Development of a Robotic Device that Performs Head Bunting to Relieve User Tension
Article No.: 20, Pages 1 - 15
Abstract
Inspired by an animal behavior called bunting, in which the animal rubs its head against other objects, including humans, we developed robotic prototypes capable of performing such bunting behaviors. Since physical contact plays an important role in therapeutic interactions between pets and their owners, we hypothesize that robot bunting can have a similar effect on its user. This article reports on the development of such a robot with a flexible neck that can change its stiffness while the robot is rubbing its head against the user. An exploratory study was also conducted with 22 human participants, on whom the developed robot performed head bunting with three different (low/high/variable) stiffness conditions. The results show that the participants’ psychological tension, as measured by the temporary mood scale (TMS), was significantly reduced (p \(\lt\) 0.001) after interacting with the robot. The difference between the three stiffness conditions was not significant in this study. Due to the lack of a control condition, we cannot confirm a clear effect of the stiffness change; however, some participants commented that the stiffness change made the robot's behavior lifelike and relaxing.
1 Introduction
Interaction with companion animals has several positive effects on humans [3]. In particular, touch interactions can reduce human stress and anxiety [23, 25] and play an important role in animal-assisted therapy [11]. However, there are risks associated with animal contact, such as the development of allergies, injuries caused by animal bites and scratches, and the development of infectious diseases, which is driving the research and development of animal-like robots for therapeutic purposes [2, 21, 33].
However, compared to companion animals, the contact behavior of these animal-like robots with humans is still limited. When an animal touches an object, the animal can adjust the joint stiffness and dynamically change the degree of force transmission depending on the contact condition. For example, in human fingers, the stiffness of the finger varies depending on the flexion or extension state of its joints [19]. Variable stiffness joints have been introduced in robots [9, 35], including the use case of the robot neck [7]. However, their utility for therapeutic purposes is unknown. On the other hand, most existing therapeutic robots only provide a soft touch or response that gives a constant sensation when in contact with a human.
An example of contact behavior in companion animals is the animal behavior called bunting, in which an animal (typically, cats) rubs its head against a social partner to express affection or for scent marking [8, 18, 20, 24]. Since the importance of touch and overt expressions of affection has been discussed in human/companion animal therapy [16], we hypothesized that robot bunting might have a similar effect on its user. Despite the lack of empirical research examining the precise effects of animal bunting on humans, analogous tactile interactions between animals and humans, such as snuggling or cuddling, have been employed in therapeutic contexts [28, 31]. With respect to the act of snuggling, a robot named SnuggleBot was developed with the objective of assisting humans in coping with feelings of loneliness [5, 6]. Although the robot was designed to be snuggled by the user and is distinct from a robot that snuggles or bunts the user, we hypothesized that robot bunting might also have a beneficial effect on its user, given the bilateral nature of haptic interactions.
In light of these insights, we undertook an exploratory development and study to investigate whether robot bunting has a beneficial effect on its user. Our first goal was to explore the design of a robot that could replicate bunting. In the process of prototyping two devices, we first identified the basic requirements for robot bunting and then created a robot with a flexible neck that can change its stiffness while the robot is rubbing its head against the user. Then, we had a research question about whether the change in stiffness made a difference in the user's mood, such as tension. To explore this research question, 22 participants were recruited to test the robot. The participants were divided into three conditions that tested three different (low/high/variable) stiffness of the flexible neck. By using the temporary mood scale (TMS), the participants’ mood was measured before/after testing the corresponding movement of the robot bunting depending on the condition. This article reports the results of all these research and development.
2 Related Works
Haptic interaction is an important topic in human–robot interaction (HRI) research. In particular, touch in HRI has been actively studied, as summarized in recent review articles [10, 27]. From the perspective of therapeutic interaction, animal-like robots have been introduced and studied [2, 21, 33]. There are also commercial robots that market their tactile qualities [13]. However, as discussed in the previous section, these existing robots can only provide a soft touch or response that gives a constant sensation when in contact with a human. There is no robot that can perform behaviors like bunting.
Hugmon [37] is a huggable robot based on a tensegrity structure of rigid elements connected by springs, which allows flexible deformation in response to external forces. Although the robot's movement (structural folding and expansion) is different from that of bunting, the physical interactions assumed in Hugmon are similar to ours: holding the robot in front of the user's stomach. Hugmon's movement is generated by six pneumatic linear actuators that expand and contract. So far, the emotional expression of these movements has been studied [37]. Although the effect on the user's mood has not yet been tested, the physical interactions assumed in Hugmon may also have a positive effect on the user's mood.
SnuggleBot is also a huggable robot with similar physical interactions [6]. The robot was built into a soft stuffed animal with a pocket for a heat pack and actuators for its flippers and tail. As the name implies, SnuggleBot was designed to be snuggled by the user. SnuggleBot has been placed in homes for extended periods of time, and its potential to help people cope with feelings of loneliness has been discussed [5]. SnuggleBot and our bunting robot are complementary, i.e., the former is designed to be snuggled by the user and the latter is designed to snuggle or bunt the user.
A therapeutic robot that changes the softness of its body surface has been developed to motivate its user for prolonged haptic interaction with the robot [36]. The robot uses a thermo-responsive hydrogel as the material of the user contact surface and can change its softness by heating or cooling the gel. However, it was reported that the speed of the softness change was slow, and sometimes participants had difficulty in perceiving the change [36]. To create robot bunting, the gel alone may not be sufficient to provide variable stiffness.
Regarding the mechanical approach for variable stiffness, Li et al. reviewed and compared several joints with adjustable compliance and then proposed a cable-driven antagonistic joint designed with variable stiffness mechanisms (VSMs) [17]. The proposed joint was also applied to a wearable elbow exoskeleton [17]. On the other hand, an endoskeleton mechanism driven by an internal tendon to control the stiffness of the whole structure was analyzed [14] and applied [22] to a dolphin-shaped robot with the VSM in its tail fin to improve its underwater thrust. The tail fin section consists of two driving wires and two stiffness changing wires, and the tail fin can be adjusted by pulling the stiffness changing wires [22]. In our study, we use a similar mechanism (but we use springs for the compliant segments) to reproduce bunting behavior.
3 First Prototype and Pilot Test
Figure 1 illustrates our envisioned bunting robot and how it interacts with the user. The robot should be sufficiently small such that it can be placed on the user's lap.
Fig. 1.
3.1 Development of the First Prototype
The first prototype [1] we developed is shown in Figure 2. It was designed with a simple appearance to eliminate visual influence. The dimensions of the robot are \(140 \mathrm{mm (W)}\times 140 \mathrm{mm (D)}\times 240 \mathrm{mm (H)}\). The weight is 524 grams, not including the circuitry, power supply, and outer skin. From observing videos of cat bunting, it was determined that the head bunting (rubbing) motion requires the execution of a motion in which the head is pressed against the user, and a rubbing motion when the head is pressed against the user. To realize these motions, one servo motor for the forward tilt and three servo motors for the rubbing motions were implemented (Figure 2(b)). The forward tilt motor is attached to the body of the robot. The motors for the rubbing motions are attached to the head. This allows the robot to perform rubbing motions in three directions: roll, pitch, and yaw. The maximum range of motion in the roll, pitch, and yaw directions is 30 (\(\pm\)15), 105 (\(\pm\)52.5), and 120 (\(\pm\)60) degrees, respectively. The robot's frame was fabricated using the Stratasys F120 3D printer [29] with acrylate-styrene-acrylonitrile filaments. The head, which is the area that touches the user, was covered with a furry polyester fabric. To ensure safe human contact, Kondo Kagaku Co., Ltd. [15] serial servo motors (KRS3302 ICS; maximum output torque: \(0.66\mathrm{N\cdot m}\)), which have current and temperature limiting functions and can feed angle information, were used as actuators. Arduino Nano Every with two serial communication paths was used as the microcontroller to communicate with a PC and the motors.
Fig. 2.
3.2 Pilot Test with the First Prototype
A pilot test was conducted with the first prototype and three graduate students to see how they felt about seven motions (Table 1), each of which is generated by a different combination of servo motors (Figure 2(b)). Participants tested all seven motions, and their impression of each motion was rated using a questionnaire. Participants were seated facing the robot, and each motion was performed on the participant's abdomen. The robot was fixed on a flat plate placed on their lap to stabilize its position, and the experimenter controlled the start and end of the motions. For each motion, the robot first slowly tilted its head toward the user using the servo motor attached to its body. Once the robot head made contact with the user's abdomen, each motor of the head performed a constant velocity reciprocating motion in the range of 34 degrees for 2 seconds to perform the rubbing motion. The motion was created using a servo library released by Kondo Kagaku Co., Ltd and smoothed by linear interpolation of the positions between poses. At the end of each motion test, the participants were asked to rate the “lifelikeness,” “cuteness,” “friendliness,” and “comfortability” of the robot's motion. After all motions were completed, the entire test was evaluated by interviewing the participants.
Table 1.
| Motors | ||||
|---|---|---|---|---|
| Neck Pitch | Head Pitch | Head Yaw | Head Roll | |
| Motion 1 | \(\circ\) | \(\circ\) | ||
| Motion 2 | \(\circ\) | \(\circ\) | ||
| Motion 3 | \(\circ\) | \(\circ\) | ||
| Motion 4 | \(\circ\) | \(\circ\) | \(\circ\) | |
| Motion 5 | \(\circ\) | \(\circ\) | \(\circ\) | |
| Motion 6 | \(\circ\) | \(\circ\) | \(\circ\) | |
| Motion 7 | \(\circ\) | \(\circ\) | \(\circ\) | \(\circ\) |
Table 1. The Composition of the Motors Used in Seven Motions
\(\circ\): The motor in use.
Due to the small number of participants, we did not perform a statistical analysis of the results. However, it was found that the “lifelikeness” and “friendliness” attributes were highly rated for motions that combined pitch and yaw motors. In addition, the “cuteness” attribute was highly rated for the head moving motion in the pitch direction. These results suggest that the head pitch and yaw axes may well influence the user's impression of the robot during the rubbing motion. Among all the motions performed, there was no difference in terms of the user's impression regarding the comfort of the motion. After all motions were tested, one participant with experience in cat ownership commented that the rubbing force was weak compared to the feeling of being rubbed against by a cat. This difference in tactile sensation may have been caused by the weakness of the rubbing force, resulting in an inadequate perception by the participant. Another participant reported feeling vibrations during the rubbing motion, which may have been caused by the fact that the robot had a rigid frame that easily transmitted the vibration of the motor directly to the participant, suggesting the need to review the structure of the robot.
4 Development of the Second Prototype
Based on the evaluation of the first prototype, we developed the second prototype, taking into account two requirements discussed in the previous section: rubbing motions in the yaw and pitch directions, and soft motions. We also considered the fact that animals can adjust the stiffness of their joints (e.g., when a human grasps an object such as a sponge, they control the stiffness of their finger joints to grasp the object without crushing it). We hypothesized that changing the stiffness of the joints during contact with the user would result in more lifelike force transmission.
The second prototype (Figure 3) we developed has a flexible neck with variable stiffness. Similar to the first prototype, the frame was fabricated using a 3D printer with acrylate-styrene-acrylonitrile filaments. Since there was a comment on the first prototype about the possible effect of the furry fabric, and we wanted to eliminate this effect in the user study described in Section 5, we decided not to cover the head of the second prototype with the furry fabric. The dimensions of the second prototype are \(160\ \mathrm{mm\ (W)} \times 156\ \mathrm{mm\ (D)} \times 330\ \mathrm{mm \ (H)}\). The weight is 870 grams, without circuit and power supply. The face part is a sphere without eyes to avoid the influence of the anthropomorphic appearance on the user's impression. In the first prototype, the desired transmission force could not be achieved because the motor torque was insufficient. Therefore, the second prototype uses a high-torque serial servo motor with metal gears (Kondo Kagaku Co., Ltd. [15] KRS-3304R2 ICS; maximum output torque: \(1.36\mathrm{N\cdot m}\)).
Fig. 3.
4.1 Flexible Neck
The required specifications for this mechanism were that it should be able to perform a flexible bending motion and that it should be able to change the joint stiffness smoothly and quickly. In addition, the number of components had to be minimized in order to keep the size and weight at a level where users could feel comfortable touching the robot. Therefore, we focused on a wire-driven continuum mechanism that can perform bending motions with two motors. A spring and wire mechanism was used for the neck of the robot. In this mechanism, rigid plates corresponding to the frame are coupled with elastic materials (springs) to allow soft bending motions. When the wire is pulled, the neck bends in the direction of the pull; when the pull is stopped, the neck returns to the neutral state due to the elasticity of the spring. This mechanism is easy to control because the amount of bending can be adjusted according to how much the wire is pulled, and the mechanism can be manufactured using relatively inexpensive materials. This mechanism has been widely used in various robots, including those for space operations and medical purposes [34], as well as commercial entertainment robots [38].
A mathematical model of this mechanism has also been discussed [26]. Ignoring the effects of torsion during bending and the effects of spring elasticity and wire stiffness, the 2-df bending property is expressed aswhere \(\alpha\in[-\pi,\pi]\) is the rotation angle and \(\theta\in[0,\frac{\pi}{2}]\) is the bending angle of the bending plane, \(l_{i}\) \((i=1,2,3,4)\) are the lengths of the driving cables, and \(r\) is the radius of the circle where the anchorage points are located [26] (see Figure 4).
\begin{align}\alpha & =\arctan{\frac{l_{4}-l_{2}}{l_{3}-l_{1}}}\end{align}
(1)
\begin{align}\theta & =2\arcsin{\frac{\sqrt{(l_{3}-l_{1})^{2}+(l_{4}-l_{2})^{2}}}{4r}},\end{align}
(2)
Fig. 4.
The bending motion of the developed mechanism is shown in Figure 5. For the rigid plate traction wire, a stainless steel wire with minimal axial elongation was used to transmit the traction force from the motor to the plates. The number of plates was set at 4, which is the minimum number that allows sufficient flexion during wire traction. In fact, there are seven cervical vertebrae in series in the feline neck. A radiographic range of motion study [12] reported that the range of motion of the feline neck in lateral flexion is approximately 19.4 degrees, and three of the seven cervical vertebrae contribute most of the range of motion. Therefore, considering the possibility of reproducing a range of motion similar to the feline flexion range of motion with fewer plates, we developed a four-plate mechanism. The plates are connected to each other by ball joints that allow flexure without directional constraints. The spring supporting the plates must have a small spring constant to suppress the effect of the low stiffness condition described below. Therefore, a spring with a free length of 22 mm and a spring constant of 0.161 N/mm was selected. To increase the range of motion during bending, the second plate from the motor side was selected as the point to be pulled by the metal wire (Figure 6). The silicone tube connecting the top three plates not only prevents the spring from being dislodged by extension but also helps the mechanism twist in place of the less rigid spring.
Fig. 5.
Fig. 6.
4.2 VSM
In order to vary the stiffness in the pitch direction of the mechanism described in Section 4.1, a VSM was developed as shown in Figure 7. Two metal wires pass between the bottom and top plates of the flexible neck in the pitch direction. The stiffness of the entire cervical region can be changed by applying tension and relaxation to these wires. A rotary plate cam mechanism is used to control the tension of the wires. The wire is suspended from the plate cam, and the cam rotates at a constant speed with a servo motor attached to the back of the robot, which in turn changes the tension applied to the wire. Since the motor rotates in the yaw direction, a bevel gear (module: 1.5 mm, number of teeth: 20) is used to change the direction of the motor rotation by 90 degrees, which is used as the rotation axis of the cam.
Fig. 7.
5 User Study
A user study was conducted to test the second prototype. In particular, we measured participants’ mood, such as tension, before/after experiencing the robot's rubbing motion, measured their impressions of the robot, and examined whether they perceived the variable stiffness of the robot.
5.1 Participants
Twenty-two participants (12 males, 10 females, age: 18–32, \(M=21.4\), SD \(=\) 2.90) were recruited from the University of Tsukuba, excluding the members of the laboratory with which the authors of this article are affiliated. Each participant was compensated with the amount of 1,380 JPY for their efforts. The study protocol was approved by the Research Ethics Committee of the Faculty of Engineering, Information, and Systems in the University of Tsukuba (2022R681), and all participants provided informed consent.
5.2 Study Design
Participants were divided into the following three conditions (between-participants design). In each condition, the robot repeated a rubbing motion (60 degrees of yaw and 40 degrees of pitch) in 4 seconds for 10 cycles, for a total of 40 seconds. The specific values for these parameters were determined based on footage of cat bunting and the impressions gained from the pilot test.
—
Condition-1 (low stiffness condition, seven participants): The cam is at the top and the stiffness is minimum (constant).
—
Condition-2 (high stiffness condition, eight participants): The cam is at the lowest point and the stiffness is maximum (constant).
—
Condition-3 (variable stiffness condition, seven participants): While the robot is performing the rubbing motion, the variable stiffness cam rotates and the stiffness changes dynamically. The cam rotates 180 degrees each time the robot performs a back-and-forth rubbing motion and the stiffness changes. In other words, the high stiffness and low stiffness states are repeated every 4 seconds.
5.3 Procedure
Participants were instructed to sit on a chair in a pre-determined position (Figure 8). They were then instructed to hold the torso of the robot with both hands. The bottom of the robot was attached to the top of a desk. The description of the robot's movements was displayed on a tablet computer so that they could review the tasks to be performed during a session. Next, an experimenter manipulated the robot to slowly lean forward toward the participant. When the robot's head made contact with the participant's abdomen, the participant reported the contact to the experimenter. The experimenter then paused the robot's motion. During this pause, the participant was asked to wear headphones to block auditory stimulation. To focus their attention on the contact with the robot, they were asked to close their eyes and listen to white noise through the headphones. Then, the experimenter started the rubbing motion of the robot for 40 seconds. After the rubbing motion was completed, the robot slowly stood up and returned to its initial position.
Fig. 8.
The entire flow chart of the user study is shown in Figure 9. The total duration was approximately 75 minutes. First, participants described their first impressions of the robot in an open-ended response. Then, after their baseline mood was measured by using TMS (see Section 5.4), the rubbing motion of the pre-allocated condition was presented. After the rubbing motion ended, participants completed questionnaires measuring their mood, impressions of the robot, and perception of stiffness. Next, participants were allowed to manually check the stiffness, i.e., they were allowed to move the robot's head by hand. They then answered open-ended questions designed to assess their impressions of the robot's appearance and their experience with pet ownership. Finally, participants were asked to test all three conditions and to evaluate the difference in perception under each condition by giving their honest opinion. The goal of this free evaluation period was to encourage spontaneous feedback from the participants by allowing them to easily interact with the robot. After giving the participants an overview of the conditions to be considered, the motions corresponding to each condition were performed. The motions were performed several times at the participants’ request. After all motions were performed, the following questions were asked of the participants via verbal communication:
(1)
Did you notice a difference in stiffness under different conditions?
(2)
What conditions have you found to be the most healing?
(3)
What are the improvements in stiffness and/or in motion?
Fig. 9.
5.4 Measurements
The TMS [32] was used to assess the mood of the participants; the TMS consists of six sub-scales: “tension,” “depression,” “anger,” “confusion,” “fatigue,” and “vigor.” Each sub-scale consists of three items, and each mood is rated on a 5-point scale (strongly disagree: 1, strongly agree: 5). The sum of the three items is the total score for the scale. This study focuses on the tension scores.
To evaluate participants’ impressions of the robot, this study used items from the Godspeed questionnaire, which is widely used as an impression evaluation index in HRI [4]. The Godspeed questionnaire is a five-level evaluation of adjective pairs representing the user's impressions and psychological state.
To rate the stiffness of the robot during contact, participants were asked to rate the following two questions on a 7-point scale:
—
Q1: Please rate the level of resistance you just received from the robot.
—
Q2: To what extent did you feel a change in force while the robot was moving?
6 Results
6.1 Tension Scores
The TMS tension scores of the participants before and after the rubbing experience are shown in Figure 10. The data met the normality assumption in Kolmogorov-Smirnov tests. A \(3\times{2}\) two-way repeated measures ANOVA was performed with a between-participants factor of condition (1/2/3) and a within-participants factor of time (before/after). The interaction was not significant (\(F(2,19)=0.410\), \(\mathrm{p}=0.670\), \(\eta_{p}^{2}=0.041\)). However, there was a significant main effect of time and participants’ tension score was significantly decreased after they experienced the rubbing motion (\(F(1,19)=18.5\), \(\mathrm{p}=0.0004\), \(\eta_{p}^{2}=0.493\)). In addition, although the condition factor was not significant, the decrease rate (\(33.3\mathrm{\%}\)) of condition-3 (variable stiffness) tended to be greater than those (\(21.3\mathrm{\%}\), \(27.3\mathrm{\%}\)) of condition-1 (low stiffness) and condition-2 (high stiffness).
Fig. 10.
6.2 Impressions of Robot
Figure 11 shows the main results (average scores on four items related to our robot) of the Godspeed questionnaire. A one-way ANOVA was performed on each item to compare scores across conditions. No significant differences were found between the three conditions. Specifically, animacy may be the most relevant to our study and its scores are as follows: low stiffness condition: \(M=3.10\), \(\mathrm{SD}=0.726\); high stiffness condition: \(M=2.92\), \(\mathrm{SD}=0.926\); variable stiffness condition: \(M=2.95\), \(\mathrm{SD}=0.550\). No significant difference was found between the three conditions (\(F(2,19)=0.112\), \(\mathrm{p}=0.894\), \(\eta^{2}=0.012\)).
Fig. 11.
6.3 Perception of Stiffness Change
In condition-3 (variable stiffness), participants were asked if they perceived a change in stiffness during the session. Three out of seven participants perceived a change in resistance. However, during the manual stiffness check period, all seven participants reported that they perceived a change in resistance.
6.4 Relation to Pet Ownership Experience
Fourteen out of 22 participants had experience with pet ownership, but we did not find a clear relationship between this experience and other results.
7 Discussion
The results showed a significant decrease in participants’ tension after experiencing the robot's rubbing motions. However, there was no significant difference between the three conditions. A possible reason may be due to the imperfect manipulation: As mentioned in Section 6.3, not all participants fully perceived the stiffness change in this study. This suggests that the stiffness change during the rubbing motion may have needed to be increased. However, during the manual stiffness check period, all seven participants reported that they perceived a change in resistance. This suggests that participants were able to perceive changes in stiffness depending on the contact position and method, even in the current motion.
With this in mind, the rate of tension reduction in the variable stiffness condition tended to be greater than in the constant stiffness conditions. The mechanism allows the neck more room to move in the low stiffness condition than in the high stiffness condition. When the robot rubs against the human in the low stiffness condition, the robot head behaves as if it is left behind the neck flexion as a result of friction with the human body surface. In contrast, in the high stiffness state, the robot head moves in conjunction with the neck flexion. In the variable stiffness condition, both movements were performed, which may have caused a slight disturbance in the contact cycle, leading to a decrease in the participants’ concentration and tension. We speculate that the irregularity in contact perception may have reduced the user's tension, even if a change in stiffness was not fully perceived.
The Godspeed results (impressions of the robot) also show no clear difference between the three conditions. However, during the free evaluation period, the following comments were made by some participants who perceived the change in stiffness:
“I thought the force was changing and seemed real; so, this behavior was the most relaxing.” “This movement felt lifelike and natural because it had stronger and weaker forces compared to the other conditions.”
These comments suggest that lifelike perception may be enhanced by perceiving a change in stiffness. In terms of lifelike perception, another participant made the following comment:
“The behavior of the robot falling toward me and rubbing its head against me gave me the feeling that the robot depended on me. Because of this feeling, I thought the robot was alive.”
The comment also introduces the unique rubbing experience of the robot. During the free evaluation period, the robot was placed on participants’ laps to perform various movements and they were asked if the weight and size of the robot were appropriate and if holding the robot was more comfortable compared to placing the robot on a desk. Most participants said that they found the robot more comfortable when it was on their lap than when it was placed on a desk. A possible reason for this may be that by placing the robot on their lap, participants felt the weight of the robot and perceived the robot as a calming presence. In addition, the robot was more unstable when placed on the lap than when placed on the desk. Therefore, participants may have been more aware of supporting the robot and focused more on the robot. This is reminiscent of the care-receiving robot [30], which was designed to be cared for by humans, resulting in a positive outcome for them.
8 Limitations
A limitation of this user study was the lack of a control condition, i.e., because we did not compare our bunting robot with other methods, we could not really identify the main source of tension reduction. Therefore, we see that the results of the current study show the feasibility of a robot bunting. To claim its exact effect on humans, additional studies with control conditions are needed. Nevertheless, we believe that reporting this first attempt to replicate bunting behavior with a robot and the results of the user study can contribute to the robotics community.
The participants in the user study were university students from a wide range of backgrounds, and no lab members participated. Other than that, we did not see any specific common background among the participants, but the results may have been biased in this demographic range.
As mentioned repeatedly, in this study, more than half of the participants in the variable stiffness condition could not fully perceive the change in stiffness. We speculate that one reason for this problem was the driving principle of the flexion mechanism. The flexion mechanism was driven by pulling down on the drive wire in the direction in which flexion was desired. Since the wires used in this mechanism are made of metal with low elasticity, when they are pulled, they stretch without deforming, and the rigidity of the entire mechanism increases. In other words, when the wire is stretched, the rigidity increases. Therefore, regardless of the stiffness condition, the stiffness during rubbing was slightly higher, which may have made it difficult for participants to perceive the change in stiffness. To overcome this problem, we need to further investigate the best balance between using elastic and rigid wires, or consider more dynamic control of the wire tension.
In this study, contact with the robot was determined by participant reports and experimenter observation. In the future, we may need to incorporate sensors into the robot to more precisely control the contact and its position.
9 Conclusion
As an interesting topic in physical HRI, this article discussed robot bunting, inspired by an animal behavior in which the robot rubs its head against the user. The ultimate goal is to have a positive effect, such as tension reduction, on the user. To this end, the article reported on the development of two robot prototypes. Based on the results and feedback from the first prototype, the second prototype introduced a flexible neck that can change its stiffness. Although the results of the user study did not show a statistically significant difference between the three different stiffness conditions, a significant decrease in the user's tension was observed after experiencing the robot's head bunting. However, due to the lack of control conditions, we have not yet been able to accurately identify the true source of this tension reduction. In addition, the VSM discussed in this article can be further improved; however, some participants did perceive the change in stiffness and their comments show the potential of robot head bunting.
References
[1]
Yuga Adachi and Fumihide Tanaka. 2022. Development of a Snuggling Robot that Relieves Human Anxiety. In Proceedings of the 2022 ACM/IEEE International Conference on Human-Robot Interaction, 662–664.
[2]
Marian R. Banks, Lisa M. Willoughby, and William A. Banks. 2008. Animal-Assisted Therapy and Loneliness in Nursing Homes: Use of Robotic versus Living Dogs. Journal of the American Medical Directors Association 9, 3 (2008), 173–177.
[3]
Sandra B. Barker and Aaron R. Wolen. 2008. The Benefits of Human-Companion Animal Interaction: A Review. Journal of Veterinary Medical Education 35, 4 (2008), 487–495.
[4]
Christoph Bartneck, Dana Kulić, Elizabeth Croft, and Susana Zoghbi. 2009. Measurement Instruments for the Anthropomorphism, Animacy, Likeability, Perceived Intelligence, and Perceived Safety of Robots. International Journal of Social Robotics 1, 1 (2009), 71–81.
[5]
Danika Passler Bates, Skyla Y. Dudek, James M. Berzuk, Adriana Lorena González, and James E. Young. 2024. SnuggleBot the Companion: Exploring In-Home Robot Interaction Strategies to Support Coping with Loneliness. In Proceedings of the 2024 ACM Designing Interactive Systems Conference, 2972–2986.
[6]
Danika Passler Bates and James E. Young. 2020. SnuggleBot: A Novel Cuddly Companion Robot Design. In Proceedings of the 8th International Conference on Human-Agent Interaction, 260–262.
[7]
Pinar Boyraz Baykas, Ertugrul Bayraktar, and Cihat Bora Yigit. 2020. Safe human-robot interaction using variable stiffness, hyper-redundancy, and smart robotic skins. In Service Robotics. Volkan Sezer, Sinan Öncü and Pınar Boyraz Baykas (Eds.), IntechOpen, 1–5.
[8]
Alexandra C. Behnke, Kristyn R. Vitale, and Monique A. R. Udell. 2021. The Effect of Owner Presence and Scent on Stress Resilience in Cats. Applied Animal Behaviour Science 243 (2021), 105444.
[9]
Junho Choi, Sunchul Park, Woosub Lee, and Sung-Chul Kang. 2008. Design of a Robot Joint with Variable Stiffness. In Proceedings of the 2008 IEEE International Conference on Robotics and Automation, 1760–1765.
[10]
Monika Eckstein, Ilshat Mamaev, Beate Ditzen, and Uta Sailer. 2020. Calming Effects of Touch in Human, Animal, and Robotic Interaction–Scientific State-of-the-Art and Technical Advances. Frontiers in Psychiatry 11 (2020), 555058.
[11]
Aubrey H. Fine (Ed.). 2010. Handbook on Animal-Assisted Therapy: Theoretical Foundations and Guidelines for Practice (3rd. ed.). Academic Press.
[12]
W. Graf, C. de Waele, and P. P. Vidal. 1995. Functional Anatomy of the Head-Neck Movement System of Quadrupedal and Bipedal Mammals. Journal of Anatomy 186, Pt 1 (1995), 55–74.
[13]
Groove X, Inc. 2023. LOVOT. Retrieved 22 October 2023 from https://lovot.life/
[14]
Tae Myung Huh, Yong-Jai Park, and Kyu-Jin Cho. 2012. Design and Analysis of a Stiffness Adjustable Structure Using an Endoskeleton. International Journal of Precision Engineering and Manufacturing 13 (2012), 1255–1258.
[15]
Kondo Kagaku Co., Ltd. 2013. KRS-3302 ICS. Retrieved 5 November 2023 from https://kondo-robot.com/product/03146
[16]
Boris M. Levinson. 1984. Human/Companion Animal Therapy. Journal of Contemporary Psychotherapy 14 (1984), 131–144.
[17]
Zhongyi Li, Weihai Chen, Jianbin Zhang, Qihang Li, Jianhua Wang, Zaojun Fang, and Guilin Yang. 2022. A Novel Cable-Driven Antagonistic Joint Designed with Variable Stiffness Mechanisms. Mechanism and Machine Theory 171, 104716 (2022).
[18]
Jill D. Mellen. 1993. A Comparative Analysis of Scent-Marking, Social and Reproductive Behavior in 20 Species of Small Cats (Felis). American Zoologist 33, 2 (1993), 151–166.
[19]
T. E. Milner and D. W. Franklin. 1998. Characterization of Multijoint Finger Stiffness: Dependence on Finger Posture and Force Direction. IEEE Transactions on Biomedical Engineering 45, 11 (1998), 1363–1375.
[20]
Bruce R. Moore and Susan Stuttard. 1979. Dr. Guthrie and Felis domesticus Or: Tripping over the Cat. Science 205, 4410 (1979), 1031–1033.
[21]
N. Jøranson, I. Pedersen, A. M. M. Rokstad and C. Ihlebæk. 2015. Effects on Symptoms of Agitation and Depression in Persons with Dementia Participating in Robot-Assisted Activity: A Cluster-Randomized Controlled Trial. Journal of the American Medical Directors Association 16, 10 (2015), 867–873.
[22]
Yong-Jai Park and Kyu-Jin Cho. 2013. Design and Manufacturing a Bio-Inspired Variable Stiffness Mechanism in a Robotic Dolphin. In Proceedings of the International Conference on Intelligent Robotics and Applications, 302–309.
[23]
Patricia Pendry and Jaymie L. Vandagriff. 2019. Animal Visitation Program (AVP) Reduces Cortisol Levels of University Students: A Randomized Controlled Trial. AERA Open 5, 2 (Apr. 2019), 1–12.
[24]
Ingo Rieger. 1979. Scent Rubbing in Carnivores. Carnivore 2 (1979), 17–25.
[25]
S. Shiloh, G. Sorek, and J. Terkel. 2003. Reduction of State-Anxiety by Petting Animals in a Controlled Laboratory Experiment. Anxiety, Stress, and Coping 16, 4 (Dec. 2003), 387–395.
[26]
Wenjun Shen, Guilin Yang, Tianjiang Zheng, Yi Wang, Kaisheng Yang, and Zaojun Fang. 2020. An Accuracy Enhancement Method for a Cable-Driven Continuum Robot with a Flexible Backbone. IEEE Access 8 (2020), 37474–37481.
[27]
Masahiro Shiomi, Hidenobu Sumioka, and Hiroshi Ishiguro. 2020. Survey of Social Touch Interaction between Humans and Robots. Journal of Robotics and Mechatronics 32, 1 (2020), 128–135.
[28]
Elisa J. Sobo, Brenda Eng, and Nadine Kassity-Krich. 2006. Canine Visitation (Pet) Therapy: Pilot Data on Decreases in Child Pain Perception. Journal of Holistic Nursing 24, 1 (2006), 51–57.
[29]
Stratasys Ltd. 2023. F120. Retrieved 5 November 2023 from https://support.stratasys.com/en/printers/fdm-legacy/f120
[30]
Fumihide Tanaka and Shizuko Matsuzoe. 2012. Children Teach a Care-Receiving Robot to Promote Their Learning: Field Experiments in a Classroom for Vocabulary Learning. Journal of Human-Robot Interaction 1, 1 (2012), 78–95.
[31]
Jane H. Thompson. 2024. Cow Cuddle Therapy: An Innovative Approach to Emotional and Physical Well-Being through Animal-Assisted Therapeutic Healing. Independently published.
[32]
Kanji Tokuda. 2011. The Validity of Temporary Mood Scale. Ritsumeikan Journal of Human Sciences 22 (2011), 1–6.
[33]
K. Wada, T. Shibata, T. Saito, K. Sakamoto, and K. Tanie. 2005. Psychological and Social Effects of One Year Robot Assisted Activity on Elderly People at a Health Service Facility for the Aged. In Proceedings of the 2005 IEEE International Conference on Robotics and Automation, 2796–2801.
[34]
Ian D. Walker. 2013. Continuous Backbone “Continuum” Robot Manipulators. International Scholarly Research Notices 2013, 726506 (2013), 1–19.
[35]
Sebastian Wolf and Gerd Hirzinger. 2008. A New Variable Stiffness Design: Matching Requirements of the Next Robot Generation. In Proceedings of the 2008 IEEE International Conference on Robotics and Automation, 1741–1746.
[36]
Motoki Yasuda, Arisa Ota, and Fumihide Tanaka. 2021. Development of a Variable-Softness Robot by Using Thermoresponsive Hydrogels for Haptic Interaction with Humans. In Proceedings of the 2021 IEEE 4th International Conference on Soft Robotics, 254–260.
[37]
Naoya Yoshimura, Yushi Sato, Yuta Kageyama, Jun Murao, Satoshi Yagi, and Parinya Punpongsanon. 2022. Hugmon: Exploration of Affective Movements for Hug Interaction Using Tensegrity Robot. In Proceedings of the 2022 ACM/IEEE International Conference on Human-Robot Interaction, 1105–1109.
[38]
Yukai Engineering Inc. 2020. Qoobo. Retrieved 26 August 2023 from https://qoobo.info/index-en/
Index Terms
- Development of a Robotic Device that Performs Head Bunting to Relieve User Tension
Recommendations
Tension reduction method for a modular cable-driven robotic arm with co-shared cables
AbstractThis paper focuses on the tension reduction method for a Modular Cable-Driven Robotic Arm (MCDRA) that consists of a number of 2-degree-of-freedom (2-DOF) cable-driven joint modules. The MCDRA has implemented a cable routing scheme in which every ...
Development of a robotic finger with an active dual-mode twisting actuation and a miniature tendon tension sensor
2016 IEEE International Conference on Advanced Intelligent Mechatronics (AIM)A robot finger is developed with enhanced grasping force and speed using active dual-mode twisting actuation, which is a type of twisted string actuation. This actuation system has two twisting modes (Speed Mode and Force Mode) with different radii of the ...
A soft gripper with variable stiffness inspired by pangolin scales, toothed pneumatic actuator and autonomous controller
AbstractAiming at combining compliant covering and rigid lifting to the object grasping, this paper presents the design principle of a variable stiffness soft gripper and carries out its structural design, gripper fabrication and controller ...
Comments
Information & Contributors
Information
Published In
June 2025
312 pages
Copyright © 2024 Copyright held by the owner/author(s).
This work is licensed under a Creative Commons Attribution International 4.0 License.
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Published: 24 December 2024
Online AM: 14 October 2024
Accepted: 26 September 2024
Revised: 07 August 2024
Received: 19 December 2023
Published in THRI Volume 14, Issue 2
Check for updates
Author Tags
Qualifiers
- Research-article
Funding Sources
- JSPS KAKENHI
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 1,361Total Downloads
- Downloads (Last 12 months)1,361
- Downloads (Last 6 weeks)1,279
Reflects downloads up to 02 Feb 2025
Other Metrics
Citations
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in