Method and system for computing a road friction estimate

Abstract

A method is described for determining a friction estimate between a road surface and a tire of a vehicle, the tire being arranged on a steerable wheel of the vehicle, and the vehicle having an axle rack pivotably attached to a linkage arm connected to the steerable wheel such that a translational motion of the axle rack causes the linkage arm to rotate about a kingpin element such that the linkage arm causes a turning motion of the steerable wheel. The method includes acquiring a plurality of rack force values; acquiring a plurality of lateral wheel force values; mapping a relation between the plurality of rack force values and the lateral wheel force values to a model; and determining the lateral friction estimate based on the mapping.

Claims

1. A method for determining a friction estimate between a road surface and a tire of a vehicle, the tire being arranged on a steerable wheel of the vehicle, and the vehicle comprising an axle rack pivotably attached to a linkage arm connected to the steerable wheel such that a translational motion of the axle rack causes the linkage arm to rotate about a kingpin element such that the linkage arm causes a turning motion of the steerable wheel, the method comprising: acquiring a plurality of rack force values indicative of rack forces on the axle rack; acquiring a plurality of lateral wheel force values indicative of lateral forces on the steerable wheel when the plurality of rack force values is acquired, the lateral wheel forces acting in a direction substantially parallel with a rotation axis of the wheel; mapping a relation between the plurality of rack force values and the lateral wheel force values to a model comprising a friction parameter, determining a lateral friction estimate based on the mapping; and providing the lateral friction estimate to an active safety system of the vehicle for use in a determination related to braking or steering of the vehicle; wherein the model comprises a plurality of predetermined relations between rack force and lateral wheel force, wherein mapping comprises matching the relation between the acquired plurality of rack force values and the acquired plurality of lateral wheel force values to one of the plurality of predetermined relations, and wherein the friction estimate is determined based on the matching predetermined relation.

2. The method according to claim 1 wherein the model is a linear relation between the lateral wheel force and the rack force, the linear relation based on predetermined data relating lateral wheel force and rack force to the friction parameter, wherein the mapping comprises: performing a linear regression to fit the acquired plurality of rack force values and the acquired plurality of lateral wheel force values to the linear relation, wherein the slope of the linear fit between the lateral wheel force values and the rack force values is indicative of the friction estimate.

3. The method according to claim 2 further comprising: determining a plurality of ratios between the rack force values and the respective lateral wheel force values, wherein the slope of the linear fit between the lateral wheel force values and the rack force values is based on the plurality of ratios.

4. The method according to claim 1 wherein the model is based on empirical data of lateral wheel force, rack force and friction between a tire of a vehicle and the road surface.

5. The method according to claim 1 wherein the lateral friction estimate is classified into different levels of nominal friction depending on the value of the friction estimate, wherein the classification levels depend on at least one of: a vehicle velocity, a vehicle velocity variation during a batch time duration, a level of wheel force excitation, a distance to interval threshold.

6. The method according to claim 1 wherein the model is a non-linear relation between the lateral wheel force and the rack force, wherein the mapping comprises: fitting the plurality of lateral wheel force values and rack force values with a non-linear model comprising the friction parameter, whereby the friction estimate is determined based on the fitting.

7. The method according to claim 6 wherein the model is tan h function relating rack force to wheel forces.

8. The method according to claim 1 wherein the model is further based on linkage arm length of the vehicle and tire pneumatic trail.

9. The method according to claim 1 wherein the lateral wheel force values and the rack force values are accumulated for a batch time duration, the method further comprising: determining a friction estimate confidence value; wherein, when the confidence value is larger than a threshold confidence, accepting the friction estimate.

10. A system for determining a friction estimate between a road surface and a tire of a vehicle, the tire for arrangement on a steerable wheel of the vehicle, and the vehicle comprising an axle rack pivotably attached to a linkage arm connected to the steerable wheel such that a translational motion of the axle rack causes the linkage arm to rotate about a kingpin element such that the linkage arm causes a turning motion of the steerable wheel, the system comprising: a force determining unit configured to determine a rack force value indicative of a rack force on the axle rack; a vehicle state estimator configured to determine a lateral wheel force value indicative of a lateral force on the steerable wheel, the lateral wheel force acting in a direction substantially parallel with a rotation axis of the wheel; and a control unit configured to: acquire a plurality of rack force values indicative of rack forces on the axle rack; acquire a plurality of lateral wheel force values indicative of lateral forces on the steerable wheel; map a relation between a plurality of rack force values and a plurality of lateral wheel force values to a model comprising a friction parameter; determine a lateral friction estimate based on the mapping; and provide the lateral friction estimate to an active safety system of the vehicle for use in a determination related to braking or steering of the vehicle; wherein the model comprises a plurality of predetermined relations between rack force and lateral wheel force, wherein mapping comprises matching the relation between the acquired plurality of rack force values and the acquired plurality of lateral wheel force values to one of the plurality of predetermined relations, and wherein the friction estimate is determined based on the matching predetermined relation.

11. The system according to claim 10 wherein the vehicle state estimator comprises an inertial measurement unit and a wheel speed sensor.

12. The system according to claim 10 further comprising a global positioning system configured to determine a velocity of the vehicle.

13. A vehicle comprising the system according to claim 10.

14. A method for determining a friction estimate between a road surface and a tire of a vehicle, the tire being arranged on a steerable wheel of the vehicle, and the vehicle comprising an axle rack pivotably attached to a linkage arm connected to the steerable wheel such that a translational motion of the axle rack causes the linkage arm to rotate about a kingpin element such that the linkage arm causes a turning motion of the steerable wheel, the method comprising: acquiring a plurality of rack force values indicative of rack forces on the axle rack; acquiring a plurality of lateral wheel force values indicative of lateral forces on the steerable wheel when the plurality of rack force values is acquired, the lateral wheel forces acting in a direction substantially parallel with a rotation axis of the wheel; mapping a relation between the plurality of rack force values and the lateral wheel force values to a model comprising a friction parameter; determining a lateral friction estimate based on the mapping; and providing the lateral friction estimate to an active safety system of the vehicle for use in a determination related to braking or steering of the vehicle; wherein the lateral friction estimate is classified into different levels of nominal friction depending on the value of the friction estimate, and wherein the classification levels depend on a vehicle velocity, a vehicle velocity variation during a batch time duration, a level of wheel force excitation, or a distance to interval threshold.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) These and other features of the present disclosure will now be described in more detail, with reference to the attached drawings showing example embodiments of the disclosure.

(2) FIGS. 1A-C schematically illustrate parameters related to friction estimation;

(3) FIG. 2A conceptually illustrates a vehicle according to example embodiments of the disclosure;

(4) FIG. 2B conceptually illustrates a chassis wheel base with wheels of the vehicle in FIG. 2A;

(5) FIG. 3 schematically illustrates a system according to an exemplary embodiment of the disclosure;

(6) FIG. 4 is a flow-chart of method steps according to an exemplary embodiment of the disclosure;

(7) FIGS. 5A-B each conceptually shows lateral wheel force values versus axle rack force and a fitted curve based on a linear model;

(8) FIGS. 6A-B each conceptually shows lateral wheel force values versus axle rack force and a fitted curve based on a non-linear model; and

(9) FIG. 6C conceptually illustrates predetermined relations between lateral wheel force and axle rack force under known friction conditions.

DETAILED DESCRIPTION

(10) As required, detailed embodiments are disclosed herein. However, it is to be understood that the disclosed embodiments are merely exemplary and that various and alternative forms may be employed. The figures are not necessarily to scale. Some features may be exaggerated or minimized to show details of particular components. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as a representative basis for teaching one skilled in the art.

(11) In the present detailed description, various embodiments of the system and method according to the present disclosure are described. However, this disclosure may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided for thoroughness and completeness, and fully convey the scope of the disclosure to the skilled person. Like reference characters refer to like elements throughout.

(12) FIG. 2A illustrates a vehicle 100 according to an embodiment of the disclosure. The vehicle 100 comprises at least one steerable wheel 102a-b. The steerable wheels 102a-b are here shown as the front wheels of the vehicle 100. The vehicle 100 further comprises rear wheels 102c-d. The wheels 102a-d each has a tire 101a-d mounted on a corresponding wheel hub 103 (not all are numbered). Furthermore, and with reference also to FIG. 2B, the vehicle 100 comprises an axle rack 104 pivotably attached to a linkage arm 106 which is connected to a steerable wheel 102a-b such that when the axle rack 106 is subjected to a force which causes a translational motion of the axle rack 104, the linkage arm 106 rotates about a kingpin 108 and thereby causes a turning motion of the wheel 102a-b. In this way, the vehicle is subject to a turning torque (T.sub.whlTurn). The vehicle 100 may for example be an autonomous car.

(13) When applying a torque to a single front wheel 102a, the resulting force (F.sub.Whl) between the wheel and the contact area (e.g., the road) generates a torque, T.sub.WhlStr, that attempts to turn the front wheel 102a around the kingpin 108. The turning torque is in turn transferred to the axle rack 104 and can be monitored as the rack force, F.sub.rack. In more detail and with reference to FIGS. 1B-C which illustrates a turning wheel (102a) at different friction conditions. The distance (x.sub.h and x.sub.1) between the resultant lateral F.sub.y is known as the pneumatic trail. The pneumatic trail (x.sub.h and x.sub.1) (together with the caster trail, here included in the pneumatic trail definition x.sub.h and x.sub.1) effectively forms a lever for the resultant lateral force F.sub.y. Thus, with the pneumatic trail (x.sub.h and x.sub.1) as the lever, the lateral force F.sub.y generates a torque on the wheel 102a, this torque is denoted T.sub.WhlStr. This torque translates via the linkage arm 106 having a length d, to the axle rack 104 where the axle rack force F.sub.rack can be measured. For torque equilibrium, the axle rack force F.sub.rack multiplied with the length d of the linkage arm should equal the torque T.sub.WhlStr. In other words, pneumatic trail*F.sub.y=d*F.sub.rack.

(14) In a low friction case (FIG. 1C), the assumed parabolic force distribution over the contact area between the tire and the road surface in the lateral direction (y) along x (.sub.highf(x)) has a lower maximum magnitude illustrated by the parabolic curve being closer to the tire 102a. Thus, in order to obtain the same resultant lateral force F.sub.y the wheel has to turn further (so that the lateral force distribution F.sub.y(x) includes more of the assumed maximum parabolic force distribution .sub.lowf(x)) thereby increasing the slip angle . This further moves the resultant lateral force F.sub.y closer to the center of the contact patch between the tire and the road, thus the pneumatic trail x.sub.1 is shorter compare to in the high friction case (x.sub.h in FIG. 1B). Since the length d of the linkage arm 106 is the same in the high friction case and in the low friction case but the pneumatic trail varies, a difference in axle rack force will thereby reflect a difference in friction between the tire and road surface.

(15) A wheel 102a subject to propulsion or braking (in FIG. 2B it is braking) is affected by lateral forces of different magnitude depending on the friction between the road surface and the tire of the wheel. The lateral force is generally a function (F.sub.y(T.sub.WhlStr, , .sub.slip)) of the wheel torque (T.sub.WhlStr), the coefficient of friction (), and the slip angle (.sub.slip), see FIGS. 1A-C. The lateral force F.sub.y is measurable via the rack force Frank as described with reference to FIGS. 2A-B and FIGS. 1A-C. Furthermore, the slip angle (.sub.slip) generally depends on the velocities of the vehicle (lateral velocity, longitudinal velocity, and yaw velocity) and the steering angle (see FIG. 1A) of the vehicle. The steering angle is the angle between a longitudinal centerline 130 of the vehicle (between rear and front of the vehicle) and the wheel hub direction W.sub.d (see FIG. 1A). The steering angle may be measured by a steering angle sensor arranged on the steering column (not shown) of the vehicle. The sensor may measure the rotational angle of the steering wheel and thereby derive the steering angle. Another way to measure the steering angle is to determine the position (translational position) of the axle rack and therefrom derive the steering angle. Accordingly, the slip angle can generally be provided as a function ((velocity, steering angle)) of the steering angle and the velocities. From the two functions F.sub.y(T.sub.WhlStr, .sub.slip) and (velocity, steering angle) the friction may be derived. For details regarding the functions (F.sub.y(T.sub.WhlStr, , .sub.slip) and (velocity, steering angle)) and relationships between the above parameters, see Tire and vehicle dynamics (Butterworth-Heinemann, Oxford, 2002) by Hans Pacejka (see in particular Chapter 1 and 3 of the Pacejka reference).

(16) Different embodiments of the disclosure will now be described with reference to FIGS. 3-6.

(17) FIG. 3 schematically illustrates a system for determining a friction estimate between a road surface and a tire of a vehicle 100 according to an example embodiment of the disclosure. The system 300 comprises a force determining unit which may be embodied as an electrical powered assisted system 302 for a steering arrangement assisting the driver with power for steering of the vehicle (i.e., to apply a force on the axle rack 104). The force determining unit is configured to determine a rack force value indicative of a rack force on the axle rack 104. The system 300 further comprises vehicle state estimator comprising an inertial measurement unit (IMU) 308 for determining a lateral wheel force value indicative of a lateral force on the steerable wheel 102a-b. The IMU 308 may also be configured to determine wheel forces in other directions such as for example normal to the plane of the ground at the contact surface between the tire and road. The IMU 308 may comprise accelerometers and/or gyroscopes such that e.g., acceleration and velocities of the vehicle in at least the dimensions in the plane of the contact surface between the vehicle 100 tire and the ground can be determined. However, the IMU 308 may determine the acceleration and velocities in all three dimensions if required. Based on measured e.g., acceleration data, the vehicle mass and inertia, the lateral force can be determined (via e.g., Newton second law relating mass and acceleration to force). There is further a control unit 310 which is configured to acquire a plurality of rack force values indicative of rack forces on the axle rack 104 from the force determining unit 302 and the lateral rack force values from the IMU 308. The control unit module 310 is further configured to map a relation between a plurality of rack force values and a plurality of lateral wheel force values to a model comprising a friction parameter, and to determine the lateral friction estimate based on the mapping.

(18) Furthermore, the system 300 comprises a wheel speed sensor 314. The wheel speed sensor 314 is configured to determine the angular velocity of the wheel and may be provided in the form of a tachometer arranged in the hub of a wheel. In addition, the system may comprise a global positioning system (GPS) 312. With the GPS 312 and the wheel speed sensor 314 the velocity of the vehicle may be determined with improved accuracy.

(19) FIG. 4 illustrates a flow chart of method steps according to an embodiment of the disclosure. In step S402 a plurality of rack force values indicative of rack forces on the axle rack are acquired. A plurality of lateral wheel force values indicative of lateral forces on the steerable wheel are acquired in step S404. A relation between the acquired rack force values and the acquired lateral wheel force values is mapped S406 to a model comprising the friction parameter. The friction estimate is determined based on the mapping S408. The model comprises a plurality of predetermined relations between lateral wheel force and axle rack force each for a different friction parameter. During the mapping, the predetermined relation that matches the acquired data best (i.e., with the highest confidence), is selected, and the corresponding friction parameter is selected as the lateral friction.

(20) In one embodiment, the model is a linear relation between the lateral wheel force and the rack force including the friction parameter. The friction estimate is determined based on a plurality of ratios (n number of ratios) between rack force values (F.sub.rack) and lateral wheel force values (F.sub.y) sampled during a batch time duration. The ratios are used in a linear regression to fit a straight line to the accumulated data (F.sub.rack vs. F.sub.y) where the slope of the straight line may be given by:

(21) $K_{\mathrm{est}}=\frac{\underset{i=1}{\overset{n}{.Math.}}{F}_{y,i}{F}_{\mathrm{rack},i}}{\underset{i=1}{\overset{n}{.Math.}}{F}_{y,i}^2}.$

(22) The sampling of data (e.g., rack force values and lateral wheel force values) may be continued until an appropriate amount of data has been accumulated. For example, it may occur that data are acquired during unsatisfying conditions (not laterally well excited), e.g., during low wheel forces, too high sample variation, or unsatisfactory turn angle of the vehicle, then the unsatisfactory data points are disqualified and the accumulation continues.

(23) When the data batch is satisfactory, i.e., laterally well excited, the variance of the data is checked. If the variance is not too large, the data is processed for determination of the friction estimate. If the variance is too large, or if K.sub.est is lower than a threshold K-value, the data batch is disqualified and the confidence is set to zero.

(24) The friction estimate may be classified into different levels of friction depending on the estimate. For example, nominal friction values may be high, medium, or low with assigned nominal friction values 0.8, 0.5, 0.3, respectively. The classification may further be based on the vehicle velocity, vehicle velocity variation during the batch time duration, level of wheel force excitation, or distance to interval threshold. The classification levels have been empirically found, thus for a given set of vehicle velocity, vehicle velocity variation during the batch time duration, level of wheel force excitation, or distance to interval threshold a certain classification is set. The level of wheel force excitation is the level of lateral wheel force, for example, if the lateral wheel force is too weak the estimation of K.sub.est is not accurate since the data spread is narrow leading to small gradient of the linear fit. The distance to interval threshold is how close the acquired fitted data is to match another model relation. For example, for the classification, the parameter K.sub.est is determined to be within a certain predetermined interval among several intervals and if the difference between the determined K.sub.est and an interval end or start point is small the confidence may be reduced. The classification is thus based on the confidence value that depends on vehicle velocity, vehicle velocity variation during a batch time duration, level of wheel force excitation, and distance to interval threshold. The confidence dependencies on these parameters and the respective tables may be empirically predetermined.

(25) FIGS. 5A-B each illustrates a graph of exemplary rack force data (F.sub.rack) versus lateral wheel force (F.sub.y). The input for the method according to this embodiment is the rack force data (F.sub.rack) and the lateral wheel force data (F.sub.y). The data batch (500 in FIG. 5A and 201 in FIG. 5B) has been accumulated for a batch time duration and are considered well excited and with sufficiently high confidence. In this case, a linear model is used and thus K.sub.est is included in the linear regression as described above. Based on the linear regression, a linear curved is fitted with the rack force data (F.sub.rack) versus lateral wheel force (F.sub.r). The linear curves 502 and 503 in FIG. 5B has gradients K.sub.est1 and K.sub.est2 respectively, which determines a classification of the nominal friction between the road and tire. Friction is classified into high, medium, and low depending on the value of K.sub.est (i.e., the gradient of the linear fit). For example, a high friction may be assigned a value of nominal friction of 0.8, medium friction may be assigned a nominal friction of 0.5, and low friction may be assigned a nominal friction of 0.3. The classification levels (e.g., high, medium, low) are predefined based on empirical data, for example it has empirically been found that when K.sub.est is above a threshold, the friction is high and thus a nominal friction of 0.8 may be assigned. Alternatively or additionally, the classification levels are found from knowledge of the linkage arm length (d) and the pneumatic trail based on torque equilibrium described with reference to FIGS. 1B-C. In the exemplary case shown in FIG. 5A, the estimated K.sub.est1 is below a threshold k.sub.3 but exceeds a threshold k.sub.2 and thus is assigned a nominal friction of 0.8 and is thus classified into high friction. In the exemplary case shown in FIG. 5B, the estimated K.sub.est2 exceeds a threshold k.sub.1 and thus is assigned a nominal friction of 0.3 and is classified into low friction. If a third K.sub.est3 is below the threshold k.sub.2 but exceeds the threshold k.sub.1 it may be assigned a nominal friction of 0.5 and is thus classified into medium friction. The relation between the thresholds may be that k.sub.1<k.sub.2<k.sub.3. With linear curve fitting in combination with the classification procedure, a convenient way to estimate the friction conditions without requiring complex modelling of the physical conditions between the tire and road.

(26) In a further embodiment, still with rack force data (F.sub.rack) and lateral wheel force data (F.sub.r) as the input to the method, a non-linear model function is used. FIGS. 6A-B each illustrate a graph of such exemplary rack force data (F.sub.rack) versus lateral wheel force (F.sub.r) for a high friction case (FIG. 6A) and a low friction case (FIG. 6B). In each of the graphs shown in FIGS. 6A-B, a fitted curve 602, 603 based on a non-linear model is shown. The model used comprises a friction parameter directly included in the model function. For example, the model may be described by the function:
F.sub.rack/F.sub.z=*tan h(k/*F.sub.y/F.sub.z),

(27) Where the k is the pneumatic trail to steering arm length ratio, F.sub.y is the lateral wheel force and F.sub.z is normal wheel force, i.e., along the normal between the road and wheel contact surface. The lateral and normal wheel forces may be measured as described with reference to FIG. 3, i.e., using an IMU 304. Thus from collected data (F rack and lateral wheel forces F.sub.y) the friction can be estimated by comparing the fitted curve with predetermined curves 604 (see FIG. 6C) which have been predetermined under known friction conditions. In other words, the fitted curve 602, 603 which is based on a non-linear model is matched with the pre-defined curves 604 determined with known friction conditions. The best match determines the friction estimate. In the exemplary cases shown in FIGS. 6A-B the fitted curve 602 matches with curve 606 which was determined under the nominal friction =0.8 and the fitted curve 603 matches with curve 605 which was determined under the nominal friction =0.3. The matching of the fitted curves with the predetermined curves may be performed using e.g., least square procedures known in the art.

(28) In yet another embodiment, the model is a physical model of the friction between the vehicle tire and the road surface is used. The lateral force is generally a function (F.sub.y(T.sub.WhlStr, x)) of the wheel torque (T.sub.WhlStr) and the pneumatic trail (x), see FIGS. 1A-C. The lateral force F.sub.y is measurable via the rack force F.sub.rack as described with reference to FIGS. 2A-B and FIGS. 1A-C. Furthermore, the pneumatic trail (x) generally depends on the velocities of the vehicle (lateral velocity, longitudinal velocity, and yaw velocity) the steering angle (see FIG. 1A) of the vehicle, and the friction between the road surface and the tire. The steering angle is the angle between a longitudinal centerline 130 of the vehicle (between rear and front of the vehicle) and the wheel hub direction W.sub.d (see FIG. 1A). The steering angle may be measured by a steering angle sensor arranged on the steering column (not shown) of the vehicle. The sensor may measure the rotational angle of the steering wheel and thereby derive the steering angle. Another way to measure the steering angle is to determine the position (translational position) of the axle rack and therefrom derive the steering angle. Accordingly, the pneumatic trail can generally be provided as a function (x (velocity, steering angle, ) of the steering angle and the velocities. From the two functions F.sub.y(T.sub.WhlStr, x) and x (velocity, steering angle, ) the friction may be derived and this model may be used as a physical model of the friction. Measured lateral wheel force values and rack force values may thereby be mapped onto predetermined relations between lateral wheel force and rack force for known friction conditions. For example, as previously discussed with reference to torque equilibrium,
F.sub.y*x=d*F.sub.rack, i.e., F.sub.rack=F.sub.y*x/d.

(29) Thus, by mapping measured rack force data and measured lateral wheel force data to a model of in this case the pneumatic trail comprising a friction parameter, a friction estimate may be obtained. For further details regarding the functions (F.sub.y(T.sub.WhlStr, x) and x (velocity, steering angle, )) and relationships between the above parameters, see Tire and vehicle dynamics (Butterworth-Heinemann, Oxford, 2002) by Hans Pacejka (see in particular Chapter 1 and 3 of the Pacejka reference).

(30) The control unit may include a microprocessor, microcontroller, programmable digital signal processor or another programmable device. It should also be noted that the system 300, force determining unit, electrical powered assisted system 302, inertial measurement unit 308, control unit module 310, GPS 312, as well as any other unit, module, system, device, arrangement or the like described herein may comprise and/or be implemented in or by one or more appropriately programmed processors (e.g., one or more microprocessors including central processing units (CPU)) and associated memory and/or storage, which may include operating system software, application software and/or any other suitable program, code or instructions executable by the processor(s) for controlling operation thereof, for providing and/or controlling interaction and/or cooperation between the various features and/or components described herein, and/or for performing the particular algorithms represented by the various functions and/or operations described herein.

(31) The person skilled in the art realizes that the present disclosure by no means is limited to the embodiments described above. On the contrary, many modifications and variations are possible within the scope of the appended claims.

(32) In the claims, the word comprising does not exclude other elements or steps, and the indefinite article a or an does not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measured cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope.

(33) While exemplary embodiments are described above, it is not intended that these embodiments describe all possible forms of the disclosure. Rather, the words used in the specification are words of description rather than limitation, and it is understood that various changes may be made without departing from the spirit and scope of the disclosure. Additionally, the features of various implementing embodiments may be combined to form further embodiments of the disclosure.