FIELD OF THE INVENTION The present invention relates in general to a method, system and a device for an advanced load identification technique that enables amongst others vehicle dynamic performance evaluation .

BACKGROUND OF THE INVENTION

In-depth vehicle dynamic performance evaluations have as yet been carried out through identification of the time-domain loads acting on a vehicle body structure. Vehicle body structure deformation can be calculated by using these time-domain loads for identifying potential weak spots in the vehicle body design. Furthermore, the transient build-up of each load can be evaluated for different vehicle configurations and/or different vehicle body configurations to understand how changes in suspension or body characteristics affect the total vehicle performance. Time-domain body loads can be used for identifying objective parameters describing vehicle performance and such parameters can be linked to subjective dri- ver ratings. Moreover, benchmarking between different

vehicles for comparing vehicle performance, specifically vehicle suspension performance, can be based on time-domain loads for e.g. identifying the optimization potential of a currently used suspension layout.

As yet, accurate time-domain load identification for vehicle dynamic performance evaluations requires high-channel -count systems, i.e. high-channel -count vehicle body instrumentation with usually more than a hundred channels and special strain- gauges for obtaining time-domain strain-data. For obtaining the relation between body load and strain, transfer- function (FRF; frequency response functions) measurements are

performed for identifying the strain to force calibration va- lues for all body loads and instrumented strain gauges. A calibration matrix comprising the identified calibration values will be used together with operational strain-data to identify the body loads. Said calibration matrix, referred to as strain-to-force calibration matrix or simply as

calibration matrix in the following, needs to be inverted for further use. A multiplication of a vector comprising the strain-data and the inverted calibration matrix finally yields the loads impacting on the vehicle body structure.

This prior art approach suffers from certain drawbacks. On the one hand it involves a large number (usually 150 to 250) of strain-gauges applied to the body structure for strain- data acquisition, with each strain-gauge having to be

installed on the body structure and wired to a measuring system. On the other hand it involves a large and complex transfer-function measurement effort for identifying the

calibration matrix. Furthermore, even with high-channel -count instrumentation and optimized transfer- function measurement approaches this method has significant mathematical limitations in case of mechanical structures with a lot of close-by acting body loads, which cannot be identified accurately due to matrix conditioning problems. Additionally, the calibration matrix needs to be re-measured for each different vehicle body variant, giving an extra potential uncertainty on the identified loads as each resulting calibration matrix can have its own limitations related to conditioning, signal-to-noise ratio etc.

Therefore there is a need for an improved method, system and device for an efficient load identification technique that can be used for vehicle dynamic performance evaluations and addresses or ameliorates these and other problems. SUMMARY OF THE INVENTION

According to a first aspect of this disclosure there is provided a method for an efficient load identification technique that can be used for vehicle dynamic performance evaluations. The relevant vehicle comprises a vehicle body and a suspension system. A plurality of sensors for acquiring strain-data is assigned to the vehicle's suspension system, e.g. by assigning each sensor to a suspension system compo- nent .

The proposed method comprises the steps of:

acquiring a system response from a plurality of sensors under dedicated conditions in a test environment, wherein the said sensors are assigned to the vehicle's suspension system; acquiring system loads by means of applying the same dedicated conditions in a simulation environment;

generating a calibration matrix based on data pertaining to the said system response and on data pertaining to the said system loads;

acquiring an operational system response from the said sensors in a test-track situation; and

using the acquired operational system response and the said calibration matrix for calculating numerical values for the loads impacting the vehicle.

In a preferred, specific embodiment the method comprises the steps of:

acquiring strain-data from a plurality of sensors by means of applying kinematics & compliances scenarios (K&C scenar- ios) in a test environment,

acquiring suspension system to vehicle body loads data by means of applying the same K&C scenarios in a simulation environment ;

generating a calibration matrix based on the acquired strain-data and on the acquired suspension system to vehicle body forces-data;

acquiring operational strain-data from the said sensors in a test-track situation; and using the said operational strain-data and the said calibration matrix for calculating numerical values for the loads impacting the vehicle. The resulting numerical values for the loads, i.e. forces, identified can be used for e.g. recommendations for suspension design and/or body design improvement and also for vehicle benchmarking . Multiple K&C scenarios are per se known in the art and applying K&C scenarios to a vehicle is usually performed using laboratory test systems, e.g. dedicated kinematics & compliance test machines and vehicle dynamic simulators. K&C testing is used to accurately establish the kinematic character- istics of a vehicle's suspension and steering system geometries and the compliance characteristics of the suspension springs, anti-roll bars, elastomeric bushes and component deformations. Knowledge of these characteristics is an essential aid for suspension engineers wishing to establish a thorough understanding of the vehicle's performance in terms of ride, impact isolation, steering and handling.

The resulting calibration matrix is defined in force over strain, implying that it does not need to be inverted anymore for the force estimation process, which strongly enhances the force estimation stability. The calibration matrix is

calculated using load-data and strain-data obtained in a K&C simulation environment and a K&C test environment

respectively. Thus the prior art measurement of frequency response functions is replaced by the use of dedicated K&C scenarios both in a test environment and a simulation environment. After the calibration matrix is calculated, it can be applied to operational strain data (E(t)) from track measurements to calculate the time-domain track forces (F(t)) acting on the car body structure or on suspension components.

According to another aspect of this disclosure there is provided a system for an efficient load identification technique that can be used for vehicle dynamic performance evaluations, comprising means for performing the said method. In an embodiment of a system for identifying loads impacting a vehicle which comprises a vehicle body and a suspension system the said system comprises a processing unit and a plurality of sensors for obtaining strain-data, wherein the said sensors are assigned to the vehicle's suspension system. The said system is adapted for acquiring strain-data from said sensors by means of applying K&C scenarios in a test environ- ment and for acquiring suspension system to vehicle body loads data by means of applying the same K&C scenarios in a simulation environment. The said system is furthermore adapted for generating a calibration matrix based on the said strain-data and the said suspension system to vehicle body loads data. The said system is still further adapted for acquiring operational strain-data from the said sensors in a test-track situation and using the said operational strain- data and the said calibration matrix for calculating numerical values for the loads impacting the vehicle.

According to a still further aspect of this disclosure there is provided a device for an efficient load identification technique, adapted for being used in the said system and/or adapted for performing the said method. An embodiment of a device according to this aspect of the disclosure is termed "processing unit" in the following.

A processing unit for identifying loads impacting a vehicle according to the approach proposed in this disclosure is adapted for

acquiring strain-data from said sensors by means of applying K&C scenarios in a test environment,

acquiring suspension system to vehicle body loads data by means of applying the same K&C scenarios in a simulation en- vironment,

generating a calibration matrix based on the said strain- data and the said suspension system to vehicle body loads data, acquiring operational strain-data from the said sensors in a test-track situation; and

using the said operational strain-data and the said calibration matrix for calculating numerical values for the loads impacting the vehicle,

or, more generally, adapted for executing the steps of the method proposed in this disclosure.

The sensors used by the processing unit and connected to the processing unit for transmitting strain-data to the processing unit are assigned to the vehicle's suspension system, e.g. in that each sensor is assigned to one component of the vehicle's suspension system. Another aspect of the present invention involves a computerized system for an efficient load identification

technique, comprising means for executing a computer program providing an implementation of the method proposed in this disclosure, for example a computerized system comprising a processing unit, wherein the processing unit comprises a memory adapted for storing a computer program and for storing data generated by and processed with the computer program, and a processor adapted for executing a computer program stored in said memory. The processing unit and its memory and processor as well as the computer program providing an implementation of the method proposed in this disclosure all constitute means for an efficient load identification

technique and the said computer program is executed by said processor when the processing unit is employed for a load identification for e.g. vehicle dynamic performance evaluations. The aforementioned processing unit, when comprising a processor, memory and a computer program implementing the method proposed in this disclosure, can be perceived as an exemplary embodiment of a computerized system for an

efficient load identification technique. Further aspects, features and advantages of the present invention will become apparent from the drawings and detailed description of the following preferred embodiments.

BRIEF DESCRIPTION OF THE DRAWINGS

The above-mentioned and other concepts of the present invention will now be addressed with reference to the drawings of the preferred embodiment of the present invention. The shown embodiment is intended to illustrate, but not to limit the invention. The drawings contain the following figures, in which like numbers refer to like parts throughout the description and drawings and wherein

FIG 1 is a simplified overview of a prior-art approach load identification technique,

FIG 2 shows a vehicle, comprising a vehicle body and sus- pension system, with sensors assigned to suspension system components,

FIG 3 shows a system enabling load identification

comprising sensors assigned to components of the sus- pension system of a vehicle and a processing unit for receiving and processing data from such sensors, and

FIG 4 is a simplified diagram of a method for efficient

load identification for vehicle operational

conditions according to the approach disclosed herein .

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

Overview This disclosure describes a method, a system and a device providing an efficient load identification technique that amongst others can be used for vehicle dynamic performance evaluations. Those skilled in the art will readily understand that the approach proposed here can be applied to various types of vehicles, e.g. passenger cars, sports cars, trucks, and also to parts of such vehicles, e.g. a soft mounted subframe, different suspension components and either the front or the rear axle. Furthermore those skilled in the art will readily understand that the approach can be used by OEM vehicle producers, test or inspection facilities and/or the racing industry.

FIG 1 shows in simplified form the prior art method for identifying loads impacting the body structure 12 of a vehicle 10. To this end, FIG 1 shows a vehicle 10, comprising a vehicle body structure 12, wherein a plurality of strain- gauge sensors 14 (only three are shown) is assigned to the body structure 12. It is to be understood that the said sen- sors 14, while being shown as being applied on the outside of the body structure 12, can in practice be applied to the outside of the body structure 12 or to internal parts of the body structure 12. The totality of the sensors 14, each being applied at its relevant location, is referred to as the load identification instrumentation, or briefly the instrumentation, in the following.

The matrix-equation shown at the topmost part of FIG 1 comprises a force vector 30, a calibration matrix 32 and a strain-data vector (strain vector) 34. The force vector 30 is unknown. The components of the force vector 30 pertain to various forces acting on the vehicle 10 and its body structure 12. Determining the force vector 30 and its components is what is aimed at when performing a load identification to enable for example a vehicle dynamic performance evaluation.

The calibration matrix 32 and its parameters are unknown, too, and cannot be measured. The strain vector 34 comprises _n

the strain response data from the strain-gauge sensors 14 acquired through measurement in vehicle operational

conditions. Once the calibration matrix 32 is known, the force vector 30 can be calculated on the basis of the

calibration matrix 32 and the strain vector 34.

However, the calibration matrix 32 comprises force over strain data, symbolically shown in FIG 1 as E[ _7JZ ] Ιε[ _1i2i3i4 ] that cannot be measured. The calibration matrix 32 has thus as yet been determined indirectly by means of transfer function testing on the body structure 12 on the basis of frequency response function (FRF) measuring. This yields strain over force relationships ( ε [i _{i 2} , ₃ ,4] /F[ _Y,Z ] ) · ^A resulting matrix

comprising said strain over force data is to be inverted for allowing a calculation of the force vector 30, as shown in the lowermost part of FIG 2. Inversion of the strain over force matrix is complex and may yield a calibration matrix 32 with limited load separation capabilities due to poor

conditioning of the strain over force matrix, wherein the latter results from strongly coupled portions of the body structure 12 of the vehicle 10 and a resulting dependency of the strain over force data.

To overcome the poor conditioning of the strain over force matrix, prior art approaches have suggested to use an

increased number of sensors 14, resulting in a large but intended overdetermination of strain measurements by e.g. using at least three strain-gauge sensors 14 per load to be identified. However, this requires an increased amount of sensors 14 to be applied to the body structure 12 of the vehicle 10 and results in large and complex instrumentation setups (often two hundred and more sensors 14) , that take a considerable amount of time to instrument and requires a system capable of processing the resulting data. The said overdetermination yields an increased dimension of the strain vector 34 and consequently an increased dimension of the strain over force matrix, which in turn makes the inversion of this matrix ever more complex. Once the strain over force matrix has successfully been inverted, the resulting calibration matrix 32 and data measured in vehicle operational conditions, i.e. the data comprised by the strain vector 34, can be used for

identifying the loads acting on the vehicle 10 by determining the force vector 30. However, the calibration matrix 32 stemming from the frequency response function (FRF) measuring performed on the body structure 12 is body dependent and thus needs to be established again and again for every new body variant that is evaluated in a project.

The approach presented in the following provides a solution for overcoming these drawbacks .

A first aspect of the approach involves that the sensors 14 are assigned to a vehicle suspension system 16, as opposed to the sensors 14 being assigned to the vehicle body structure 12 in prior art (FIG 1) . To this end, FIG 2 shows the vehicle 10 of FIG 1, comprising a vehicle body structure 12 and a suspension system 16, the latter being exemplified by a front and a rear suspension spring. The sensors 14 (only two are shown) are assigned to the suspension system 16 and its individual suspension components, e.g. suspension springs, late- ral suspension links, stabilizer bars, dampers, etc., as indicated in FIG 1. In an exemplary embodiment at least one sensor 14, e.g. a strain-gauge sensor 14, is applied to each major component of the suspension system 16. FIG 3 shows a system 20, comprising the sensors 14 of FIG 2 each being assigned to one suspension system component, and a processing unit 22 for processing strain-data obtained by the said sensors 14. The sensors 14 are communicatively connected to the processing unit 22 in an appropriate manner, i.e.

wire-bound or wireless. The processing unit 22 thus functions as a measuring system for receiving and processing the relevant strain-data. Since the strain response from a component of the suspension system 16 is the result of a relatively limited set of forces compared to a strain response from the vehicle body structure 12, processing the strain-data obtained from sensors 14 assigned to the suspension system 16 is a far less complex problem to solve and less overdetermination, i.e. number of responses required to identify one force, is needed. For example, nine forces pertain to a lateral link of the suspension system 16, i.e. three forces acting on each end and three forces acting in the middle of the link from a strut that is supported on the link. In contrast, the strain on the vehicle body structure 12 results from forty to sixty forces. Consequently, a significantly reduced number of sensors 14 are required in the instrumentation, i.e. as part of the sys- tern 20, when the sensors 14 are assigned to the suspension system 16.

A second aspect of the approach presented here involves that the prior art transfer- function calibration measurements are replaced by a number of measurements obtained in unique kinematics & compliance scenarios, e.g. vehicle bounce, vehicle roll, lateral parallel or opposed inputs at the tire patches, aligning torque parallel or opposed inputs at the tire patches, longitudinal inputs at the tire patches etc., in both a test environment and a simulation environment.

In a test environment, the suspension strain responses are measured during execution of a number of said kinematics & compliances scenarios, termed K&C scenarios for the sake of brevity in the following.

In a simulation environment, the suspension to body loads are identified for the same K&C scenarios applied in the test environment. The simulation environment uses a multi-body simu- lation model known per se in the art.

A simulation in the simulation environment employing a K&C scenario is termed K&C simulation in the following and a test in the test environment employing a K&C scenario is termed K&C test, accordingly. The strain responses and the suspension to body loads are denominated as E and F, respectively. The K&C tests and K&C simulations provide strain-data E on the one hand and load-data F on the other hand. Both are acquired under identical boundary conditions in the test environment and the simulation environment, respectively.

Knowing the loads F in defined conditions (the K&C scenarios) and having the suspension responses - the strain data E - in those same identical conditions enables identification of the link between the loads F of interest and the suspension strain responses E. Said link can be written as H = F * E ^"1 , wherein H is the calibration matrix 32, F is the force vector 30 and E is the strain vector 34 of the matrix equation shown in FIG 1 (F = H * E <=> H = F * E ^"1 ) and the determination of the calibration matrix 32 on the basis of the load-data F and the strain-data E is a matrix multiplication of a force matrix based on the load-data F with the inverse of the strain- data matrix based on the strain-data E.

As multiple K&C tests and K&C simulations are applied, wherein each K&C scenario excites unique phenomena in the suspension system 16, this approach results in a robust rela- tion between the loads F of interest and the suspension strain response. The resulting calibration matrix 32 can be used to identify the time-domain loads when performing full vehicle operational tests on a test-track. Through this different load calibration approach there is no need for body- side transfer- function measurements and since the identification process of the calibration matrix 32 is mathematically different from the approach in the prior art, the calibration matrix 32 does not need to be inverted anymore, resulting in a far more robust and accurate load identification.

The system 20 (FIG 2) performs the following steps 40-48, shown in FIG 4, for identifying the loads acting on a vehicle 10. In a first step 40 the system 20 acquires strain-data E from the sensors 14 applied to the suspension system 16 and generates a strain-data vector (single K&C scenario) or a strain-data matrix 35 (multiple K&C scenarios) . This data is obtained by applying at least one K&C test on the relevant vehicle 10.

In a second step 42 the system 20 calculates suspension-to- body forces F by applying the same K&C scenarios as applied in the first step 40 in a simulation environment and genera- tes a force vector (single K&C scenario) or a force matrix 31 (multiple K&C scenarios) .

On account of a vector constituting a special matrix, i.e. a nxl-matrix, a strain-data vector or a strain-data matrix 35 as well as a force vector or a force matrix 31 are jointly termed strain-data matrix 35 and force matrix 31

respectively, or briefly matrix 35, 31.

In a third step 44 the system 20 generates the calibration matrix 32 on the basis of the matrices 35, 31 generated in the first and second step 40, 42. The calibration matrix 32 results from a combination (matrix multiplication) of the matrix 35 and the matrix 31 or alternatively from any other applicable option for reconstructing the calibration data from a number of known forces, e.g. forces as comprised in the force matrix 31, and a number of known responses, e.g. responses as comprised in the strain-data matrix 35.

With the first, second and third step 40-44 being performed by means of the system 20, the mathematical tools for

identifying time-domain body loads acting on the body structure 12 of a vehicle 10 have been prepared.

In a fourth step 46 and subsequent steps these tools are now applied. In a fourth step 46 track measurements (as opposed to the K&C test applied in the first step 40) are performed for acquiring operational strain-data from the sensors 14 applied to the suspension system 16 and a strain-data vector 34 comprising operational strain-data is composed with this data .

In a fifth step 48, using the calibration matrix 32 generated in the third step 44 and the operational strain-data in the operational strain-data vector 34 generated in the fourth step 46, the time-domain body loads, i.e. the elements of the force vector 30 resulting from a matrix multiplication of the calibration matrix 32 and the operational strain-data vector 34, are determined.

The fourth step 46 and the fifth step 48 can be repeated as often as necessary and desired for identifying loads, i.e. forces, acting on the vehicle body 12 according to this novel approach presented here.

The steps 40-48 shown in FIG 4 can be perceived as a

simplified representation of a computer program 50 providing an implementation of the method for efficient load identifi- cation proposed in this disclosure. Such a computer program 50 will be stored in a memory (not shown) of the processing unit 22 of the system 20 shown in FIG 2 or a similar device for efficient load identification and will be executed by means of a processor (not shown) of the said processing unit 22 for efficient load identification according to the approach presented in this disclosure.

By employing the method proposed in this disclosure, more particularly by employing a computer program 50 providing the said method, load identification can be performed faster and at lower costs. The proposed method requires less instrumentation efforts on account of the fewer required sensors 14. The load calibration matrix 32 is not dependent on the body structure 12 of the relevant vehicle 10.

Generally, the load calibration matrix 32 is generated by means of K&C tests and simulations, wherein at least one dedicated K&C scenario, usually a number of dedicated K&C scenarios, e.g. scenarios known as "vertical parallel",

"vertical opposed", "lateral parallel", "lateral opposed", "aligning torque parallel", "aligning torque opposed", "longitudinal" etc. in professional terminology, are applied. This replaces the fully test-based prior art approach of measuring the transfer- functions between strain-responses and body forces. As the calibration matrix 32 resulting from the K&C tests and simulations does not need to be inverted anymore, the subsequent load identification procedure (fifth step 48) does not suffer from the potential conditioning problems of the prior art load identification approach.

It is to be understood that the embodiment used for

describing the approach proposed in this disclosure is but one example and not to be construed as limiting. Therefore, while vehicle dynamics performance analysis has been used as the basis for the present disclosure, the proposed load identification method can be used for a wide variety of applications, such as vehicle primary ride evaluation, in-depth sus- pension performance evaluation, acquisition of objective data that can be linked to subjective perceptions of drivers etc. Furthermore, while the present disclosure has been based on K&C scenarios for ascertaining a dedicated unique boundary condition for the relevant system, e.g. a vehicle 10, it is to be understood that providing such unique boundary

conditions (dedicated conditions) is the key aspect and that these may be provided by other means, such as an application of known loads to the physical vehicle or to a vehicle component, e.g. the engine block, while acquiring system responses (e.g. strain responses) that can be made proportional to the loads of interest. In a simulation environment, similar load applications can be applied to the system, where the system interface loads of interest are calculated - for example the engine to body loads. Therefore, the approach proposed in this disclosure can be generalized as being a method, system, apparatus etc. for identifying loads impacting a vehicle 10, wherein said vehicle 10 comprises a vehicle body 12 and a suspension system 16, and wherein the method comprises and the system or apparatus executes or is adapted for executing the steps of:

acquiring a system response (e.g. strain-data 35) from a plurality of sensors 14 assigned to the vehicle's 10 suspen- sion system 16 under dedicated conditions in a test environment (e.g. under conditions resulting from applying K&C scenarios in a test environment) ;

acquiring system loads (e.g. suspension system to vehicle body loads data 31) by means of applying the same dedicated conditions in a simulation environment (e.g. by applying conditions resulting from applying the same K&C scenarios as used when acquiring the system response in a test environment) ;

generating a calibration matrix 32 based on data pertain- ing to the said system response (e.g. the said strain-data

35) and on data pertaining to the said system loads (e.g. the said suspension system to vehicle body forces-data 31) ;

acquiring an operational system response (e.g. operational strain-data 34) from the said sensors 14 in a test-track situation; and

using the said operational system response (e.g. strain- data 34) and the said calibration matrix 32 for calculating numerical values for the loads impacting the vehicle 10. The present invention provides a multitude of advantages to the supplier and system integrator. First and foremost is the more robust and accurate load identification, which is due to the previously required matrix inversion having been rendered unnecessary by the approach disclosed here. Secondly, significantly less sensors 14 are required to enable the load identification saving both instrumentation and data analysis time. The developed approach requires instrumentation for the forces of interest only, as there is no cross-influence between forces in the calibration matrix 32. The calibration matrix 32 is identified for each force (each row)

individually and the presence of one force in the calibration matrix 32 does not influence the calibration value of another force. Third, the time consuming transfer function measurements for load calibration are replaced by short duration K&C tests and simulations giving another timing advantage. Overall, the total throughput time from instrumentation time to identified loads is significantly reduced compared to the classic load identification approach. Also, with the sensors 14 being applied to the vehicle's suspension system 16 and not to the vehicle body 12, the calibration matrix 32 does not need to be re-measured for different vehicle body variants and once generated, the calibration matrix 32 can be re-used for the subsequent calculation of numerical values for the loads impacting a vehicle 10.

In addition to the embodiments of the aspects of the present invention described above, those of skill in the art will be able to arrive at a variety of other arrangements and steps which, if not explicitly described in this document,

nevertheless embody the principles of the invention and fall within the scope of the appended claims.