MediaLAB Amsterdam is now the Digital Society School! You are viewing an archive of MediaLAB projects. Content on this website may be outdated, and is without guarantee.

Women Safety (NL)

Team

Akarsh Sanghi

Interaction Designer, Developer


Laura Mul

Researcher


Shinichiro Ito

Creative Technologist, Media Producer


Commissioner:

Description

SANTA : SENSOR | CLAUS : CALIBRATION

“Christmas is weird. What other time of the year do you sit in front of a dead tree and eat candy out of your socks? “

Working during the so called “Christmas break” was fun because you get to break away from the regular routine and work from home with a free mind. Specially for our project, since it was going through a phase of technical difficulties. There were a lot of problems that were coming up during the execution of the PostureAroma necklace, specially in terms of the mechanism. If i try to explain in simple terms of what we are trying to execute is essentially :

1) Taking a triple axis accelerometer (ADXL 335) and calibrating it to detect change in movement of the upper body, i.e. shoulders. The change should be to a certain degree which essentially is an improper posture and is slouching.

2) When the sensor detects this change in value, it triggers the aroma mechanism and the aroma is released to warn the user of their bad posture.

Sounds simple enough ???

Its not……..

After 3 weeks of playing around with the sensors, e-cigarettes etc. and when things did not work the way we wanted them to, i decided to take a break and read about the types of bad posture and how we can measure them. So here is a summary of my readings on posture:

Good Posture or Neutral Spine (scientific name)

It is defined as the stance which is attained when the joints of the body are not bent and the spine is aligned and not twisted. This refers to the 3 natural curves that are present in a healthy spine (image):

curves of back

Neutral spine gives rise to the idea of achieving “ideal posture”, which indicates the proper alignment of the body’s segments in such a manner in which least amount of energy is required to maintain a desired position.

Bad Posture:

Bad or poor posture occurs when certain muscles of the back tighten up or while others lengthen and become weak which is often due to one’s daily activities, i.e. sitting in an office chair all day long. There are different factors which have an impact on the posture of a person which include occupational activities (e.g. truck drivers, labourers) or biomechanical factors like force and repetition.

To correctly identify the posture which can be detected via the sensor, it was a good idea to make a preselection of a few Bad Posture’s.

The following are examples of common behaviour that lead to a bad posture.

  • Slouching with the shoulders hunched forward

  • Lordosis (also called “swayback”), which is too large of an inward curve in the lower back

  • Carrying something heavy on one side of the body

  • Cradling a phone receiver between the neck and shoulder

  • Wearing high-heeled shoes or clothes that are too tight

  • Keeping the head held too high or looking down too much

  • Sleeping with a mattress or pillow that doesn’t provide proper back support, or in a position that compromises posture

 

Screen Shot 2014-01-06 at 05.01.07 am

Reading various case studies and statistics(Reference: www.spine-health.com), the above mentioned are some of the most common situations in which people can have a bad posture. Keeping the technical limitations and the duration of the project in mind, we as a team decided to choose “Slouching i.e.shoulders hunched forward” as the bad posture. Now the task at hand was to calibrate and code the sensor in such a way that detects the angle movement from the normal position of the body (specific to every user) and trigger the aroma when the angle crosses a certain threshold value. To achieve this we decided to place the sensor at the neck, which would be inside of the 3D printed cases of the necklace design, and would accordingly detect the values. We had already been working with this theory for weeks and had not achieved many positive results. Doing some further research in posture detection I came across the following conclusions:

1) In almost every posture detection/correction medical device that exists in the market as a consumer product uses the lower spine as the reference point to detect the “bad posture”. Therefor many of those devices come as belts which can be worn around the waist and provide accurate values when the spine is curved more than the natural curve (as mentioned above). Since our solution was a specifically targeted for women and was in the form of a necklace, we had to stick to detecting the posture from the neck which is much more tough since getting accurate values from a sensor like the ADXL 335 is quite a daunting task.

2) Theoretically our solution is very much possible and can be used perfectly for detecting the bad posture, but executing it technically is being difficult because of the following reasons:

  • The sensor is extremely sensitive and gives a high deflection in values even when moved slightly. Therefor being a necklace worn around a woman’s neck it would keep giving values since the necklace is in constant movement while the user is doing any activity.

  • Placing the sensor on the neck made it more tougher to detect the change in angle of the upper body movement. This was primarily due to the reason that the normal deflection in values of the sensor when the user is doing some activity (i.e. in motion etc.) and is slouching was not much. Therefor it became very hard to find the threshold value after which we could trigger the aroma mechanism to correct the posture, since the bad posture was not detected properly in the first place.

  • Placing the sensor on the front part of the necklace (i.e. at the pendant etc.) was also not a very good idea since the front part of the necklace is more vulnerable to random movements etc. Therefor the sensor would give a large deflection of values and finding the threshold value amongst them would be impossible.

Initially while testing the sensor, we were using it in an ideal environment for an accelerometer, i.e. well placed on a breadboard and kept on a table static for a long period of time so the values can be static and it would only give the values when it is tilted in a particular axis and after crossing a certain threshold value. We were easily able to calibrate this and test it out in this particular environment, but as soon as we started testing the sensor by placing it on the neck, the deflection of values increased exponentially and we were not able to find a threshold value, even while standing static for 10 minutes. Therefor calibrating the sensor for usage in the scenario where the user is wearing the necklace was not achieved. In my understanding, ignoring technical details and specifications during the concept phase of the project led us to this situation. As a team if we had given more thought on “HOW” this interesting idea can be executed and prototyped, the status of the project would have been more positive currently. Therefor the biggest learning experience from this phase of the project has been, to always try out things by getting your hands dirty (experimenting and failing in the initial stages of the project i.e. concept phase ) rather than failing at the final steps. Experimenting with ideas in a tangible form  is always a better idea instead of making elaborate plans on paper and feeling content with them. And as they say:

“Plans have a way of going awry”            

                                                                                 -Unknown

Post a Comment

You must be logged in to post a comment.