum air resistance for a vehicle setup, or empirical data, which is typically found in handbooks of the equipment (truck, trailer, deflector), and can be expected to provide information on which deflector position is optimal for a certain trailer height and/or width for a certain truck. Instead of providing a window a-h which represent the optimum position, the reference data series 151-157 may comprise specific signal values to be used as optimum position.

From the method 100, a resulting reference data set 150 comprising a plurality of reference data series 151-157 is provided. For each reference data series 151-157 an optimum position of the air deflector 3, 4 is identified.

Such reference data set 150 may then be used for positioning an air deflector 3, 4 at an optimum position using a method 200 as illustrated in fig. 7. The method 200 for setting the air deflector 3, 4 to an optimum position for guiding an airflow around a vehicle during driving comprises a first step of moving 202 the air deflector 3, 4 from a first position p1 to a second position p2. The air deflector 3, 4 is moved by means of the electric actuator 31. The air deflector 3, 4 is moved between these positions during movement of the vehicle in the forward direction X. During the movement 202 of the air deflector 3, 4, a data series is recorded 204 which represents a signal related to the electric actuator 31. The signal may for instance be the electric current used by the electric actuator 31 during the movement of the air deflector 3, 4. The data series may thereby be a signal function of current over actuator extension. The data series provides current values for each actuator extension value, or alternatively for each time instant or air deflector position, during the recording period. Such recorded data series may be provided in the same way as the current curve illustrated in fig. 4 and discussed above. Similarly, as discussed above for the recorded data series for a reference data series, the signal values may be normalized, and the first and last portions of the signal values over the actuator extension may be excluded.

In the next step of the method 200, the recorded data series is sampled 206 by dividing it into at least one window. Each window contains a plurality of signal values related to a range of positions. This forms a sampled data series. Such sampled actual data series may be provided in the same way as the sampled data series illustrated in fig. 5 illustrating the diagram with sample windows a-h of the data series in fig. 4. The corresponding sample windows in the sampled actual data series may in this exemplified embodiment be denoted i-viii. The sample windows i-viii each comprises an interval of signal values of the recorded data series.

Further, the method 200 comprises a step of extracting 208 a feature from each sample window i-viii. The feature is selected from the group of a mean value of the signal values, a standard deviation of the signal values, a maximum value of the signal values, a minimum value of the signal values, a mean of down sampled signal values, a standard deviation of down sampled signal values, a mean of a first derivative of the signal values, a standard deviation of a first derivative of the signal values, a mean of a second derivative of the signal values, and a standard deviation of a second derivative of the signal values. The extracted features for the sample windows i-viii together form a feature actual data series 250 as illustrated in fig. 8. Correspondingly as the illustrated example of fig. 5 for the reference data series, each of the eight sample windows i-viii for the feature actual data series provide a single value feature.

Next, the method 200 comprises a step of receiving 210 a reference data set 150 comprising at least one reference data series 151-157 formed of at least one feature representing a corresponding sample window a-h representing a plurality of positions of the air deflector, and identifying a window, or a position within a window as representing an optimum position of the air deflector. The illustrated reference data set 150 in fig. 6 provides for each reference data series 151-157 a window a-g identified as the optimum position. The reference data set 150 may be received as a look up table. The reference data set 150 may be a reference data set determined using the method 100 as discussed above. The feature used for forming the reference data series 151 -157 in the reference data set 150 is preferably the same feature as extracted for each sample window i-viii in the sampled actual data series and forming the feature actual data series 250.

Further in the method 200, the feature actual data series 250 comprising a set of feature values for the sample windows i-viii therein is compared 212 to at least some of the reference data series 151 -157 in the reference data set 150. I.e. the values of each instant in the feature actual data series 250 in fig. 8 is compared to the corresponding values in each reference data series. The feature value of instant i in the feature actual data series 250 is compared to the value of instant a in a reference data series 151-157. Correspondingly, the values in instants ii-viii are compared to the respective values of instants b-h in the reference data series 151-157. For each complete set of feature values, i.e. a reference data series 151-157, the amount of deviations between the values of the feature actual data series 250 and the values of the reference data series 151-157 are determined.

Following such determination, the method 200 comprises a step of determining 214 which reference data series 151-157 that provides the best fit with the feature actual data series 250. I.e. which reference data series 151-157 which provides the least deviations from the values of the feature actual data series 250 or provides deviations which according to predetermined conditions will provide the best fit.

The reference data set 150 may comprise a large number of reference data series. There may be hundreds or thousands of reference data series providing optimum positions for a large number of vehicle setups. The determination 214 of which reference data series that provides the best fit may in such situation be a complex operation. To effectively perform such operation, a classification algorithm can be used. Classification algorithms are used in machine learning and statistics for identifying to which of a set of subpopulations a new observation belongs. In this case, the reference data set forms a so called training set comprising observations in the form of the reference data series. The known optimum positions for the reference data series may then be transferred to the new data series, the feature actual data series. There are many known classification algorithms known that can be used. Such techniques may be Bayesian based classifiers (e.g. naive Bayes), boosted trees, random forest or support vector machines.

The next step of the method 200 provides a positioning 216 of the air deflector 3, 4 to a position in the window a-g indicated in the reference data series 151-157 determined to be the best fit as the optimum position.

In the illustrated embodiment, steps 202-208 of the method 200 provides the feature actual data series 250 as in fig. 8. In step 210 the reference data set 150 as illustrated in fig. 6 is received. The feature actual data series 250 is thereby compared to each of the seven reference data series 151-157 in the reference data set 150. It would then be found that reference data series 154 provides the best fit since it is found that there are some deviations in values at instants a, c and h, but none or very small deviations at instants b, d, e, f and g when compared to the values at instants i-viii in the feature actual data series 250. When reference data series 154 has been determined to provide the best fit, the optimum position value of window d provided by the reference data series 154 is considered, and the air deflector 3, 4 is positioned to a position corresponding to an extension value in window d of the electric actuator 31. When the reference data series 154 has been determined as providing the best fit, the air deflector 3, 4 is positioned to a position within sample window iv corresponding to window d of the reference data series 154. In the sample window iv comprising an interval of positions, a position may be selected according to a predetermined configuration, for instance a center value in the interval. Alternatively, the reference data series 154 may provide a specific signal value or actuator extension value as optimum position, and the deflector may be positioned to a position corresponding to that extension value of the electric actuator 31.

In one embodiment, the sample windows a-h are given different weights. Fig. 9 illustrates the reference data series 154 in which each sample window a-h is given a weight value between 0-1. When comparing the feature actual data series 250 with the reference data series 154 and determining whether it provides the best fit, the weights of the sample windows a-h, and thereby also of the feature values therein, are used. The sample windows a-h may be differently important when determining the best fit reference data series that would provide the most correct optimum position of the air deflector. Fig. 9 illustrates the reference data series 154 as being provided with sample window weights. However, the weights may as well be provided in the feature actual data series i-viii, or to both.

In the drawings and specification, there have been disclosed preferred embodiments and examples of the invention and, although specific terms are employed, they are used in a generic and descriptive sense only and not for the purpose of limitation, the scope of the invention being set forth in the following claims.

Claims (17)

1. A method (200) of positioning an air deflector (3, 4) for guiding an airflow (5) around a vehicle (1, 2) during driving, the method comprising: a) moving (202) the air deflector by means of an electric actuator (31) from a first position (p1) to a second position (p2), while the vehicle is moving in a forward direction (X), b) recording (204) a data series representing a signal related to the electric actuator, during said moving of the air deflector, c) sampling (206) the data series by dividing it into a plurality of windows (i-viii), each containing a plurality of signal values related to a range of air deflector positions, thereby forming a sampled actual data series, d) extracting (208) at least one feature from each window of the sampled actual data series, thereby forming a feature actual data series (250), e) receiving (210) a reference data set (150) comprising at least one reference data series (151-157) formed of at least one feature representing a sample window (a-h) representing a plurality of positions of the air deflector (3, 4), and identifying a feature or a position within such sample window (a-h) as representing an optimum position of the air deflector, f) comparing (212) the feature actual data series with the reference data series in the reference data set, g) determining (214) which reference data series provides the best fit with the feature actual data series, and h) controlling (216) the electric actuator to set the position of the air deflector to correspond to the optimum position identified in the determined reference data series to provide the best fit.

2. The method as claimed in claim 1, further comprising normalizing the recorded data series.

3. The method as claims in any one of the claims 1 -2, wherein the step of comparing (212) comprises comparing the values of each of the features in the feature actual data series (250) with the values of the corresponding features in the reference data series (151-157) in the reference data set (150).

4. The method as claimed in any one of the claims 1-3, wherein the feature actual data series (250) and the reference data series (151-157) comprises at least two windows (i-viii, a-h), and wherein the at least two windows of the reference data series are given individual weights, and wherein said weights are used in the step of determining which reference data series provides the best fit with the feature actual data series.

5. The method as claimed in any of claims 1-4, wherein said comparing (212) comprises using a classifier algorithm to determine which reference data series (151-157) provides the best fit with the feature actual data series (250).

6. The method as claimed in any of claims 1 -5, wherein the reference data set (150) is received in the form of a look up table.

7. The method as claims in any of the claims 1 -6, wherein the extracted feature in the sampled actual data series and the feature in the reference data series comprises at least one of: a mean value of the signal values, a standard deviation of the signal values, a maximum of the signal values, a minimum of the signal values, a mean of down sampled signal values, a standard deviation of down sampled signal values, a mean of a first derivative of the signal values, a standard deviation of a first derivative of the signal values, a mean of a second derivative of signal values, a standard deviation of a second derivative of the signal values an oscillation of the signal values, and a peak value of the signal values.

8. The method as claimed in any of the claims 1-7, wherein the reference data series (151-157) in the reference data set (150) each identifies an interval of signal values or a plurality of signal values representing a set of possible optimum positions for the air deflector (3, 4), and wherein the reference data series further provide criteria for determining one optimum position out of the set of possible optimum positions to be used based on a value of an additional feature for each window or a combination of windows, wherein the method further comprises a step of extracting an additional feature for each window of the sampled actual data series, and a step of, when a reference data series providing the best fit has been determined, determine which of the possible optimum positions in the set of possible optimum positions for the reference data series to be used based on the extracted additional feature, and wherein the step of setting the position of the air deflector is performed with the determined optimum position based on the additional feature.

9. A method (100) of providing a reference data set for determination of an optimum position of an air deflector (3, 4) for guiding an airflow (5) around a vehicle (1, 2) during driving, the method comprising: a) moving (102) the air deflector by means of an electric actuator (31) from a first position (p1) to a second position (p2), while the vehicle is moving in a forward direction (X), b) recording (104) a data series representing a signal related to the electric actuator, during said moving the air deflector, c) sampling (106) the data series by dividing it into at least one window (a-h), each containing a plurality of signal values related to a range of positions, thereby forming a sampled data series, d) extracting (108) at least one feature from each window of the sampled data series, e) forming (110) the reference data set (150) to comprise a reference data series (151-157) formed of said features for said windows, and f) identifying (112) one of the at least one window, or a signal value in a window of the recorded data series, as representing an optimum position of the air deflector.

10. The method as claimed in claim 9, further comprising normalizing the recorded data series.

11. The method as claimed in any of the claims 9-10, wherein the sampled data series comprises at least two windows for which at least one feature is extracted for each window, and wherein each of the windows are given an individual weight.

12. The method as claimed in any of the claims 9-11, further comprising repeating recording (104) at same speed and same vehicle setup a plurality of times and updating the reference data series (151-157).

13. The method as claimed in any one of claims 9-12, wherein the reference data set (150) is formed as a lookup table.

14. The method as claimed in any one of the claims 9-13, wherein the extracted feature comprises at least one of: a mean value of the signal values, a standard deviation of the signal values, a maximum value of the signal values, a minimum value of the signal values, a mean of down sampled signal values, a standard deviation of down sampled signal values, a mean of a first derivative of the signal values, a standard deviation of a first derivative of the signal values, a mean of a second derivative of the signal values, a standard deviation of a second derivative of the signal values an oscillation of the signal values, and a peak value of the signal values.

15. The method as claimed in any one of the claims 9-14, further comprising forming a plurality of reference data series (151-157) for different vehicle setups, each reference data series comprising at least one feature for at least one window of the sampled data series and an identified window or signal value representing an optimum position for that reference data series, and wherein the reference data set (150) comprises the plurality of reference data series.

16. The method as claimed in any one of the claims 9-15, wherein the step of identifying (112) an optimum position comprises identifying an interval of signal values or a plurality of signal values representing a set of possible optimum positions for the air deflector (3, 4), and wherein the reference data series (151 -157) further is provided with criteria for determining one optimum position out of the set of possible optimum positions based on a value of an additional feature for each window (a-h) or a combination of windows.

17. An air deflector system (300) of a vehicle (1, 2) for guiding an airflow (5) around the vehicle during driving, the air deflector system comprising an air deflector (3, 4) which is moveably attachable to the vehicle, an electrically drivable actuator (31) configured to adjust the position of the air deflector relative to the vehicle, a controller (30) configured to perform any of the methods according to any of the claims 1-16.