4
You will find the figures mentioned in this article in the German issue of ATZ 11I2007 beginning on page 1058. Optimierung des Lenkverhaltens – Kundenwünsche und Zielwerte Optimization of Steering Behavior Customer Requirements and Target Parameters Authors: Bernhard Schick, Masaki Yamamoto, Ikuo Kushiro, Stefan Resch, Robert Matawa and Naoki Hagiwara The multi-dimensional field of driving dynamics offers a multitude of interpretations of the type of vehicle the cus- tomer would wish for. Between the two poles customer’s emotions and facts and figures required by engineers, the application of the method “Quality Function Deployment” (QFD) is intended as an effective tool for translation. TÜV Süd Automotive captured together with Toyota Motor Corporation with the method “Improve by QFD” the cus- tomer’s wishes at the example steering behavior accord- ing to statistical principles, translated these information into target parameters for the R&D engineers, realized them in a prototype vehicle and finally validated the fulfill- ment of the customer’s expectations successfully. 1 Introduction All car manufacturers strive to differenti- ate their products in order to distinguish themselves from potential competitors and to offer a product to their customers, which is as attractive as possible. In product devel- opment, outside of chassis engineering, methods such as bottle-neck engineering, TRIZ, QFD and others are widespread. On the one hand, the aim is to realize custom- er wishes as closely as possible and on the other, to minimize development as well as production costs. Usually, a known custom- er wish serves as the origin of a new pro- duct or functional feature, which is directly derivable from this wish. The multi-dimensional field of driving dynamics offers a multitude of interpreta- tions of the type of vehicle the customer would wish for. In this context several, mu- tually influencing product aspects such as the steering system, kinematics design of the chassis or perhaps merely the selection of suitable tires could be attributable to a customer wish, which might have been worded as “direct and precise steering.” The specification of target parameters for the items named as the basis for the design process is the job of the engineers. Between these poles, that means the customer’s emotions and the facts and figures required by engineers, the application of the meth- od “Quality Function Deployment” (QFD), a tried and tested process, is intended as a tool. Because for all the different interpre- tations of the handling the quotation of Bernd Pischetsrieder applies equally: “If only engineers identify differences, there is no value for the customer” [1]. The method “Improve by QFD” (devel- oped by TÜV Süd Automotive GmbH) cap- ATZ 11I2007 Volume 109 15

Optimization of steering behavior

Embed Size (px)

Citation preview

Page 1: Optimization of steering behavior

You will find the figures mentioned in this article in the German issue of ATZ 11I2007 beginning on page 1058.

Optimierung des Lenkverhaltens –

Kundenwünsche und Zielwerte

Optimization of Steering Behavior

Customer Requirements and Target

Parameters

Authors:Bernhard Schick, Masaki Yamamoto, Ikuo Kushiro, Stefan Resch, Robert Matawa and Naoki Hagiwara

The multi-dimensional field of driving dynamics offers a multitude of interpretations of the type of vehicle the cus-tomer would wish for. Between the two poles customer’s emotions and facts and figures required by engineers, the application of the method “Quality Function Deployment” (QFD) is intended as an effective tool for translation. TÜV Süd Automotive captured together with Toyota Motor Corporation with the method “Improve by QFD” the cus-tomer’s wishes at the example steering behavior accord-ing to statistical principles, translated these information into target parameters for the R&D engineers, realized them in a prototype vehicle and finally validated the fulfill-ment of the customer’s expectations successfully.

1 Introduction

All car manufacturers strive to differenti-ate their products in order to distinguish themselves from potential competitors and to offer a product to their customers, which is as attractive as possible. In product devel-opment, outside of chassis engineering, methods such as bottle-neck engineering, TRIZ, QFD and others are widespread. On the one hand, the aim is to realize custom-er wishes as closely as possible and on the other, to minimize development as well as production costs. Usually, a known custom-er wish serves as the origin of a new pro-duct or functional feature, which is directly derivable from this wish.

The multi-dimensional field of driving dynamics offers a multitude of interpreta-tions of the type of vehicle the customer would wish for. In this context several, mu-

tually influencing product aspects such as the steering system, kinematics design of the chassis or perhaps merely the selection of suitable tires could be attributable to a customer wish, which might have been worded as “direct and precise steering.” The specification of target parameters for the items named as the basis for the design process is the job of the engineers. Between these poles, that means the customer’s emotions and the facts and figures required by engineers, the application of the meth-od “Quality Function Deployment” (QFD), a tried and tested process, is intended as a tool. Because for all the different interpre-tations of the handling the quotation of Bernd Pischetsrieder applies equally: “If only engineers identify differences, there is no value for the customer” [1].

The method “Improve by QFD” (devel-oped by TÜV Süd Automotive GmbH) cap-

ATZ 11I2007 Volume 109 15

Page 2: Optimization of steering behavior

tures the customer’s wish according to sta-tistical principles and processes this infor-mation into target parameters for engi-neers, using a QFD approach, which has been specifically adapted to the aspect of handling and drivability. The investment of time and costs required to obtain the outcome, a scenario for fulfilling customer wishes through respective engineering, is optimized by the method itself. In collabo-ration with Toyota Motor Corporation [8] the “Improve by QFD” method was to be used for the first time to show the custom-er’s preferred choice in the feature “steer-ing behavior” in a prototype to be subse-quently verified/confirmed again by the end user in a final customer test ride. Previ-ous studies had shown that the complex subject of steering behavior contributes sig-nificantly to an attractive driving experi-ence and is one of the top priorities on cus-tomer’s wish lists. For the program, the lower and upper mid-sized car segments were selected and shown, using a suitable benchmark of a total of eight vehicles. The objective of the project was to create an at-tractive design of steering behavior.

2 Schematic Presentation of “Improve by QFD”

Central element of the methodology pre-sented in this paper is the House of Quality (HoQ). Quality in this context is defined as the fulfillment of customer wishes. Figure 1 shows a principle schematic of the House of Quality. While Figure 1 shows in the ver-tical frame the engineers speech, the hori-zontal frame shows the customer language. This is where the customer’s statements are captured using statistical-psychological tools, and processed into a development ob-jective from the customer’s perspective, the “3b customer targets.” This result ob-tained from steps 1 to 3 is intended to pro-vide answers to a question, which might be formulated like this: “How would we like to have customer’s rate our new product, or evolution of an existing product, during the next survey?” As a suitable tool for por-traying these customer targets the so-called Market Opportunity Map (MOM) has estab-lished itself.

In the vertical frame of Figure 1 the en-gineers speech is shown, as mentioned be-fore. In the steps 4 and 5 the engineering options for product optimizations based on the stringent, methodical QFD approach are listed and linked to the collection of customer wishes. The steps 6 to 9 produce the effort and cost-optimized generation of

engineering measures needed to achieve a product evolution.

3 Customer Wishes, Weighting and Customer Targets

The entire customer survey illustrated by the example of steering behavior was performed by chassis specialists with specific training in interviewing techniques at selected locations in Europe. The locations were selected so as to ensure a representative set of findings. Over 200 end customers drove the benchmark vehi-cles while the specialist guided the interview and recorded the responses from the passen-ger’s seat. To complete the picture, a similar program was conducted with a group of 18 European media representatives from auto-motive magazines to TV journalists.

The survey was conducted in a three-level matrix with basically open questions and neutral key words. This was designed to avoid any influence by the specialists be-ing exerted on the customer’s choice of words, which would have adulterated the customer’s language.

Furthermore, this method offers the major benefit of being able to gather cus-tomer wishes which would not be reflected in a closed questionnaire. In addition, ex-perience has shown that the compulsory response to questions which exceed the perception and ability to make judgments of some customer’s reduces the signifi-cance of the result.

Among other things, the key words are derived from a functional model “steering behavior / steering feeling” created by a workshop team of specialists during the preparatory phase. This functional model reflects the interactions, which may direct-ly or indirectly influence the focal area of the study. As such, the seating position

considerably influences the steering feel-ing that is generated within the driver, al-though this aspect is difficult to influence by the chassis developer.

As the matrix level rose during the course of the interview, the customer’s pre-vious responses were questioned and ana-lyzed and thus became increasingly de-tailed. At this point, the value of the ap-proach of using chassis specialists, who are native speakers, to perform the survey was confirmed. This proved to be a prerequisite for capturing the nuances (“reading be-tween the lines”) of the customer’s state-ments, in other words recognizing the cus-tomer’s true intentions and evaluating and allocating the respective driving situation and vehicle response.

The Table shows a section of the three-level matrix. On the left-hand side of the Table the customer wishes expressed dur-ing the three, ascending interview stages and their weighting (Imp = importance) is shown. The weighting scheme corresponds to a statistical spread of the total naming of items standardized to a column total of 100. On the right-hand side of the Table the fulfillment of the listed wishes by the vari-ous vehicles has been documented. Large numbers represent good levels of customer satisfaction. A good overview of these sur-vey results is provided by the market op-portunity map (MOM). This map groups all customer wishes as well as the rating awarded to the various vehicles within the four quadrants of the axial cross, consist-ing of “Satisfaction” and “Weighting.” Figure 2 depicts a respective MOM. The num-bers in the symbols represent customer wishes. In our example the “13” stands for the wish for “precise and direct steering”.

The MOM is suitable for use as a “quick to grasp” presentation to establish customer targets following the analysis of the surveys.

Table: Three-level matrix with section of results of the survey

DEVELOPMENT

ATZ 11I2007 Volume 10916

Chassis

Page 3: Optimization of steering behavior

The various quadrants contain the customer wishes with different priorities for optimi-zation. The most important quadrant, no doubt, is the one on the upper left-hand side (1. Action necessary). This is where customer wishes have been identified as being impor-tant, but are not being met sufficiently by the reference car. The items, which tend to be less important in the second quadrant (2. Action useful), which might put the finish-ing touches on the product from the cus-tomer’s perspective, are contrasted by the third quadrant (3. Cost reduction possible). This quadrant shows customer wishes, which tend to be less important to the cus-tomer but are being met excessively by the reference car – an indicator for possible “over engineering.” In addition, the exam-ple shows the customer targets, which have been entered using the symbols of the target car on the right-hand side. They mark the objective of the optimization.

The survey result shown in Figure 2 is a component of the horizontal frame in the HoQ as well. Figure 3 shows the survey result, which in addition, for a better spread in col-umns 8 to 13, is multiplied with the custom-er’s weighting. To ensure that the informa-tional content of this presentation remains equivalent to the MOM, the importance at-tributed by the customer is visualized in col-umn 15, and the customer target recorded in column 16. This means that all relevant information has been captured in a trans-parent format. The most important custom-er wish by far, in terms of steering behavior, can be seen clearly in line 5 in this format as well: “precise and direct steering”. In col-umn 16 (actions) is noted: “Yes, action planned; objective: ‘best in class’”.

4 Linking Customer Emotions to Technical Specifications

The objective is to link the emotions and wishes of the customer’s to the Technical Specifications. For this purpose, the inter-secting area of the horizontal customer frame is processed with the vertical engi-neer’s frame. The conventional QFD method exclusively uses directly quantifiable values [5]. The complex subject of handling and the steering feeling, however, are described by means of a large number of individual pa-rameters [3, 4] and, considering the custom-er’s very emotional and “compound” wishes cannot be addressed by the conventional QFD method. The significance of the identi-fied point, “precise and direct steering”, is very complex in its meaning due to its very emotional component.

The evolution of the conventional QFD method into “Improve by QFD” was there-fore the next consistent and logical step. To perform the links within the matrix the col-umns were filled with supra-ordinate terms for entire groups of technical parameters. A crucial additional specialty is the inclusion of evaluation criteria from the subjective handling evaluation [2]. A robust method to form the subjective judgment of the test drivers was used in order to ensure the re-producibility of the evaluation. If the cus-tomer’s wish has been worded, “I want a smaller steering wheel,” this wish can be linked directly with the technical feature “steering wheel diameter”. For “precise and direct steering” the QFD evolution “Improve by QFD” first performs a translation into the subjective engineer’s language. Thus de-emotionalizes the wish, before it is translat-ed into objective parameters and specifica-tion values in further translation steps.

When linking emotions and specifica-tions, Figure 4, for each intersecting point, a question worded something like this: “How strong is the influence of the technical specification” (column 5, steering response) on the customer’s wish of “precise and di-rect steering” (line 5)? The level of influ-ence is expressed on the following scale:– 0 = no influence– 1 = low – 3 = medium– 9 = strong.It becomes obvious quickly, that an increas-ing number of technical specifications re-sults in a disproportionate increase in the number of linking questions required. This justifies the approach of using supra-ordi-nate terms.

After the linking questions, a character-istic value is identified for each technical specification, called “Calculation Relation-ship” in this case. For each column, the linked values are multiplied by the impor-tance of the customer wish, and a total for the column is created. The higher the value of a column, the more directly a change to this particular technical specification can serve to influence the sum total of all cus-tomer wishes. This procedure enables an optimization of costs and efforts to be achieved. If the technical specifications are sorted by the points score, this will result in a ranking, as shown by example in Figure 5. These rankings are created in ac-cordance with each translation step – from the customer’s wish to the subjective engi-neer’s evaluation, from the subjective engineer’s evaluation to vehicle perform-ance properties and from vehicle per-formance properties to design parameters.

Now all the information is available, which is required to answer the question: “What technical criteria influence customer wishes in the most positive way?” Highly in-teresting at this point were the differences in the ranking after the translation from the customers’ language into the engineers’ lan-guage compared with the media language into the customers’ language. In this case the top 4 engineers’ criteria in terms of naming, and even sequence, were identical, Figure 6.

The evaluation of the correlations of the various translation steps within the engi-neer’s language is done in the “Customer Criteria Analysis” matrix (CCA matrix). The CCA matrix shows those terms in the engi-neer’s language, which positively influence customer’s wishes at the highest level of ef-ficiency and the technical parameters from the individual functional groups, which – in turn – have the major influence on them. Now the most important customer wish, “precise and direct steering”, has been effectively linked with the technical criteria identified from the individual functional groups (tire, brake, steering, K&C, vehicle performance properties, etc.).

5 Technical Benchmark – First Heat

Using the correlations in Figure 4 and Figure 5, it is now possible to derive changes from this data for the technical specifications with the strongest influences. As a result of work-ing with supra-ordinate terms such as “tire characteristics” in the specifications, a pro-gram for a technical comparison (bench-mark) was put together, which enabled an objective description of the parameters rele-vant for the steering behavior. This selection of technical parameters of the comparison vehicles was determined in order to obtain a complete picture for the subsequent defini-tion of the parameters in the action plan.

It was evident, though, that the conven-tional methods for objective evaluation of the most important criteria for steering be-havior had to be insufficient. For this pur-pose, an expanded standard for objective measurement was developed. Aside from standard tests to determine the vehicle’s transfer performance, kinematics and elas-to-kinematics as well as tire parameters, tests to assess important design factors such as ditch effect and steering system hysteresis became a component of the test matrix.

Using the results of the benchmarking, it was now possible to fill the development stage matrix, which is an important tool in the “Improve by QFD” process. This matrix is depicted in Figure 7. In the horizontal lines it

ATZ 11I2007 Volume 109 17

DEVELOPMENTChassis

Page 4: Optimization of steering behavior

contains all technical criteria sorted by the individual functional groups. The vertical columns reflect the development process from the initial tests (benchmarking) on the left-hand side to the first prototype at the far right-hand side. Colored indicators show the degree of compliance of the respective technical criterion from one development step to the other and whether a criterion qualifies for very effective, effective or inef-fective modification.

6 What Criteria Must be Improved? Action Plan

The action plan, as well, is tied to the matrix in Figure 7. It specifies target values and cor-ridors for the individual technical specifica-tions which have been identified. This prepa-ration of a specifications book kicked off the development of the prototype which then had to pass the test of the customer’s critical eye. Prior to that, however, there was another step to be completed: The identification of the items, which truly required improve-ment. The “Improve by QFD” method initially reflects the most critical and most relevant technical specifications, which have the most significant influence on customer satisfac-tion. At the same time, the objective bench-mark shows the objective values of many pa-rameters. Now it is time to identify those pa-rameters, which a) still leave major room for improvement and b) do not degrade custom-er wishes, which have already been met, as the result of a negative interaction. The project team discussed these issues and de-cided what functional groups need to be worked on to achieve the objectives.

The content of the action plan, for one, consisted of an assessment of the target values and corridors to be complied with regarding the technical parameters of the supra-ordinate terms, shown in Figure 5. For the other, the requisite investment of time and money was assessed. This resulted in an action plan, which provided the basis for preparing a spec-ifications book pertaining to this optimiza-tion of steering performance. The target val-ues, as well, are listed in the matrix.

7 Development of the Protoype

The following functional groups to be modi-fied were identified as being highly effective:– seat– front axle kinematics– steering kinematics– steering ratio– spring/damper adjustment.

The kinematics changes identified in the QFD process were developed using the Ad-ams/Car simulation tool, a decision support-ed by the desire to make efficient use of the project time. In addition to the option of changing the steering ratio, the modifica-tions made to the adjustable front axle shown in Figure 8, include an adjustment op-tion of the instantaneous center of rotation and the steering kinematics (toe-in vs. spring travel). Furthermore, the project team – in accordance with the output of the QFD proc-ess – changed the suspension set-up (spring/damper) and the roll stiffness.

Items, which were very important to the customer, such as the ride comfort, were not impaired by the various chassis-related meas-ures. Focusing on the specifications, which according to the QFD process show the high-est level of relevance, was responsible for the high level of efficiency of the measures taken, as the final customer test ride was to show.

8 Technical Benchmark – Second Heat

In order to be able to assess the change of the technical specifications and to enable a comparison with the target values, a sec-ond technical benchmark had to be per-formed. Figure 9 by means of an example in the abscissa shows the most important items identified in the engineer’s language sorted by their importance and, in the ordi-nate, the degree of compliance achieved by the individual vehicles. Criteria of particu-lar importance, in terms of customer wish-es, should have a high degree of compli-ance, and thus a good performance from the customer’s perspective. The improve-ment in these relevant items, documented in Figure 9, suggests that there is a good starting basis for the customer test ride.

Additionally high performance, particu-larly in the important items, is shown by Figure 9. At least the engineers are now con-vinced of the project’s success. Yet the final success regarding improved and more attrac-tive/delighting steering behavior can only be confirmed by the customer him- or herself. The end customer should have the final word in a concluding customer test ride and con-firm the value of the process and the project.

9 Verification in the Final Customer Test Ride

Whether or not a significant improvement will actually be perceived by the customer, is a question, which, ultimately, can only be confirmed by the critical and unadulter-

ated view of the customer in a final cus-tomer test ride. The group of interviewers as well as the methodology, interviewing strategy and the type of questions and ques-tionnaires used are identical to those of the customer test ride, conducted at the begin-ning of the project. This ensures a maxi-mum comparability.

The vehicle matrix consists of the origi-nal OEM vehicle and the designated bench-mark vehicle of a competitor and, of course, the “Improve by QFD” prototype, which, if not earlier, now has to pass the test of the critical eye of the consumer. This determines the success or failure of the project. To back the results by statistical means, a minimum of 25 participants in the designated custom-er segment has to look at the vehicles and, of course, drive them extensively. The cus-tomer wishes of the final customer test ride, as well, are displayed in the three-level ma-trix during the QFD process, and the ques-tionnaires evaluated accordingly.

How are the items, which are important to the customer, complied with by the three comparison vehicles? The two MOMs of Figure 10 provide information about the de-gree of compliance with the individual cus-tomer wishes for “Improve by QFD” proto-type and original OEM vehicle. Criterion 1, for example, is the customer wish, “precise and direct steering”. The direct comparison shows, that in terms of important customer criteria the “Improve by QFD” prototype, Fig-ure 10 top, has seen a crucial improvement compared with the original OEM vehicle, Figure 10 bottom. The comparison with the designated benchmark vehicle analogue to Figure 8, as well, shows the success of the project.

References[1] Pischetsrieder, B.: Beim Wort genommen. VDI-Nach-

richten, 10. Dezember 2004, S. 25

[2] Heißing, B.; Brandl, H. J.: Subjektive Beurteilung des

Fahrverhaltens. Vogel-Buchverlag, Würzburg 2002

[3] Matschinsky, W.: Radführungen der Straßenfahrzeuge

– Kinematik, Elasto-Kinematik und Konstruktion. 2.

Aufl., Springer-Verlag, Berlin 1998

[4] Reimpell, J.; Betzler, J.: Fahrwerktechnik. Grundlagen.

5. Aufl., Vogel-Buchverlag, Würzburg 2005

[5] Klein, B.: QFD – Quality Function Deployment. Expert-

Verlag, Renningen-Malmsheim 1999

[6] Becker, K.; et al.: Subjektive Fahreindrücke sichtbar

machen. Expert-Verlag, Renningen-Malmsheim 2000

[7] Becker, K.; et al: Subjektive Fahreindrücke sichtbar

machen II. Expert-Verlag, Renningen-Malmsheim 2002

[8] Yamamoto, M.; Schick, B.; Kushiro, I.; Resch, S.; Hagi-

wara, N.: Optimisation of Steering Behaviour by Sys-

tematic Implementation of Customer Requirements in

Technical Targets on the Basis of Quality Function De-

ployment. In: Proceedings of Fisita Conference 2006,

Yokahama, Japan 2006

DEVELOPMENT

ATZ 11I2007 Volume 10918

Chassis