Keywords

1 Introduction

Gesture belong to the natural way of human interaction like language and has been an important part of human-computer interaction [1]. For more than a decade,, researchers have been advocating a new interface that will allow drivers to operate smoothly while driving without increasing cognitive load [2]. According to the study of Wahl, in the environmental interaction era which we’ll face right away, the traditional ways of interaction rely on the display and keyboard to communicate with the user may will be eliminated, we need a Interactive mode which will be able to link the technology and the users without the third party [3]. As a natural way of interaction, gestures can greatly reduce the cost of cognitive and visual communication, which have great potential for development.

Driving modern cars is a complex task. Drivers must constantly observe traffic and instrumentation while dealing with various random distracted driving tasks, such as telephone rings or conversation with passengers, etc. [2]. However, the number of in-car devices has reached the limits of the driver can bear, especially a large number of traditional input devices such as knobs, buttons and handles [4]. Therefore, safety and cognitive load reduction are the primary concerns for all car interaction designs. Compared with the traditional visual interaction, gesture interaction can keep the driver’s visual attention on the driving task [5]. At the same time, Geiger et al. found that gesture interaction can effectively reduce the driver’s distraction [6]. As a result, a number of studies have attempted to bring gestures into car design [7, 8].

However, there are still some problems in the design of interactive gesture. For example, the vast majority of car interior interaction design and external styling design are basically independent of each other, which has resulted in the design of car interior design and external design split. On the other hand, the current interactive design research is relatively concentrated in the parametric test [9], but pay little attention to cultural and natural factors. Therefore, in order to solve the above problems, this paper intends to study the car gesture interaction design by drawing on the knowledge of car styling and culture rules.

2 Lines and Car Gestures

In the early period of car styling design, it is often used to construct the car styling with the line sketch, and the car styling feature line is also considered as an important representation of car styling [10]. Zhao has characterized the car styling line summarized as 20 [11], and used it to guide the car shape design. In order to keep interior styling and car exterior styling coordinated, designers often look for the modeling feature from the external modeling characteristic line and apply it to the configuration of various parts of the interior (Fig. 1). Therefore, the lines can be used to characterize the knowledge of car modeling.

Fig. 1.
figure 1

Car exterior and interior styling lines

Full size image

For the “Whole-hand gestures” [12], people usually perceive gesture behavior by capturing the movement of the hand, i.e., by observing the “fingertip curve” [13]. This finger movement forms a dynamic line on a time dimension. At the same time, lines are also an important way to construct various cultural products in the world [14], especially in China, which emphasizes image culture. Therefore, taking the line as the design ontology, it is possible to make the car gesture design conform to the cultural aesthetic characteristics and to benefit the whole design of the car by drawing the rules of the car modeling and cultural field.

3 Gesture Feature Experiments

The purpose of the service design is to make the properties of the design product converted from the “Useful” and “Available” and other basic needs to the “Satisfaction” and “Easy to use”, and other high-level needs. In order to provide satisfactory services, in the service design the designer needs to focus on the user experience inspection and understanding. The commonly used hand movements in the car reflect the behavior patterns and cognitive habits of users in driving situations. Take these hand movements as the gesture design reference object, you can design interactive products which are more suitable for user habits. In order to obtain the types and characteristics of hand behavior in the car, the feature experiment was carried out.

3.1 The Experimental Process

20 users (9 females) were selected in the experiment. The flow of the experiment is as follows: 1. First of all, through the questionnaire to obtain the user’s basic information and driving tasks make the hand position changes (relative to the normal driving hand position). 2. Based on the driving tasks obtained through the questionnaire, a number of major test scenarios were designed to allow the user to complete the relevant driving behavior in the test scenario. The researchers used the observation and video recording methods to capture their behavior. 3. Got typical hand behavior characteristics through the behavior classification statistics.

In the experiment, 3 common driving tasks were identified by questionnaire: overtaking; answering telephone calls; adjusting (audio or air conditioning, etc.). Then, according to the experimental design of Pellegrino [15], we selected a straight road section in Sichuan University. The experiment scene of overtaking was simulated by the auxiliary car of the experimental staff, and simulated the experimental scenarios of answering phone calls and adjusting. After that, we found that users have two typical characteristics of hand movements through observation and video analysis. The first feature occurs in the hand movement of a long path, such as when the driver shifts, the right hand leaves the steering wheel to the vicinity of the gear lever, and then back to the steering wheel; When the driver adjusts the volume knob, the right hand will also reach the center position, then back to the steering wheel; In answering the phone, the driver also takes the phone from the central position of handrail to the head position and so on. The common feature of these behaviors is that the trajectory of the hand is long; the driver will not have long-term eye-tracking (without looking at the movement of the hand, or looking at it and immediately returning to watching the road); In the experiment, the users adopt random, comfortable and inaccurate hand movement; Eventually, the hands return to the steering wheel position (which is determined by the primacy of the driving task); All users use the right hand to perform tasks. The second feature occurs in the hand movement of a short path. For example, the driver’s right hand leave the steering wheel to press the horn; Hand movement in the shift; And, after answering the phone many users will not directly put the phone back in situ, but holding the phone from the ear shift to the steering wheel position to pause (then the right hand will contact the steering wheel for a short driving time), and then return the phone to the center armrest and so on. The common features of these hand movements are short trajectories; And the iterative repetitive movement often occurs. For example, the driver often presses the horn of the car more than once; After the phone, some users will once again lift the phone back to the ear to listen to confirm and then put it back; As well as the fine-tuning of reciprocating motion and so on.

3.2 Experimental Results

It can be found that the basic gesture behavior can be divided into two categories through the experiment: Long path gestures and short path gestures.

3.2.1 Long Path Gesture Design Features

The main design features of long path gestures are: As the driver in the driving process will not give a lot of visual attention to hand behavior, the hand can not be too close to the center console and other entities, so the designer should give the gesture to a large redundant space to avoid interference. At the same time, in order to facilitate the capture of sensors, Akyol et al. have found that car hand movements need to be limited to higher than gearshift [7]. Thus, in very limited space, long-path gestures that are highly variable are unlikely to occur; Secondly, because of the long path gestures with casual, comfortable features, therefore, the gesture design needs to meet the ergonomic requirements of the human body. In McNeill’s study, the comfort of a series of gestures has been summarized, Such as “when the elbow flexion more than 90° the muscles will be uncomfortable; try to avoid the wrist and arm rotation” and so on [16], these can be used as the reference of long path gesture design. In addition, the hand will eventually return to the steering wheel, so the end of the path should be located in the vicinity of the right side of the steering wheel, which means that the end of the gesture design should also be located near the right side of the steering wheel, so the gesture designer only needs to consider the starting position.

3.2.2 Short Path Gesture Design Features

The main features of short path gestures are as follows: Short path gestures often occur in iterative and repeated paths. Through the user interviews we know that the main reason for this phenomenon is due to the ower cost of conduct of short hand gesture. And many users doubt the accuracy of the sensor, so the users always do a few more repeated action to confirm. At the same time, there are users that only do short-path gesture for one time feel “too little work,” so subconsciously repeated several times. In short, both for the user’s psychological needs or task requirements, iterative short-path gesture design needs to be taken seriously.

4 Implementation

Most of the time the service is invisible, but the user will feel the full range of services from multiple sensory channels. and the invisible service can also be displayed based on the subconscious experience design of the line. The main purpose of the design practice part is to design a gesture product for a target model (see Fig. 2). Based on the Loehmann study, the content of car gestures can not be too much [17]. At the same time, because the most important purpose of car interaction is to drive safely, our design goal is to design a car-assisted driving products with no more than 5 kinds of gestures. Based on this goal, the following sections detail how to design long path and short path gestures.

Fig. 2.
figure 2

Target model

Full size image

4.1 Long Path Gesture Experience Design

Usually, straight line trajectories with long distances and gesture behavior with sharp kink traces will result in discomfort, so gesture lines are usually designed with smooth over-curves. Curves tend to give people the feeling of soft and elastic, but car styling often pursue the image of muscle, strength and speed, which requires a more gentle curve adjusted to a tension curve. This requires a curve with a sense of tension. Tension is a psychological effect, which is a important way to create strength of the car modeling line. Arnheim argues that the purpose of tension is to make “the static object has a dynamic posture” [18]. Bangor created the flame surfacing for BMW because Aristotle said that “the shape of the flame is the most active of all shapes, because the shape of the flame is most conducive to produce a sense of movement.”

Arnheim describes the way in which tension is constructed as “a temporary equilibrium or dynamic equilibrium with a thrust, thrust, and pull against each other” [18]. In the car styling line, designers tend to use the appropriate tilt and bending deformation to create imbalances, so as to provide tension to lines. Similar to car styling features, dynamic gesture path lines also require tension. However, it should be noted that the inclination and deformation are the consequence of the interaction between the applied force and the resistance force. Appropriate deformation and tilt like a bow, can be used to accumulate strength, But the excessive deformation means that the resistance is too weak, which is detrimental to the tension; On the contrary, not enough deformation means that the force is not strong enough.

In order to design a gesture that matches the car body shape image, it is necessary to study the characteristics of car body modeling. Therefore, according to the characteristic line theory, the researchers extracted the main feature lines of the target model from the front and side view angles, and analyzed and integrated the local variation of the lines. As shown in Fig. 3, to make gestures more natural, the gesture trajectory must be smooth and excessive without straight-line angle, so the straight line angle added Bezier curve. Therefore, the straight-line corners are changed to over-curves based on the Bezier curve. The results show that the lines can be divided into three categories (Fig. 4): Class A is asymmetric large-curvature lines; Class B is symmetrical large-curvature lines; and Category C is acute-angle small-curvature lines. In order to further obtain the optimal degree of deformation of each type of line, we set 5 deformation degree of different curves in each type of deformation curve (Fig. 4). The Likert scale questionnaires were designed from three aspects: the degree of visual tension, the degree of matching with exterior shapes, and the esthetic level of strength and speed. On the Internet, 50 users (24 women) were scored, and the highest score was obtained: A3. A3 curve is a kind of Class A curve, and the deformation degree is in power savings and restraint. It is also fit with the body waist line. Therefore, A3 curve is taken as the basic path form of the long path gesture.

Fig. 3.
figure 3

Target car modeling line feature extraction

Full size image
Fig. 4.
figure 4

Three types of basic lines

Full size image

4.2 Short Path Gesture Experience Design

Similar to the gesture, a traditional culture - Chinese calligraphy is also used lines to build aesthetic [14]. It depicts emotions and images through the lines with the movement of the nib that is squeezed with ink. It uses lines to build beauty in many ways. Because short-path gestures often use repeated lines to enhance its sense of presence, in order to find out the factors that help to construct the short path gestures, the researchers analyzed more than 20 famous works of calligraphy, It is found that there are a lot of works in the use of repeated ways to enhance the performance of local lines of the phenomenon like short path gestures. In Liang’s study, this phenomenon is called a mass effect [19]. Based on this, the researchers drawn and summarized the repeated features of the lines in the calligraphy works, and obtained a series of local characteristics of the calligraphy lines (Fig. 5).

Fig. 5.
figure 5

Characteristics of calligraphy lines

Full size image

Next, the researchers also need to carry out these lines for further screening and improvement. As the car gestures can not be too complex, so we first remove some of the lines which are too complex. For the remaining lines, we invited 20 users in the laboratory to simulate gestures along the line path. According to the smoothness of the operation, visual effects, and the user’s own evaluation results after simulation, the optimal line is B by scoring. The experiment also received additional information: most of the users think that the basic line type at least need to be iterated more than 2 times (Fig. 4, B-line iterative 3 times) to form a gesture. While the upper limit is 5 times.

4.3 Car Gesture Products and Testing

Based on the acquisition of the long path and short path gestures described above, the designers devised a car gesture product. The main purpose of this product is to be able to use gestures to interactively switch the map on the HUD and the central control screen display, and can manually adjust the display brightness. Service design not only needs to consider user needs, but also take into account the needs of the designer. Through the face-to-face communication with the typical designers and users, to record the key processes in the service. At the same time, detail is the key to the service design, so the first need is to sort out the touching points. The contact points between the user and the product were used as the touching points. In order to reduce the driver’s operating frequency and reduce cognitive load in the driving situation, the number of touching points were reduced as far as possible without affecting the user’s operation. Finally, through a series of comparisons and design iterations. The number of the touching points were reduced to 3. Based on the contact, the designer has designed the product tasks. And because the Sequencing of the tasks is very important in the service design, the designer also considered the order of the three functions. The final three tasks are: open; stow; adjust the brightness. Since the function of opening and stowing is the opposite, in order to reduce the cost of learning and cognition, the gestures of stowing and opening are also designed as opposite lines, so that users can learn one kind of gesture and then deduce the other. In addition, since the right hand needs to return to the steering wheel position after completing the gesture, the end point of the gesture points to the right side of the steering wheel. The starting point is located in the original picture of the transferred image in the position of the central control display screen, gesture trajectory design using long path gestures (Fig. 6). Conversely, the hand gesture is reversed when the opposite “stow” task is performed (Fig. 6).

Fig. 6.
figure 6

Gesture design

Full size image

Similarly, using the resulting short path gesture, the adjusted luminance gesture is designed as shown in Fig. 6. And because the required degree of brightness is different, brightness adjustment function includes three modes: bright, normal and dark, It is switched by the number of basic iterations. With reference to previous studies, the minimum number of iterations is twice, at most 4 times.

Subsequently, the researchers conducted user research. As the prototype has not yet been produced, it is impossible to perform accurate quantitative testing. But we let users to simulate the operating experience in the car. That means, after the action is performed, the feedback is given on the screen of the car, which is controlled by the experimenter. The main purpose is to obtain the user’s response and information which can optimize the product. First, let the user in accordance with the gestures we designed to simulate the operation, and then let the user in accordance with their own ideas to design their own gestures, the final interview, eventually interviewed them. Through the experiment we found that the gesture interactive products can significantly stimulate the user’s interest and concern, and they can learn in a short period of time, most of the tasks have been completed. Users believe that gesture design and car styling and cultural factors are more fit, and get a great sense of satisfaction and pleasure. However, there are also problems such as a deviation in the position of the user’s operation when performing the brightness adjustment gesture. Someone like to operate in the vicinity of the right side of the steering wheel, and someone like to operate in the center console position. These require further user research; There are some users have doubts of the implementation ability of non-contact gesture interaction, so some users keep giving visual attention in the operation. Although most users feel the gesture track is very beautiful, but if you do not tell this is related to calligraphy and car styling, most users can not find on their own. These need further consideration of the follow-up design.

5 Conclusion and Future Works

As Donald Norman said, “Good behavior comes from careful design.” In this study, we propose the design of car gesture interaction based on line experience design. In order to make the car interaction design consistent with the external design, as well as to meet the user’s cultural and cognitive habits, the modeling principles of car styling lines and calligraphic lines are utilized to design long path and short path gestures for car, and designed the interactive product based on experience design and service design, which provides a possibility for the design of hand gesture and the design of car interactive products.

Through the current design and testing, the research has found some factors that should be paid attention to. Such as iterative design features of short path gestures; spatial constraints of long path gestures; and final hand position problems of gestures and so on. In addition, there are still some problems, such as the continuing interest of users mentioned by Coskun et al. [20]. After 2 months, when the freshness subsided, can users still stick to the gesture interaction product? As well as the above mentioned, how to conduct a unified gesture design for users with different behavior habits? And how to make the user feel the content and image of the gesture is more intuitive and interesting? At the same time, the current research is mainly qualitative exploratory research, and further work is needed to enhance the credibility and availability of the study. These subsequent quantitative tests will have to wait until the prototype is finished. Will there be new discoveries with the participation of quantitative data? And whether the current findings will be amended? All of these need to be further studied.