(a) aSpire, a clippable pneumatic-tactile feedback device with 3 soft actuators. (b) Control UI that allows users to select/create different tactile patterns. (c) User test of aSpire on passengers in on-road commuting environment. aSpire clipped on; (d) a seat-belt for breathing guidance and providing comfort for vehicle passengers, (e) a back pack strap during walking.

 1. We implemented a mobile GUI (Graphic User Interface) for allowing users to personalize their own tactile patterns delivered by the aSpire and control it.

2. Evaluation on its tactile intensity control system has been done

3. User study protocol for evaluating the aSpire system in various daily activities (walking, driving, desk work) other than commuting in cabin as a passenger has been submitted to the institutional review board and planning to run it at HMC R&D center.

4. Thanks to the member collaboration with HMC (NVH Research Lab, PI: Jinmo Lee), we deployed the aSpire on a Hyundai toy car for mitigating stress and anxiety of pediatric patients waiting for getting a surgery in SJD Barcelona Children's Hospital (Project Video in YouTube)

Thank you for visiting the project page of aSpire!

Hi! I'm Kyung Yun Choi (please just call me Yun), first year PhD student at the Tangible Media group and PI of the aSpire project. My background is Mechanical Engineering (BS) and Aerospace Engineering (MS), and my research efforts revolve around creating a new tangible interface that extends the space of user experience by enabling the transmission of emotional and tactile information.

I am excited to introduce our collaboration project with ML member company Hyundai Motor Company (PI: Jinmo Lee), the Affective Computing group (Dr. Neska El Haouji, Prof. Rosalind W. Picard) and Jessie Wang (undergraduate student in M.E), which was initiated to provide physiological and physical comforts to passengers in a future self-driving car. However, we are not only constrained in the on-road car application, but also have been exploring the potential deployment of this research in daily activities, as shown in the figure above as a few examples. 

Let me explain more details below; 1) how we started to come up with this idea,  2) how we implemented this compact and mobile tactile feedback system, 3) how we evaluated this system, and conducted a user study and its result, and 4) lastly, potential applications that could be benefit from using aSpire in daily routines. 

Growing Attention on Improving Mental Health by Enhancing Breathing Experiences

Regular practice of slow and deep breathing has gained lots of attention thanks to its health benefits [33] such as enhancing ventilation efficiency of respiratory system, heart rate variability, and lowering of blood pressure [22, 63, 27, 49].  Breathing exercise also has physiological benefits such as reducing mental stress, anxiety, and depression, and promoting a generalized state of relaxation [27, 7].  Since breathing provides an efficient way to voluntarily and indirectly control our physiological state, including heart rate and heart rate variability [27, 65], it has been known as the oldest stress-reduction technique [15]. Various types of techniques that involve slow, deep breathing such as yoga, Tai Chi, and some forms of meditation, have been practiced throughout history to increase well-being [61, 41].

A growing number of studies in Human-computer Interaction (HCI) have developed and explored new interfaces to assist people to regain and sustain attention to their inner body through breathing [46]. 

However, despite making many advances, most of the devices lack mobility and a convenience that enables them to be deployed and tested in natural daily-life environments. Also, their studies have neither examined pneumatic-tactile feedback for guiding breathing rate (BR) control, nor have they evaluated effectiveness to allow people to have various options for getting tactile biofeedback. Lastly, it is difficult to find a tactile biofeedback system that provides adaptive intensity to adjust to different body shapes and user’s preferences, due to the lack of feedback control ability.

To promote breathing exercises that are convenient to test in daily life, we introduce aSpire, a mobile tactile feedback device that enables users to control their BR while maintaining a high energetic and pleasant level, without evoking stress.

 Contributions

• We explore a pneumatic-tactile feedback system as an effective and engaging tactile guidance for effortlessly regulating breathing rate (BR), and introduce a clippable design which turns a commonly-worn belt/strap (e.g., waist-belt, seat-belt, straps of backpack and cross-bag) into a tactile feedback interface.
• We present a customized stretchable pressure sensor and its simple and cost-effective fabrication process, which enabled us to implement a closed-loop pressure feedback control to provide adjustable and uniform contact pressure regardless of different body shapes.
• We introduce a compact pneumatic control system that individually controls the shape deformation of the soft actuator with various speed using a single motor pump, which enables the system to guide both breathing phase and rate.
• We conducted a user study in a real-world commuting setting by clipping aSpire on a seat-belt to evaluate its engagement, acceptance level, and perception by passengers riding in a car on open roads.

Design Goals

Our goal is to develop a non-intrusive tactile feedback device that guides users to regulate their BR during daily activities without inducing extra mental stress. We focused on the tactile modality as
opposed to other sensory modalities for influencing BR. To achieve our goal, we have set the following design requirements:

1.  Personalizable: The device should be able to provide users with ability of adjusting the tactile stimulation intensity and pattern to accommodate their different BR, preference
   and sensitivity. 
2. Comfortable: The tactile stimulation from the device should not evoke any physical discomfort and mental stress; it should be close to organic with a non-machine-like comfortable
   feeling. 
3. Mobile: The device should be able to be carried or worn on the users’ body and its location on the body should be easily adjustable.

Implementation

aSpire consists of three soft air pouch actuator modules and a pneumatic control system module. Each actuator is paired with a 3-way solenoid valve. The DC motor air pump  is connected to two solenoid valves that control the open/close of the outlet and inlet of the motor. Based on these 5 solenoid valves’ ON/OFF state combination as shown in the state table (figure below (b)) and its ON/OFF duration, the system can individually control the deformation speed of
the actuator using a single motor. 

Hardware Specification of the aSpire

The maximum speed of the DC motor air pump is 18.5 ml/sec for inflation, and 21.4 ml/sec for deflation. This enables the system to deliver up to 15 bpm with various tactile patterns. When the motor was running at the maximum speed, the noise level was measured as 62 dB from the 1 cm distance, and 38 dB from the 30 cm distance. This maximum noise level is relatively low given the noise from a conversation in a restaurant or office may be around 60 dB. When the motor was running at the average speed, the noise level was measured as 43 dB from the 1 cm distance while it was 28 dB from 30 cm distance (ambient noise was 24 dB). The maximum normal contact force that the actuator can handle was measured as 17.7N. The dimension of the single soft air pouch actuator is 58W x 67L x 15H mm. The dimension of the pneumatic system module is 99W x 67L x 38H mm. The total weight of the device is 252 g. The silicone we used for fabricating the sensor and actuator has a 900% elongation at break, which led us to achieve a maximum z-direction (height) strain of 32 mm, and the maximum volumetric displacement of 40 ml for inflation and 17.5 ml for vacuuming from the default state. At the maximum-inflated state, the pressure inside the actuator goes up to 1054.57 hPa on average (10.6 Mohm) where the average atmospheric pressure was 1010.89 hPa measured from the barometric sensor. When the motor fully vacuums the actuator, the pressure of the actuator drops down to 872.73 hPa (490.8 Kohm).

Stretchable Pressure Sensor for the Closed-loop Pressure Control of the Soft Actuator 

To develop the device compact and mobile, we created a stretchable pressure sensor that can be integrated as a part of the actuator material, instead of introducing an extra barometric pressure sensor tethered to another air channel tube branching from it connecting between actuator and pump, which is commonly used in a pneumatic system. We embedded the sensor system on the air pouch actuator membrane that responds to the volumetric deformation of the actuator by varying its resistance value.

Technical Evaluation: Customized Stretchable Pressure Sensor Response to Various Force Inputs

We ran the inflation/deflation cycle of the soft air pouch actuator with 12 bpm speed for 160 min to have it run for 1920 cycles for a durability test of the stretchable sensor and for inspecting the peak-to-peak value variance. After running the 1920 cycles, the barometric pressure sensor’s the peak-to-peak value root-mean-square error (RMSE) was 8.86% while the stretchable pressure sensor had the RMSE of 7.69%. We used the data acquired from the cycle test for a sensor calibration. 

Animations shown below present the sensor response to different force inputs.

User Study

Most people in the U.S spend 43 - 53.2 minutes on average per day on the road to commute in a car or train [29] and this average commute time has kept growing since 2009 [8]. Reflecting on this trend, we chose to evaluate the potential real-world daily situation of using aSpire while riding in a car.

We drove a four-door midsize sedan on a designated round-trip driving route alongside a river in Cambridge. The driving route was a two-lane road with a speed limit of 35 miles per hour and 9 mile distance in one way, which took 30 minutes of driving. We divided the route into four segments to test four different conditions of the aSpire operations for 5 min each. We conducted the study between 10 AM to 4 PM on weekdays and weekends to avoid the rush hour traffic for minimizing the uncontrolled variables.

A total of 18 participants ranging from 20 to 38 years old (M = 28:56 years, SD = 5:06 years) took part in the study, including 8 females and 10 males. The first three participants were pilot testers.

In summary, 

1.   80% of the participants were able to decrease 24.8% of their average BR when they were experiencing the  sequential tactile feedback pattern (TS_SQ) compared to the average baseline extracted from the TS_N condition., while reporting the same energy and pleasantness levels and without any significant extra stress. 
2.  40% of participants preferred the TS_SQ mode compared to 33% who preferred the TS_SC. 
3. 20% of participants were able to make the immediate connection between breathing control and the device's purpose before disclosing the information. 
4.  93% of participants reported that they would recommend using the device for stress relief and meditation/yoga practice. 
5. The electrodermal activity (EDA), which was measured for verifying if there was an effect of the device usage on the users' affective arousal level, suggests that the synchronized tactile stimulation pattern (TS_SC) is the least arousing. When comparing the total number of peaks for the four segments, no significant difference is reported (Friedman chi-squared = 5.36, df = 3, p-value = 0.15).

Remarks

Eight participants commented on the feelings and emotions induced by the use of the device, four of
them reported the feeling of a pet or person presence, while four reported more focus, calmness and more comfort when using the device in the seat belt. 

One participant highlighted the usefulness of such device attached to the seat belt to appreciate the wear of the seat belt while it is usually uncomfortable to put. 

Three participants comments can be classified related to the use of the device, two recommended using it while driving and another as a device to help kids to sleep. 

One participant reported that it would be more relaxing to have warmth from the device while getting the tactile stimulation which could make this feeling more similar to hugging someone. 

Only one reported a remark related to the design, wishing to have the device along the whole seat belt, proposing then to enlarge the contact surface with the body.

Applications

Commuting in Cabin

A 73% of participants reported that the aSpire would be effective for stress relief, we think that applying aSpire for a car or bus seat-belt, or perhaps an airplane seat-belt (this would need more testing because of the lap position), could enhance comfort and relaxation during travel and commutes. Some visions of the future include autonomous cars with the potential of the car cabin becoming a place for greater relaxation and well-being. Since most participants perceived the sequential tactile pattern as massaging, and the synchronized tactile pattern as soothing, aSpire may help enable such a future vision.

Given the device’s demonstrated potential to support reaching a lower goal breathing rate (GBR), we think that setting a higher GBR might also be effective. A few participants reported that condition (sequential tactile stimulation pattern) felt like awakening them, and the work by Russo et al.  presented benefits of increasing the BR as alerting and increasing focus. Thus, future work should evaluate the potential of aSpire’s programmability, using different settings, to help users achieve different types of desired benefits.

Potential Application of Enhancing the Connectedness over Distance

As we currently developed two aSpire devices, we hope to explore the potential usage which is for enhancing communication over distance/in a same space via tactile feedback. aSpire could be used for enhancing the physiological and emotional connectedness of a couple who are communicating through video conference while walking on a road/staying in a room by delivering each other's BR phase and rate.  Not only for this physiological connectedness, but the aSpire could be used as a playful and interactive tactile communication tool- as TMG showed several examples such as inTouch, and PegBlocks- via distributed aSpire's symmetric/asymmetric connected inflation and deflation motions in response to users' touch and pushing force.  Detailed concept of physiological connected ness via biofeedback interface can be found in my MAS thesis (BioResonant Interfaces: Tangible, Subliminal Biofeedback to Regulate Physiological States, MIT) page 62 - 65.