An Ergonomic Study on the Operation Method and In-Vehicle Location of an Automotive Electronic Gearshift

Abstract

This study conducted a survey to identify the best ergonomic operation method, in-vehicle location, and the effects of their combination on electronic gearshifts. A total of 15 different design alternatives were derived through combinations of three operation methods (lever slide, button push, and dial rotation) and five in-vehicle locations (left wheel spoke, right wheel spoke, upper center fascia, lower center fascia, and center console). A total of 40 respondents with diverse ages and driving experiences evaluated the 15 different design alternatives across nine ergonomic evaluation measures (accuracy, efficiency, rapidity, learnability, intuitiveness, safety, preference, memorability, and satisfaction). The study results indicated that: (1) the lever slide and button push were superior to dial rotation for the operation method; (2) the lower center facia and center console were superior for the in-vehicle location, and (3) implementing the lever slide method in the center console location was found to lead to the best combination of the operation method and in-vehicle location, while implementing the button push method in the right wheel spoke or upper center fascia location also showed relative superiority. The study findings are expected to contribute to the ergonomic design of electronic gearshifts that can enhance the driver’s gear-shifting experience, thereby improving driving performance and safety.

1. Introduction

Electronic transmission systems based on shift-by-wire technology have been developed and widely implemented to meet the market demands for lightweight vehicles and spacious interior spaces [1,2,3]. One of the significant advantages of electronic transmission systems is their flexibility in gearshift design, allowing for a wide range of designs in terms of operation methods, in-vehicle locations, appearance, size, and more [3,4,5]. This design flexibility can be utilized to develop systems that enhance the driver’s gear-shifting experience, thereby improving driving performance and safety.

Meanwhile, among the various design variables for electronic gearshifts, the two most important and representative ones are their operation method and in-vehicle location. Operation methods can include lever slide, button push, and dial rotation, while in-vehicle locations can be near the steering wheel, on the center facia’s upper/middle/lower part, or at the center console. Currently, various automotive manufacturers and their production car models have adopted and applied various combinations of operation methods and in-vehicle locations, as shown in

Table 1. Operation methods and in-vehicle locations of electronic gearshifts adopted by various automotive manufacturers and their production car models.

	Operation Method	Lever Slide	Button Push	Dial Rotation
In-Vehicle Location
Near the steering wheel			Not adopted yet
Manufacturer (Model)		PORSCHE (TAYCAN and PANAMERA [2024~])		BMW (I3), CHEVROLET (ID4), HYUNDAI (CONA [2022~], GRANDEUR 7, IONIQ 5, IONIQ 6, SONATA [2024], and SANTAFE [2024]), and KIA (EV9)
Center facia upper/middle part
Manufacturer (Model)		KIA (LAY), LAND ROVER (DEFENDER 130), NISSAN (QUEST), and TOYOTA (SIENNA)	CHEVROLET (SUBURBAN and TAHOE), GMC (YUKON), HONDA (ACCORD and ODYSSEY), and LINCOLN (CONTINENTAL, MKC, MKZ, and MKX)	CHRYSLER (200, 300, PACIFICA, and VOYAGER) and RAM (2500, 3500, 1500, and CHASSIS CAB)
Center facia lower part				Not adopted yet
Manufacturer (Model)		GENESIS (G90 [2021]), HONDA (ACCORD [~2018], CR-V, PILOT [~2015], and ODYSSEY [~2017]), HYUNDAI (CASPER, STAREX, and STARIA), and KIA (CARNIVAL [~2014])	GMC (ACADIA and TERRAIN), HONDA (ODYSSEY), HYUNDAI (SANTAFE, GRANDEUR, and STARIA), LINCOLN (AVIATOR, CORSAIR, NAUTILUS, and NAVIGATOR), and JAGUAR (I-PACE)
Center console
Manufacturer (Model)		CHEVROLET (SPARK, MALIBU, TRAX, and TRAILBLAZER), DODGE (CHALLENGER, DURANGO, and CHARGER), FORD (MUSTANG), GENESIS (G70), HYUNDAI (AVANTE, VELOSTER, and KONA N), JAGUAR (F-TYPE, F-PACE, E-PACE, and X-TYPE), JEEP (CHEROKEE, COMPASS, and WRANGLER), KIA (K3, K9, SPORTAGE [~2014], STINGER MEISTER, and CARNIVAL), LAND ROVER (DEFENDER, DISCOVERY, NEW RANGE ROVER, EVOQUE, and VELAR), LINCOLN (MKT, MKS, and NAVIGATOR [~2017]), and PORCHE (CAYENNE, TURBO S, and 911)	CHEVROLET (CORVETTE C8 STINGRAY and VOLT), HONDA (ACURA NSX, PILOT, and TLX), and HYUNDAI (AZERA, TUCSON, PALISADE, SONATA, CONA [~2021], NEXO, and TUCSON)	GENESIS (G80, G90, GV70, GV60, and GV80), DODGE (DURANGO [2014~2017]), FORD (EXPLORER and EXPEDITION), JAGUAR (XF, XJ, XK, and XE), JEEP (GRAND CHEROKEE, WAGONEER, and GRAND WAGONEER), KIA (K5, K8, SELTOS, SPORTAGE, CARNIVAL, NIRO EV6, and SORENTO), and LAND ROVER (LR4, RANGE ROVER, EVOQUE [~2019], and VELAR [~2020])

.

However, despite their design flexibility and diversity, poorly designed electronic gearshifts have the potential to cause driver confusion and errors when shifting gears [6]. Indeed, safety issues and accidents have occurred due to this, as was the case in the United States in 2015 when a train and a car collided at a station near Valhalla—the unfamiliar operation method of the electronic gearshift was identified as one of the causes of the accident [7]. Furthermore, it is important to note that drivers have frequently expressed complaints about the inconvenience and risks associated with using electronic gearshifts.

Relatedly, a few previous studies have developed various gearshift HMI design alternatives and conducted comparative evaluations. Andersson and Lenshof [8] conducted a study aiming to design an intuitive interface for electronic gearshifts. They generated promising gearshift design alternatives using various ideation methodologies and subsequently compared them based on different criteria, including accessibility, uniqueness, ergonomics, intuitive interaction, shift pattern understandability, interactive display, and feedback. As a result, a joystick-type gearshift with a touchscreen mounted on the top surface, located in the center console, was found to be the best alternative; this was partly explained by its similarity to traditional gearshifts in terms of the operation method and in-vehicle location. Bladfält et al. [4] conducted a comparative evaluation of four types of gearshifts (button shifter, polystable rotary shifter, joystick shifter, and monostable stalk shifter) in a scenario including driving between cones through a simulation experiment. The results showed that while the button shifter and polystable rotary shifter were found to be intuitive, with lower error rates compared to the other two types, the participants most preferred the joystick shifter. Bladfält et al. [9] conducted a simulation experiment in a scenario involving gear shifting to comparatively evaluate a total of 12 joystick shifter types derived by combining four shift patterns (I- or J-shaped monostable shift pattern, toggle functioned monostable shift pattern, J-shaped polystable shift pattern, and I-shaped polystable shift pattern) and three complexity levels based on the number of gear positions to be selected (three {RND}, four {RNDM}, and five {PRNDM]}. The results indicated that the I or J-shaped monostable shift pattern with PRNDM shift sequence showed the longest task completion time and the lowest subjective ease-of-use, which was explained by the lack of visibility of the gear shifter state and haptic feedback. Choi et al. [10] derived a total of 12 shift button layouts for button-type gearshifts based on four ergonomic layout design principles (consistency, functional similarity, compatibility, and stereotype) and conducted a survey to evaluate them in terms of seven ergonomic evaluation measures (accuracy, convenience, rapidity, learnability, intuitiveness, safety, and preference). As a result, a layout, to which the principles of consistency, functional similarity, and stereotype were applied together, was recommended as the best (most ergonomic) layout design.

Despite these previous research efforts, however, there still exists a notable gap in our understanding of how the operation method, in-vehicle location, and their combinations for electronic gearshifts impact driver performance, comfort, and safety. Given that various combinations of operation methods and in-vehicle locations have been applied across various manufacturers/models (Table 1), such a knowledge gap poses a significant obstacle to preventing the abovementioned safety issues and accidents due to driver confusion and errors.

As a response to this gap, the primary objective of this study is to identify an ergonomic design for electronic gearshifts that can prevent driver confusion and errors, ultimately enhancing driver safety and comfort. In an effort to achieve this goal, the current study focuses on addressing the following three unexplored research questions:

Research Question 1: What is the best ergonomic operation method for electronic gearshifts?

Research Question 2: What is the best ergonomic in-vehicle location for electronic gearshifts?

Research Question 3: What is the best ergonomic combination of operation method and in-vehicle location for electronic gearshifts?

To address the three research questions, this study considered three operation methods (lever slide, button push, and dial rotation) and five in-vehicle locations (left wheel spoke, right wheel spoke, upper center fascia, lower center fascia, and center console), resulting in a total of 15 design alternatives based on their combinations. A total of 40 respondents, varying in age and driving experience, assessed each design alternative across nine ergonomic evaluation measures (accuracy, efficiency, rapidity, learnability, intuitiveness, safety, preference, memorability, and satisfaction). The collected data were analyzed to conduct a comparative evaluation, allowing for the identification of an ideal ergonomic operation method, in-vehicle location, and their ideal combination for electronic gearshifts, along with some ergonomic design characteristics that should be considered in gearshift design.

2. Materials and Methods

2.1. Respondents

This study aimed to recruit respondents to encompass diverse ages and driving experiences to ensure that the study results are not confined to a specific driver group but are applicable across various driver groups. To achieve this, the study recruited respondents from two age groups and two driving experience groups, applying the following criteria: first, for age groups, ‘young adulthood’ was defined as individuals aged between 19 and 34, while ‘middle-aged’ were individuals aged between 35 and 70 [11,12]. Note that, considering the inherent nature and challenges of online surveys, individuals aged 70 and above were excluded from this study. Additionally, referring to some previous studies that classified/characterized drivers according to their driving experience [13,14,15], respondents’ driving experience was categorized according to their total driving distance in the last five years, with those driving less than 5000 km defined as ‘novice drivers’, and those with more than 10,000 km as ‘experienced drivers’. A total of 40 respondents participated in the survey, and their age and driving experience are summarized in

Table 2. Characteristics of respondents.

Group			Age (Years)			Driving Experience (Driving Distance in the Last Five Years {km})
Experience	Age	N	Mean	SD	Range	Mean	SD	Range
Novice	Young adulthood	15	23.3	2.40	19–26	731	1372	0–5000
	Middle-aged	6	47.8	11.2	37–63	2500	2739	0–5000
	Total	21	30.3	13.1	19–63	1237	1965	0–5000
Experienced	Young adulthood	9	26.9	4.20	21–34	65,889	49,994	13,000–150,000
	Middle-aged	10	40.6	4.38	36–49	73,500	47,458	25,000–160,000
	Total	19	34.1	8.18	21–49	69,895	47,457	13,000–160,000
Total		40	32.1	11.1	19–63	33,849	74,404	0–160,000

. Prior to participation, each respondent provided informed consent.

2.2. Design Alternatives

2.2.1. Operation Method

This study considered three typical operation methods adopted by various automotive manufacturers and their production car models: the lever slide (referred to as M_LS), button push (referred to as M_BP), and dial rotation (referred to as M_DR), as shown in

Table 3. The three operation methods considered in the current study.

Lever Slide (MLS)	Button Push (MBP)	Dial Rotation (MDR)

.

2.2.2. In-Vehicle Location

This study considered three in-vehicle locations commonly adopted in various currently available production cars: the upper center fascia (referred to as L_UCF), lower center fascia (referred to as L_LCF), and center console (referred to as L_CC). Additionally, two-wheel spoke locations, left wheel spoke (referred to as L_LWS) and right wheel spoke (referred to as L_RWS), were included based on their potential ergonomic benefits, specifically in regards to reducing the eye-off-the-road time and, thereby, the risk of accidents [16,17,18,19,20,21]. The five locations considered in this study (i.e., L_LWS, L_RWS, L_UCF, L_LCF, and L_CC) can be distinguished by the hand travel distance from the steering wheel, following the recommendation in SAE J1138 that higher priority hand controls, such as the gearshift, be within reach of a driver wearing a lap and shoulder restraint.

Table 4. The five in-vehicle locations (shaded in blue) considered in the current study.

Left Wheel Spoke (LLWS)	Right Wheel Spoke (LRWS)	Upper Center Fascia (LUCF)	Lower Center Fascia (LLCF)	Center Console (LCC)

illustrates the five in-vehicle locations considered in this study.

2.2.3. Design Alternatives

A total of 15 design alternatives were derived by combining the three operation methods (M_LS, M_BP, and M_DR) and five in-vehicle locations (L_LWS, L_RWS, L_UCF, L_LCF, and L_CC), as illustrated in

Table 5. The 15 design alternatives derived in this study.

	Operation Method	Lever Slide (MLS)	Button Push (MBP)	Dial Rotation (MDR)
In-Vehicle Location
Left wheel spoke (LLWS)
Right wheel spoke (LRWS)
Upper center fascia (LUCF)
Lower center fascia (LLCF)
Center console (LCC)

.

Among the 15 design alternatives, only 8 (M_BPL_UCF, M_DRL_UCF, M_LSL_LCF, M_BPL_LCF, M_DRL_LCF, M_LSL_CC, M_BPL_CC, and M_DRL_CC) have been implemented in production car models (Table 1). Relatedly, during the initial phases of developing new products, the absence of functional prototypes often hinders the assessment of user experience through tangible interactions with the product [22,23]. During such early stages of the design process, the anticipated user experience (AUX) assessment method, wherein participants envision the product design concept and anticipate future experiences, has proven to be a viable and practical approach, before actual product experiences occur [22,23,24,25]. Similarly, ISO 9241-210 [26] also defined user experience as a “person’s perceptions and responses resulting from the use and/or anticipated use of a product, system or service”. The AUX assessment is commonly conducted through methods such as freehand design sessions or survey/interview techniques. The efficacy of AUX assessments in gaining insights into user needs and their practicability and association with actual experience-based evaluations has been previously demonstrated and discussed [22,23,24,25,26,27]. Therefore, the current study also employed the AUX assessment method, providing respondents with illustrated images of each of the 15 design alternatives (Table 5) and instructing them to evaluate each design based on their imagination of future interactions and experiences with it.

2.3. Questionnaires

In this study, to perform a comparative ergonomic evaluation of the 15 design alternatives derived above, nine ergonomic evaluation measures were employed which were: accuracy, efficiency, rapidity, learnability, intuitiveness, safety, preference, memorability, and satisfaction. These measures were adopted with reference to some key ergonomic design considerations/principles for in-vehicle displays/controls [8,28,29,30,31], along with the subjective evaluation measures employed by Choi et al. [8] in their evaluation of the ergonomic performance of 12 different layouts of automotive electronic shift buttons. Each of the evaluation measures and its description (question) given to the respondents is listed in

Table 6. The nine ergonomic evaluation measures and their corresponding descriptions (questions).

Measures	Description (Question)
Accuracy	This design will help me shift gears accurately without errors.
Efficiency	This design will help me shift gears efficiently in terms of eye and hand movements.
Rapidity	This design will help me shift gears quickly.
Learnability	This design can be learned without much effort.
Intuitiveness	This design meets my expectations.
Safety	This design will help me drive safely.
Preference	I like this design (I am willing to use/buy it).
Memorability	This design will be easy to use again even after a period of not using it.
Satisfaction	I am generally satisfied with this design (this design is excellent).

.

For each of the 15 design alternatives, the respondents answered the questions for each of the nine measures on a 7-point Likert scale [32,33] ranging from “Strongly disagree” (1) to “Strongly agree” (7), with the midpoint being “Neutral” (4).

2.4. Freehand Design Session

Following the completion of the aforementioned subjective evaluations, a freehand design session was conducted to explore the respondents’ design preferences beyond the 15 presented/evaluated design alternatives. The respondents were presented with an image of a vehicle interior without a gearshift (Figure 1) and instructed to choose their preferred operation method from the three options (M_LS, M_BP, and M_DR) and subsequently place it freely at their preferred location within the vehicle. The final picture created by each respondent was collected.

2.5. Survey Procedure

The survey was conducted on an online platform (Survey Monkey). This online survey spanned approximately two weeks, allowing the respondents to engage in the survey at their preferred time and location.

The procedural details of the survey are outlined as follows:

• Before the survey, an introduction was provided, explaining the research background and objectives, along with the survey questionnaire structure.
• After obtaining informed consent from all respondents, detailed explanations were provided for each of the 15 design alternatives, the nine ergonomic evaluation measures, and the associated rating scale.
• Respondents evaluated the 15 design alternatives across the nine measures, with the presentation order of the alternatives randomized for each respondent to minimize potential order effects.
• Following the evaluation of all design alternatives, a freehand design session was conducted.

2.6. Data Analyses

A two-way repeated measures ANOVA was conducted to test the effect of operation method (three levels: M_LS, M_BP, and M_DR) and in-vehicle location (five levels: L_LWS, L_RWS, L_UCF, L_LCF, and L_CC) on each of the nine evaluation measures. The main and interaction effects of the independent variables on the dependent measures (accuracy, efficiency, rapidity, learnability, intuitiveness, safety, preference, memorability, and satisfaction) were examined. In cases of a significant interaction effect between the independent variables, the simple main effects for each independent variable were investigated. Mauchly’s test was used to test whether the data met the assumption of sphericity. In cases where sphericity was violated, the degrees of freedom were corrected; the Greenhouse–Geisser correction was used when the Greenhouse–Geisser estimate of sphericity (ɛ) was less than 0.75; else, the Huynh–Feldt correction was used [34]. In cases where an ANOVA identified a significant effect, post hoc multiple comparisons with Bonferroni corrections were conducted to determine the pairs with significant mean differences. The statistical analyses were conducted at an alpha level of 0.05 using IBM SPSS Statistics 26.0.

Regarding the responses collected during the freehand design session (i.e., the pictures created by the respondents), an affinity diagram analysis was carried out on the group pictures with similar design characteristics. The frequency associated with each group, indicating the number of respondents who drew similar pictures, was recorded.

3. Results

3.1. Overall Scores for the 15 Design Alternatives

Table 7. The mean and standard deviation values for each dependent measure across the 15 design alternatives.

	Measure	Accuracy	Efficiency	Rapidity	Learnability	Intuitiveness	Safety	Preference	Memorability	Satisfaction
Design
MLSLLWS		2.65 (1.82)	3.38 (2.07)	3.56 (1.99)	3.30 (1.94)	2.68 (1.67)	2.40 (1.55)	2.45 (1.80)	3.63 (2.03)	2.50 (1.69)
MLSLRWS		3.08 (1.70)	4.00 (2.08)	4.45 (1.99)	3.75 (1.92)	3.28 (1.74)	3.00 (1.65)	2.98 (1.91)	4.05 (1.83)	3.03 (1.83)
MLSLUCF		4.18 (1.85)	3.40 (1.85)	3.58 (1.62)	4.25 (1.84)	3.75 (1.68)	3.43 (1.62)	2.90 (1.55)	4.30 (1.70)	2.90 (1.55)
MLSLLCF		4.90 (1.52)	4.50 (1.59)	4.48 (1.58)	5.03 (1.72)	4.68 (1.56)	4.80 (1.34)	4.35 (1.64)	5.13 (1.51)	4.33 (1.46)
MLSLCC		5.60 (1.26)	4.90 (1.60)	5.13 (1.36)	5.88 (1.28)	5.55 (1.54)	5.18 (1.47)	5.08 (1.58)	5.70 (1.20)	5.03 (1.54)
MBPLLWS		2.80 (1.67)	3.73 (1.97)	4.10 (1.98)	3.60 (1.91)	2.93 (1.69)	2.73 (1.63)	2.60 (1.68)	3.70 (1.83)	2.73 (1.65)
MBPLRWS		3.53 (1.69)	4.58 (1.99)	4.50 (1.89)	4.20 (1.77)	3.68 (1.85)	3.33 (1.76)	3.50 (1.85)	4.33 (1.83)	3.75 (1.88)
MBPLUCF		4.15 (1.67)	4.23 (1.69)	4.00 (1.69)	4.50 (1.71)	3.93 (1.54)	3.90 (1.60)	3.40 (1.65)	4.55 (1.54)	3.43 (1.55)
MBPLLCF		4.48 (1.48)	4.55 (1.52)	4.45 (1.41)	4.98 (1.31)	4.58 (1.34)	4.53 (1.30)	4.38 (1.58)	5.00 (1.15)	4.35 (1.41)
MBPLCC		4.83 (1.43)	4.33 (1.58)	4.55 (1.48)	4.90 (1.57)	4.90 (1.53)	4.70 (1.44)	4.50 (1.83)	5.03 (1.44)	4.48 (1.74)
MDRLLWS		2.73 (1.68)	3.08 (1.95)	3.50 (1.93)	3.28 (1.95)	2.65 (1.72)	2.43 (1.58)	2.40 (1.75)	3.50 (1.85)	2.40 (1.72)
MDRLRWS		2.80 (1.49)	3.78 (1.98)	3.65 (1.90)	3.33 (1.89)	2.80 (1.70)	2.70 (1.57)	2.65 (1.75)	3.45 (1.87)	2.60 (1.58)
MDRLUCF		3.60 (1.78)	3.93 (1.90)	3.83 (1.71)	4.00 (1.78)	3.28 (1.69)	3.38 (1.75)	3.13 (1.76)	4.08 (1.67)	3.18 (1.78)
MDRLLCF		4.25 (1.58)	4.10 (1.60)	4.18 (1.47)	4.30 (1.67)	4.05 (1.45)	4.03 (1.49)	4.00 (1.48)	4.60 (1.52)	3.85 (1.44)
MDRLCC		4.20 (1.62)	4.28 (1.71)	4.33 (1.58)	4.53 (1.57)	4.33 (1.69)	4.08 (1.70)	4.05 (1.88)	4.63 (1.51)	4.15 (1.67)

provides the mean and standard deviation values for each of the nine dependent measures for each of the 15 design alternatives.

3.2. Main Effects of Operation Method

The ANOVA results revealed that the main effects of the operation method were statistically significant for all measures except for preference: accuracy F(2, 78) = 6.44, p < 0.01; efficiency F(2, 78) = 5.27, p < 0.01; rapidity F(2, 78) = 3.20, p < 0.05; learnability F(2, 78) = 6.36, p < 0.01; intuitiveness F(2, 78) = 7.13, p < 0.01; safety F(2, 78) = 6.38, p < 0.01; preference F(2, 78) = 2.97, p = 0.06; memorability F(2, 78) = 6.36, p < 0.01; and satisfaction F(2, 78) = 4.34, p < 0.05. The mean and standard deviation values of each operation method for each of the dependent measures are shown in Figure 2, with asterisks indicating the statistical significance of the post hoc comparisons.

The statistically significant differences between the operation methods are as follows:

• M_LS showed significantly higher mean scores than M_DR in accuracy, learnability, intuitiveness, safety, and memorability.
• M_BP exhibited significantly higher mean scores than M_DR in efficiency, learnability, intuitiveness, safety, memorability, and satisfaction.

3.3. Main Effects of In-Vehicle Location

The ANOVA results revealed that the main effects of in-vehicle location were statistically significant for all nine measures: accuracy F(2.66, 104) = 26.2, p < 0.001; efficiency F(2.83, 110) = 4.43, p < 0.01; rapidity F(2.88, 112) = 4.19, p < 0.01; learnability F(3.14, 122) = 18.1, p < 0.001; intuitiveness F(3.08, 120) = 23.5, p < 0.001; safety F(2.62, 102) = 25.5, p < 0.001; preference F(2.71, 106) = 22.4, p < 0.001; memorability F(2.77, 108) = 15.9, p < 0.001; and satisfaction F(2.78, 108) = 24.1, p < 0.001. The mean and standard deviation values of each in-vehicle location for each dependent measure are shown in Figure 3, with asterisks indicating the statistical significance of the post hoc comparisons.

The statistically significant differences between the in-vehicle locations are as follows:

• L_RWS showed significantly higher mean scores than L_LWS in efficiency and satisfaction.
• L_UCF exhibited significantly higher mean scores than (1) L_LWS in accuracy, learnability, intuitiveness, safety, and memorability; and (2) L_RWS in accuracy.
• L_LCF had significantly higher mean scores than (1) L_LWS across all measures except rapidity; (2) L_RWS across all measures except efficiency and rapidity; and (3) L_UCF in accuracy, intuitiveness, safety, preference, memorability, and satisfaction.
• L_CC obtained significantly higher mean scores than (1) L_LWS across all nine measures; (2) L_RWS across all measures except efficiency and rapidity; and (3) L_UCF across all measures except efficiency.

3.4. Interaction Effects between the Operation Method and In-Vehicle Location

The ANOVA results indicated that the interaction effects between the operation method and in-vehicle location were statistically significant for all nine measures: accuracy F(8, 312) = 3.46, p < 0.01; efficiency F(8, 312) = 2.65, p < 0.05; rapidity F(8, 312) = 2.51, p < 0.05; learnability F(8, 312) = 3.50, p < 0.01; intuitiveness F(8, 312) = 2.50, p < 0.05; safety F(8, 312) = 2.58, p < 0.05; preference F(8, 312) = 2.45, p < 0.05; memorability F(5.76, 224) = 2.53, p < 0.05; and satisfaction F(5.66, 220) = 3.05, p < 0.01. The mean and standard deviation values of each combination of operation method and in-vehicle location (i.e., design alternative) for each dependent measure are shown in Figure 4, with asterisks indicating the significant simple main effects.

The statistically significant simple main effects of the operation method are as follows:

• When the in-vehicle location was L_RWS,

  ▪

  M_LS showed significantly higher mean scores than M_DR in rapidity and memorability.

  ▪

  M_BP exhibited significantly higher mean scores than (1) M_LS in satisfaction and (2) M_DR across all nine measures.

• When the in-vehicle location was L_UCF,

  ▪

  M_BP showed significantly higher mean scores than M_LS in efficiency.

• When the in-vehicle location was L_LCF,

  ▪

  M_LS showed significantly higher mean scores than M_DR in safety.

  ▪

  M_BP exhibited significantly higher mean scores than M_DR in learnability.

• When the in-vehicle location was L_CC,

  ▪

  M_LS showed significantly higher mean scores than (1) M_BP in accuracy, learnability, and memorability and (2) M_DR across all nine measures.

The statistically significant simple main effects for in-vehicle location are as follows:

• When the operation method was M_LS,

  ▪

  ▪ L_UCF showed significantly higher mean scores than L_LWS in accuracy, intuitiveness, and safety.

  ▪

  L_LCF exhibited significantly higher mean scores than (1) L_LWS and L_RWS across all measures except efficiency and rapidity and (2) L_UCF across all nine measures.

  ▪

  L_CC had significantly higher mean scores than (1) L_LWS and L_UCF across all nine measures; (2) L_RWS across all measures except efficiency and rapidity; and (3) L_LCF in learnability and intuitiveness.

• When the operation method was M_BP,

  ▪

  L_RWS showed significantly higher mean scores than L_LWS in intuitiveness, preference, and satisfaction.

  ▪

  L_UCF exhibited significantly higher mean scores than L_LWS in accuracy, intuitiveness, and safety.

  ▪

  L_LCF had significantly higher mean scores than (1) L_LWS across all measures except efficiency and rapidity; (2) L_RWS in accuracy and safety; and (3) L_UCF in preference and satisfaction.

  ▪

  L_CC obtained significantly higher mean scores than (1) L_LWS across all measures except rapidity and (2) L_RWS in accuracy, intuitiveness, and safety.

• When the operation method was M_DR,

  ▪

  L_UCF showed significantly higher mean scores than L_LWS in accuracy and safety.

  ▪

  L_LCF exhibited significantly higher mean scores than (1) L_LWS and L_RWS across all measures except efficiency and rapidity and (2) L_UCF in preference.

  ▪

  L_CC had significantly higher mean scores than (1) L_LWS across all measures except rapidity; (2) L_RWS across all measures except efficiency and rapidity; and (3) L_UCF in intuitiveness, preference, and satisfaction.

3.5. Freehand Design Session

A total of 43 pictures were collected during the freehand design session, with 37 of the 40 respondents providing one picture each and the remaining 3 respondents submitting two pictures each. These pictures were grouped into 15 distinct groups through affinity diagram analysis. It was observed that in 10 groups, which comprised 35 pictures, the gearshift was placed at the location considered in this study—in other words, these 35 pictures were the same as or similar to 10 out of the 15 design alternatives under study. In the remaining five groups, which included eight pictures, the gearshift was placed in different locations, specifically on the left or right side of the steering wheel.

Table 8. The frequency associated with each picture group collected during a freehand design session.

In-Vehicle Location Considered in This Study	Operation Method						Total
	MLS		MBP		MDR
LLWS	1		0		1		2
LRWS	1		2		0		3
LUCF	1		0		0		1
LLCF	8		6		0		14
LCC	9		3		3		15
Total	20		11		4		35
In-Vehicle Location ‘not’ Considered in This Study	Operation method						Total
	MLS		MBP		MDR
	Picture	N	Picture	N	Picture	N
Left side of the steering wheel		2		2		1	5
Right side of the steering wheel		1		2		0	4
Total	3		4		1		8

shows the frequency (number of pictures) associated with each of these 15 picture groups.

4. Discussion

In this study, 15 design alternatives were derived by combining three operation methods and five in-vehicle locations for electronic gearshifts. A total of 40 respondents, varying in age and driving experience, assessed each design alternative using a 7-point Likert scale across nine evaluation measures (accuracy, efficiency, rapidity, learnability, intuitiveness, safety, preference, memorability, and satisfaction). The major findings can be summarized as follows—note that Major findings A and B pertain to the main effects of the operation method and in-vehicle location, respectively, while Major finding C represents the results that did not surface in the main effects but emerged through the interaction effects between the two design variables.

• Major finding A. M_DR was evaluated as significantly inferior compared to M_LS and M_BP (Figure 2 and Figure 4).
• Major finding B-1. L_LWS and L_RWS were evaluated as significantly inferior compared to L_UCF, L_LCF, and L_CC. L_LWS was also evaluated as significantly inferior compared to L_RWS (Figure 3 and Figure 4).
• Major finding B-2. L_UCF was evaluated as significantly inferior compared to L_LCF and L_CC (Figure 3 and Figure 4).
• Major finding C-1. When the operation method was M_LS, L_CC was evaluated as significantly superior compared to L_LCF. When the in-vehicle location was L_CC, M_LS was evaluated as significantly superior compared to M_BP (Figure 4).
• Major finding C-2. When the in-vehicle location was L_RWS or L_UCF, M_LS was evaluated as significantly inferior compared to M_BP (Figure 4).

Major finding A: the finding that M_DR was evaluated as inferior to M_LS and M_BP aligns with actual accident cases associated with the use of M_DR. For example, in 2016, an accident occurred in a vehicle equipped with M_DR when a driver intended to shift into park (P) but mistakenly engaged reverse (R)—believing the vehicle was stationary, the driver exited the car; however, since it was in R, the vehicle moved backward and tragically ran over the driver, resulting in a fatality. Approximately 39,000 such accidents related to the use of M_DR were reported from 2016 to 2020. An investigation conducted by the National Highway Traffic Safety Administration (NHTSA) in the United States revealed that many of these incidents were not due to vehicle/transmission system defects but were more likely caused by driver confusion regarding the operation method [35]. Indeed, in this study, comments such as ‘inconvenient and unfamiliar operation method’ and ‘higher risk of driver error compared to other operation methods’ were collected regarding M_DR from the respondents. Furthermore, the results from the freehand design session (Table 8) also indicated that, among the three operation methods, the number of respondents selecting M_DR as their preferred operation method was the lowest (5 out of 43 respondents).

One possible reason for driver confusion and errors in M_DR can be explained by the lateral design of its operation direction (i.e., the rotation direction of the dial knob) and the arrangement of the gear positions (i.e., P, R, N, and D). Firstly, the operation direction of M_DR, involving rotation in a counterclockwise or clockwise direction, is not aligned with the vehicle’s longitudinal (forward/backward) travel direction determined by the use of the gearshift—in other words, M_DR does not satisfy movement compatibility [36]. In general, a higher degree of compatibility is associated with faster learning, shorter reaction time, fewer errors, and higher user satisfaction [37,38]. Additionally, the lateral arrangement of gear positions of M_DR, contrary to the longitudinal arrangement applied by automotive manufacturers over previous decades, can be considered a less satisfying driver stereotype [36]. Relatedly, the importance of satisfying driver stereotypes in designing in-vehicle controls and gearshifts has been emphasized in previous studies [10,29]. Therefore, future research needs to empirically investigate the ergonomic benefits of the longitudinal design of the operation direction and arrangement of gear positions of the dial-type gearshift.

Major finding B-1: firstly, the finding that the steering wheel spoke locations (L_LWS and L_RWS) were evaluated inferior can be attributed, at least partially, to the position of the gearshift outside the vehicle center, specifically, to the left of it—note that for the other three locations (L_UCF, L_LCF, and L_CC), the gearshift is positioned at the vehicle center. Traditionally, mechanically linked gearshifts have been centered for reasons such as maintaining the vehicle’s center of gravity and ensuring efficient power transfer—thus, the departure from this familiar placement in L_LWS and L_RWS could contribute to their inferior evaluations. Relatedly, it is recognized that the use of unfamiliar systems can potentially compromise drivers’ safety [39,40,41].

The inferiority of L_LWS and L_RWS can also be attributed to the continuous changes in the position and orientation of the gearshift, which results from the driver’s rotation of the steering wheel for lateral vehicle control. Indeed, some comments such as ‘it may be confusing due to the rotation of the steering wheel’ were collected regarding L_LWS and L_RWS in this study. The mental rotation of objects to align with one’s own left/right perspective is known to be cognitively challenging for many individuals [42]. Moreover, in this limited area of wheel spoke locations, the reduction in the size of the gearshift would be inevitable. Considering the findings from Chen et al. [43] and Feng et al. [44] that small in-vehicle controls were associated with longer reaction times and lower accuracy, the reduced size of the gearshift may affect the driver’s performance in gear-shifting tasks. Indeed, comments such as ‘I do not prefer the gearshift located on the steering wheel spoke, due to its much-reduced size,’ were also collected from the respondents in this study.

Meanwhile, the finding that L_LWS was evaluated inferior to L_RWS can be explained not only by the aforementioned fact that L_LWS is relatively farther from the vehicle center, but also by the reversal of the gearshift position relative to the ego center (driver’s midline). Since the gearshift is predominantly used on one side from the driver’s seat (for example, on the right side in the case of left-hand-drive {LHD} vehicles), positioning the gearshift on the opposite side could lead to driver confusion and errors, potentially resulting in safety issues. Indeed, relevant previous research has warned that the reversed and unfamiliar position of controls, such as gearshifts, can increase the risk of crashes [45]. However, it should be noted that the illustrations provided in this study depict LHD vehicles, and all respondents were drivers from LHD cultures. Therefore, future research should validate the results of this study by examining whether drivers from right-hand-drive (RHD) cultures would actually evaluate L_LWS as superior to L_RWS in RHD vehicles.

Major finding B-2: the finding that L_UCF was evaluated as inferior to L_LCF and L_CC can be explained by the increased physical strain imposed on drivers when the gearshift is positioned at L_UCF, requiring them to raise and extend their arm for gear shifting. In this regard, comments such as ‘it is preferable because of the easy-to-reach location’ and ‘it is comfortable because the gearshift is reachable with the arm lowered in a natural position’ were exclusively collected for L_LCF and L_CC. The findings from the freehand design session (Table 8) further support this, as only 1 out of the 43 collected pictures featured the gearshift in L_UCF, indicating that L_UCF is a non-preferred location. Relatedly, some design guidelines for in-vehicle controls also recommend placing controls in easily reachable locations by the driver’s hand [29]; additionally, empirical evidence has demonstrated that as the target distance and height from the driver’s seat increase, driver discomfort also increases [46].

The inferiority of L_UCF could also be interpreted in that the driver’s actions for using the gearshift in this location (e.g., reaching out for or operating the gearshift) could visually obstruct the information provided through the in-vehicle information system (IVIS) typically located in L_UCF. This interpretation aligns with design recommendations for IVIS that in-vehicle controls should be located so that the driver’s hand does not block the view of an in-vehicle display [47].

As a promising alternative that reduces the physical strain imposed and avoids the visual obstruction of IVIS, the ‘near the steering wheel’ location, as adopted by Porsche (Table 1), may be suggested. The freehand design session results (Table 8) also indicate a preference for this location, as all eight pictures with the gearshift placed at the location not considered in this study featured the gearshift near (on the left or right side of) the steering wheel. Considering that the ‘near the steering wheel’ location also reduces the driver’s eye and hand movement during driving, which is an aspect crucial for lowering the risk of road traffic accidents [16,19,20], further investigation into its effectiveness is warranted.

Major finding C-1: the finding that when the in-vehicle location was L_CC, M_LS was evaluated superior to M_BP and M_DR, and when the operation method was M_LS, L_CC was evaluated superior to L_LCF can be attributed to the widespread adoption of the combination of M_LS and L_CC (i.e., M_LSL_CC) in gearshift designs by many automotive manufacturers over several decades. In other words, it seems that M_LSL_CC has become the most familiar combination for drivers, leading to the most favorable evaluations. Indeed, M_LSL_CC was the most frequently collected picture (9 out of 43, 20.9%) from the freehand design session, with mentions of reasons such as ‘seeking a familiar operation method and in-vehicle location’. This aligns with the general design principle that consistently placing the same items in the same locations induces instinctive and automatic human behaviors [36]. However, it is essential to note that this perceived superiority might result from accessibility bias, a phenomenon that researchers generally believe should be approached with caution [48,49,50]. Consequently, careful and empirical examination is required to determine whether the design of this M_LSL_CC combination best serves driving performance and safety.

Major finding C-2: the finding that when the in-vehicle location was L_RWS, M_LS was evaluated inferior to M_BP can be explained in several ways. First, when the gearshift is located at L_RWS, gear shifting is mainly performed with the use of a finger, making M_LS, which involves relatively complex movements (e.g., the grasping and pushing/pulling of a lever), less favorable compared to the simple and easy-to-operate M_BP. Relatedly, a study investigating the driver preferences for the operation method and in-vehicle location of in-vehicle secondary controls also showed that the push button operation method was most preferred to be implemented in the steering wheel spoke location [51]. Additionally, given that hand contact frequently occurs in the area of the steering wheel spoke during driving, the protrusion of M_LS may be considered less desirable than the flatness of M_BP in terms of the possibility of unintended operations of the gearshift. Indeed, in this study, comments such as ‘unintended contact/operation may occur’ were collected regarding M_LSL_RWS from the respondents. Moreover, M_LSL_RWS also seems undesirable when considering that it increases the risk of critical injuries due to the impacts with its protruding part in the steering wheel spoke in the event of a collision—relatedly, Shaw et al. [52] found that impacts with steering control devices mounted on the wheel rim could cause severe chest, eye, and facial injuries in accidents.

On the other hand, another finding that when the in-vehicle location was L_UCF, M_LS was evaluated inferior to M_BP seems interpretable in terms of the physical strain imposed on the driver as well. As mentioned earlier in interpreting Major finding B-2, this location requires raising and extending the arm for gear shifting, increasing physical strain. Under this physically challenging condition, compared to M_BP, involving a simple finger press for shifting, M_LS requiring larger movements of the arm and shoulder, along with more uncomfortable hand and wrist motions for manipulating the lever, would make the physical strain induced by this location even worse. This interpretation is supported by empirical evidence that, at shoulder height, fingertip pushing posture was associated with lower in-vehicle reach discomfort compared to three-finger- and five-finger-grasp postures [53].

5. Conclusions

This study derived a total of 15 electronic gearshift design alternatives by combining three operation methods (M_LS, M_BP, and M_DR) and five in-vehicle locations (L_LWS, L_RWS, L_UCF, L_LCF, and L_CC). The respondents evaluated each design alternative across nine ergonomic evaluation measures. The results indicated that the best ergonomic operation method, in-vehicle location, and their combination were M_LS and M_BP, L_LCF and L_CC, and M_LSL_CC, respectively. A summary of the major findings and answers to the three research questions is presented in

Table 9. The major findings and answers to the research questions.

Major Finding	Research Question	Answer (Illustration)
A (MDR < MLS, MBP)	What is the best ergonomic operation method for electronic gearshifts?
		MLS		MBP
B-1 (LLWS < LRWS < LUCF, LLCF, LCC)	What is the best ergonomic in-vehicle location for electronic gearshifts?
		LUCF	LLCF		LCC
B-2 (LUCF < LLCF, LCC)
		LLCF		LCC
C-1 (MLSLCF, MBPLCC < MLSLCC)	What is the best ergonomic combination of operation method and in-vehicle location for electronic gearshifts?
		MLSLCC
C-2 (MLSLRWS < MBPLRWS, MLSLUCF < MBPLUCF)
		MBPLRWS		MBPLUCF

.

This study provides practical and theoretical implications/contributions as follows. First, through the inferior evaluation of M_DR (Major finding A), this study identified its HMI design issues, suggesting the need for a redesign of the M_DR that has been currently adopted with a uniform design featuring a lateral operation direction and lateral arrangement of gear positions. As a potential design solution for innovation/improvement in the HMI design of M_DR, a longitudinal design of the operation direction and arrangement of gear positions can be proposed, though its effectiveness will need to be investigated in future research. Second, through the superiority of the L_LCF and L_CC (Major findings B-1 and B-2), it could be recommended that when designing the location of electronic gearshift, consideration should be firstly given to the existing/traditional ones. Third, the superiority of M_LSL_CC (Major finding C-1) confirmed the excellence of the combination of the traditional operation method and the traditional in-vehicle location in designing electronic gearshifts. However, the relative superiority of M_BPL_RWS and M_BPL_UCF (Major finding C-2) implies that non-traditional operation methods could be superior when implemented in non-traditional in-vehicle locations, which provides opportunities for HMI design improvement/innovation for electronic gearshifts. Finally, the ergonomic design characteristics of electronic gearshifts derived in this study can serve as design/improvement guidelines to address the frequently reported safety issues associated with using electronic gearshifts. Furthermore, it is anticipated that these guidelines could be applied to the design of the operation method and in-vehicle location of electronic gearshifts for new types or future modes of transportation.

The limitations and future research directions of this study are as follows. Firstly, although the number of respondents in this study was sufficient to conduct statistical tests without concerns about statistical power, increasing the sample size would be necessary in future research to ensure more accurate results. In doing so, it can be recommended to recruit an additional group of drivers, beyond the two age groups (young adulthood and middle-aged) and two driving experience groups (novice and experienced) recruited in this study. For instance, an ‘older driver group’ (aged 70 and above) for age and an ‘intermediate driver group’ (falling between novice and experienced) for driving experience could be recruited. The inclusion of such additional driver groups may help reveal meaningful differences in the study results among various driver segments. Secondly, since approximately half of the 15 design alternatives derived in this study have never been implemented in practical settings (production car models), the respondents evaluated each design alternative in an illustrated form, relying solely on their imagination of future interactions/experiences with it. Thus, future research should consider providing respondents with actual prototype/product-based experience of using each design alternative before the survey to ensure more accurate evaluations. Thirdly, the current study findings obtained through the survey would need to be validated through a laboratory (simulator) or field (on-road) experiment where the physical interaction with each design alternative could be evaluated using various objective behavioral and physiological measures. Especially, when designing/conducting simulator experiments, the use of VR technology should be considered—indeed, it is known that implementing a VR environment can offer an immersive and interactive experience, making the virtual gearshift feel quite similar to the real gearshift [54,55]. Finally, besides the operation method and in-vehicle location considered in this study, there may be various other design variables (e.g., the sequence/layout of ‘P’, ‘R’, ‘N’, and ‘D’ gear positions and the dimension of the gearshift) that could be targeted for the ergonomic design improvement/innovation of electronic gearshifts. Therefore, future research could explore the design space of electronic gearshifts and empirically investigate the effects of some key design variables on driving performance and safety.