**5. Experiment design**

communication interfaces of I2C (Inter-Integrated Circuit) or SPI (Serial Peripheral Interface). The received signals are then filtered and calculated to generate the tilt angles of the posture

The social-networked posture training CSPT App subsystem is the main interface to posture training users and their peers. The CSPT App provides a number of functions including gateway of sensor data, posture data processing, fuzzy logics and posture determination, biofeedback initiating, data feeder to the cloud and presentation or rendering the historical and analytic posture information. The CSPT App is also a social networking graphical user interface (GUI) for posture training and information sharing among individuals and their peers. As the key element in the CSPT system, the CSPT App executes and manages many tasks, including signal processing, posture determination, biofeedback and data management to cloud computing. It receives, processes and further transmits posture angles. It determines the good posture and decides to notify when biofeedback is needed. The CSPT App renders the analytic data streams from the cloud, manages identity and access control and does encryption/decryption of the data and user ID, as the platform for chatting, messaging and

The CSPT App is also a notifier for people to receive alerts or warnings so that they can correct the poor posture immediately. When a biofeedback is enabled, the CSPT App plays default or customized sound or music and makes the smartphone vibrate for a short period of time to notify the users to change their bad posture. The smart feature of the CSPT App enables the intelligent detection of wearable devices so that no sound, music or vibration is made when the wearable devices are not attached or out of their operating space like being left on the desk. To avoid annoying, the notification period increases when the bad posture continues.

The friendly GUI of the CSPT App is the core to people. The CSPT App renders the analytic data streams from the CSPT Cloud. People can watch and be aware of the real-time status of their postures and realize how good or poor their head/neck and low back postures are. People can also glance over their peers' posture training performance. Based on the historical data from the cloud, people can explore their analytics, including the percentages of maintaining good postures, total wearing times and average response times. The resolution of time scales spreads from day, week, month, to year. **Figure 5** depicts two screenshots of the CSPT App GUIs for the analytic report in the day (right) and for poor posture (left), respectively.

The light speed advance of mobile technologies always makes the smartphone markets and products dazzling. Backward compatibility is a non-negligible issue in developing smartphone apps, especially to the Android platform. Not every smartphone is fresh new and up-to-date. For young kids and teens, some of them may use cheap-but-obsolete styles or their parents' used smartphones. These legacy smartphones usually use old operating system (OS) and supporting interfaces that are even unable to upgrade. In order to maximize the compatibility to most existing smartphones, the development of the CSPT App subsystem has to consider various and many smartphone models from different manufacturers with different OS and software versions. As compared to the development of other subsystems, it is so tedious and challenging.

monitoring hardware with Eq. (1).

48 Biofeedback

file sharing of social network functions.

**4.2. Social-networked posture training CSPT App subsystem**

To validate the effectiveness of the proposed CSPT posture training framework, we design several experiments considering a group of six teens in a middle school in San Jose, California, USA, whose families are similar in their race of Asian Americans, socioeconomic status, occupational status, family size, housing, geographic location, ethics and morals. The six teens denoted as "C," "E," "GC," "GW," "H" and "L," by the first letter(s) of their names, respectively, are all 8th graders in the school and friends to each other. They study, play and chat in very close proximity to each other of the group.

## **5.1. Experiment method**

Our design of experiments has four objectives as follows:


**Date Good posture (%) Wearing time (min) Scenario C E GC GW H L C E GC GW H L** Oct. 17 65 85 80 78 74 87 60 61 60 63 61 74 S1

Collaborative, Social-Networked Posture Training with Posturing Monitoring and Biofeedback

http://dx.doi.org/10.5772/intechopen.74791

51

Oct. 18 68 88 82 76 77 85 61 64 60 62 61 77

Oct. 20 90 92 95 92 95 99 61 68 68 61 64 82 Oct. 21 85 95 90 93 92 98 61 69 65 60 61 86

Oct. 25 95 100 99 100 99 100 87 151 90 111 120 149 Oct. 26 94 100 100 99 100 100 102 155 121 160 158 156 Oct. 27 99 100 100 100 100 100 115 169 162 176 159 180 Oct. 28 99 100 99 100 100 100 152 192 170 179 167 190

**Table 1.** Experimental results.

**Figure 6.** Comparison of results without and with biofeedback.

Oct. 19 89 95 96 95 90 98 60 72 62 62 62 75 S2

Oct. 24 92 100 98 100 98 100 62 113 85 96 89 127 S3


The following three scenarios are designed for the experiments as follows:

**Scenario S1:** Both the biofeedback and social network functions are DISABLED.

**Scenario S2:** Biofeedback is ENABLED but social network function is DISABLED.

**Scenario S3:** Both the biofeedback and social network functions are ENABLED.

We conducted the experiments in each student's home with the guidance of students' parents. Each student was equipped with a wearable training headset, a training lumbar belt and the posture training CSPT App. They were asked to wear the training headset and lumbar belt for at least 60 min a day. The experiments were carried out from October 17th to 21st and from October 24th to 28th, 2016, detailed as follows:

Scenario S1 was conducted first on October 17th and 18th. Scenario S2 was followed and conducted on October 19th, 20th and 21st. Scenarios S3 was conducted from October 24th to 28th. Before each experiment, all the teens do not know about any details of the three scenarios. For the first 2 days (October 17th and 18th), each teen was asked to wear the headset and belt without knowing anything about biofeedback and social network functions. They were told and became aware of the biofeedback music, voice and vibration in Day 3 (October 19th). On October 24th, the teens were asked to download App's social network function where they can glance at their friends' training scores. During the experiments, all the teens knew who of their peers are participating in the experiments. Teens can share their observations and knowledge to each other during the entire experiment period.
