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 Content on this website may be outdated, and is without guarantee.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)


Akarsh Sanghi

Interaction Designer, Developer
Laura Mul

Shinichiro Ito

Creative Technologist, Media Producer



Wrapping up the project

Last week the final presentations took place. First we presented our concept after which all visitors could come to our stand. Here, people could try out the necklace, watch the video, look at our poster and smell various scents. These were the scents that were used during the experiment. The necklace that could be tried on was the one made from felt which is not the final design, but was the only working prototype that we had.

foto 3

foto 1

foto 2

foto 4

Scent experiment

A scent experiment was conducted to find out what scents can trigger certain emotions. The aim of the experiment was to find out what emotions are evoked when participants smell 11 different scents. Another aim is to find out whether the different scent experiences are connected to the feeling of safety


These are the scents that were used. Water is used as a control scent:

  • Neroli (natural)

  • Lemon grass (natural)

  • Mandarine peel (natural)

  • Pheromones

  • Mixed one

  • Water

  • Mint (natural)

  • Apple (natural)

  • Cappucino (natural)

  • Strawberry (synthetical)

  • Vanilla (synthetical)


6 people participated of a mean age 41.


Step 1

  • Let the participants fill in the form in which they describe how much they smoke and whether they have respiratory problems. This is done, because it might influence the results. Also, participants will be told that their answers are not used for other purposes than the experiment. They were also asked about where they grew up, since their memories can easily influence their experience of scent.

Step 2

  • The participant will smell the first bottle. Each bottle is labeled with different numbers like 1, 2, 3, 4 and 5, so the participant doesn’t know what is in the bottle. All jars will have the same colour, so it will not have influence on what the participants smell.

Step 3

  • After the first bottle is smelled, participants are asked to show how much they feel 12 emotions by using a specially designed tool. Then these questions are asked:

    • What is the emotion connected to this?

    • What do you associate the scent with?

    • Do you think of a place or experience?

Step 5

  • This is repeated for all the smells



  • Independent variable – smells.

  • Dependent variable – how people experience smells.

  • Control variable – smell in room.

After the experiment, some questions about how safe one feels are asked. This is done to find out if someone’s scent experience is connected to someone’s feeling of safety.



Calibration for you.

From the last time, we improved the mechanism.

The improvements:

・To add an external battery for Arduino

・To add a switch to turn the Arduino on and off

・To activate the LED on the board as a visual feedback

・To add the calibration function and button
Because of some trouble, I changed the Arduino Nano to Arduino Micro.

I explain about how we calibrate for every user.

Logic flow:

1 | When the user pushes the calibration button first time, it registers the good posture value (angle).  At that time the Arduino registered the X/Y/Z values from sensor as good posture value.

2 | When the user pushes the calibration button second time, it registers the bad posture value (angle).  At that time the Arduino registered the X/Y/Z values from sensor as a bad posture value.

3 | After 1 and 2, it compares the absolute values of the difference between good posture value and bad posture value of X/Y/Z.

4 | It chooses one axis from X/Y/Z which has the biggest absolute value of the difference.

5 |  If  the absolute of the difference between the current angle of the chosen axis and good posture value is bigger than the absolute of the difference between the current angle of the chosen axis and bad posture value, it fire the filament and LED on the board and emit the smell.

Tomorrow we will try the user testing.


Engineering for Women’s safety

Today I did about engineering part of our necklace.

I succeeded to make the circuit work that we suffered.
The problem was simple. The pin arrangement of the NPN power transistor was different from that they thought.

The lesson is “Check the data sheet.
Here you can see the data sheet of the transistor.

You can see what I did here:

I confirm that aroma part that heats filaments and emits the aroma and sensing part that detect posture are working with Arduino Nano.

This is  the circuit I used this time.

Figure of Circit


・There should be enough wet around filaments. If it isn’t, it doesn’t get smoke.
・I used a tube for filament to keep the aroma in a small area. It gives a strong smell.
・Detecting posture precisely is the next step as sensing part.

Here I took a video when a tube with aroma and filament emits a smell:



Spending consecutive nights trying to make the PostureAroma Necklace work, being awake for long hours, being charged up by constant supply of caffeine, nicotine and indian food, Shubho and I put some serious effort to make this thing functional. We hunted alternate ways to make the sensor work accurately, viewed various forums, youtube videos, blogs etc. to find methods which could give us accurate values in the scenario in which we wanted it to work, i.e. the sensor being placed on the neck (inside the 3d printed case of the necklace). But we never came across any solution because we were making one fundamental mistake of “not keeping the sensor in a perfectly horizontal position”. Inside the 3D printed case of the necklace, the sensor was not always in a static horizontal position which made it impossible to generate a threshold value from it, which could be used to trigger the aroma mechanism. Some things which have to be taken care of while working with prototyping materials like Arduino, Sparkfun Sensors etc. is that they never function in the exact way we want it to work. There are limitations and errors which occur due to certain miscalculations we make during the process which go unnoticed. This makes it even more hard to debug and rectify the error we are making. For example, while working with the sensor initially we were using pin headers with jumper wires attached to the sensor. The pin headers were not perfectly attached with the sensor which in turn was giving us false values on our preview monitor. This error was easily rectified by soldering the pin header to the sensor, a valuable advice given to us by Corne (iFabrica guy).

Once we soldered the sensor and the pin header, we could reduce the noise from the sensor by a large amount but still it was not enough to get an accurate threshold value using which we could trigger the aroma mechanism. Taking Corne’s help, we were able to figure out the circuit for the aroma, but it became very difficult to connect the aroma circuit and the sensor circuit together. The current required or the power required to make both the things work simultaneously was not being achieved without a 9Volt battery. Even with a 9Volt battery we were not sure, that if it would work perfectly or not since, excess power or excess current could also damage the circuit/sensor/filament/arduino. Going through this long and frustrating process without any positive outcomes, Shubho and i decided to start all over again, with an Arduino Uno and test out the sensor with LED’s. In order for the necklace to work, the Arduino and the aroma mechanism should be hooked up with a battery source which is adequate enough to power both, the arduino+sensor and the filament.

Current status:

The Arduino Uno was used with the ADXL 335 Accelerometer, and was tested out using LED’s along different axis. The code was written in a way in which the LED would light up when the sensor was tilted along a particular axis. So currently the LED is working with all the 3 axes. Depending upon the movement of the sensor, it is giving constant readings but at the same time it also gives a lot of garbage values which are of no use. By modifying the code, the Z-axis was figured out and was giving the desired values but the other 2 axes were still giving out random values. Therefor with the LED, the response from the sensor is positive and gives consistent readings as per the tilt of the sensor. In trying to make this process work on the correct power source-all LED’s got burnt. In order for the necklace to work, instead of the LED, the aroma mechanism should be triggered with a tilt in the sensor. But that is still not working because there is not enough current to power the aroma circuit. Therefor in theory, the principle of the project is working, but the actual aroma mechanism is not working due to inadequate power/current supply.

DSCN6969 DSCN6973

Necklace Design getting us into Dutch Movies ??

Making the choice between Brigit and Amber as to who would design our necklace was a tough call to make since both had their pro’s and cons. In the end we chose a designer who suited our specifications of working in an iterative process and evolve the designs as per the continuous changes being made in the mechanical part. Working with Amber has been a very interesting experience for the entire team since we all were on the same page in terms of what we wanted this project to be and how we wanted to execute it. Not being a trained jewellery designer, Amber very sucessfully showed us, how skills learned in one profession/domain can be easily merged and applied to another to create something unique and beautiful. The design of our necklace is a perfect example of “Create by intuition” thought process, since we did not have any decided or finalized design on paper or a CAD model etc. We did take the help of 3D printers to print the cases, but bringing all those plastic objects together in the form factor of a necklace and weaving them together with the right material and color was the real artistic genius behind the design. Going analog in this digital age i.e. doing everything by hand instead of using any machine was a good choice made by Amber because it has brought out the ethnic and raw feel to the necklace.

photo 2photo 1

Meanwhile we also came across a very interesting opportunity of being a part of the Dutch Film circuit. Jette Van den Berg from SPEKTOR had heard about our project and offered us to be a part of the advertising campaign she was conducting for the upcoming movie “KENAU”. Kenau Simonsdochter Hasselaer has been very important part of the Dutch History and her name and legacy lives on in colloquial Dutch. Over the years her name has become a symbol and synonym of female bravery. The advertising campaign being conducted by Jette involves a lot of aspects of women empowerment in the Netherlands. In our meeting with her, she mentioned how it could be interesting to have our necklace as a part of the advertising campaign since the core of our project is also about women, and their self confidence and safety. This meeting boosted the morale of the entire team and sort of reinstated the value and importance of our project in all of us. Since our project is still very much a work in progress, we could not commit anything to her, but if time permits and the necklace works as expected there is a possibility of creating a special edition necklace around the theme of the movie, and making it a part of the campaign.

So guys watch out !!!! There is a chance you might see one of the cast members sporting the PostureAroma Necklace on the red carpet and feeling more safe, confident and empowered……..!!



“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:, 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”            


Choosing a designer (December 6)

The options

After having met Brigit, we went to meet Amber, an student from the Amsterdam Fashion Institute. When we met here she was very enthusiastic and wanted to make something that is down to earth and feels good on the body. This sounded very nice, but the problem was that she never made any jewelleries. After we talked to both designers, we set out the pro and cons for each.

If we would pick Brigit, we would be sure that a great design would be created, since we saw a lot of her work which looked very nice. It would also be done fast, which would mean around 1 week and we know that it will go right, since she is a professional. The cons are that we need to give the specifics about the mechanism very quickly and we can not go into an iterative process. When Brigit would make something, it would be a final design as it will not be easy to edit it anymore. This is because the process of making it consists of using water and soap to make felt out of wool. After the felt has been made, it is stuck and changing it would mean that a whole new design has to be made. This is difficult for us, since we would like to change the design according to the change of the mechanism. The mechanism and design should grow together and during every step the design should be changed a little bit.

If we pick Amber, it would mean that we choose someone that never designed jewellery, but is very creative and has a fashion background. At the moment she is making her collection for the upcoming fashion show, which takes up a lot of time. When we met her, she said that she always prototypes her designs to see if they look good in real life. This is useful for us, because if we do that, we can change the design after we see it in real life. She wants to make a skeleton and then create textiles around it. This way, it feels good on the body and the mechanisms fit in there.

The decision

In the end, we decided to choose Amber, because we can work together with her closely as she can come to the MediaLAB to work. In this way, we can keep on changing the designs and try printing those at Ifabrica. Hopefully this will evolve into a goodlooking design where the mechanism fits into.

Meeting Brigit Daamen (December 3)

During our search for designers, we stumbled upon Brigit Daamen, who is a felt designer. She was very enthusiastic about meeting us, since she makes a lot of spatial felt designs where our mechanisms would actually fit in.

Why did we actually choose a necklace?

When we arrived we started talking about the concept, which provoked some critiques. Why didn’t we choose for a scarf? Why didn’t we choose for a clip on the bag? How would it work if it is mostly for women on the street when the necklace is actually underneath the coat? We found that we couldn’t answer these questions immediately and there were some serious design challenges coming up. This made us think about brooches or extensions of the aroma part of the necklace that could be clipped on one’s coat.

Interesting designs by Brigit


During the meeting, Brigit showed some interesting designs of which one can be seen in the picture on the left. The necklace does not close which would make it easier to wear on a coat. Also, the pendants are made of rubber in the inside with felt on the outside. Our technical parts could go in there. Felt could also be created around objects. This is because the felt is made from wool and adding water and soap will create the felt, which means that around the technical parts, felt can be created.


Influence of felt making process on our concept

One of the problems that comes up here, is that the technical parts are not water proof, so we have to think of a casing to put around it. Also, if it gets surrounded in Felt, we can not open it up anymore to fix the smell mechanism or change batteries. We could choose to work together with Brigit, but we will also meet AMFI students who are also willing to help us. Eventually one of the designers will be chosen.



Ifabrica lasercutting: first print of necklace

3D printing

The necklace doesn’t only have to work correctly, but women must also be willing to wear it. First of all, the design was expected to be based on 3d printing as this would be an innovative approach and it would be relatively easy to put our technical parts in the necklace. Although this sounded nice in the beginning, we changed our minds after having talked with a 3D designer who thought that our instant focus on 3d printing was not obvious. Why not try other materials?




Felt was the next idea as it is easy to use and it feels nice on the skin. Also, it can be very spatial which makes it easy for the technical parts to fit in. As can be read in one of our previous blogposts, we made one case for one of the electronic part that has to fit in the necklace. It would be great to combine the 3D printing to make cases and then put felt around it.



In order to cut the felt, the lasercutter at Ifabrica would be perfect to use. This is why I went to attend the lasercutting workshop.  The evening started with a nice soup where the workshops participants were introduced to each other. The lasercutting workshop was attended by one lamp designer and one woman who wanted to make an interactive aquarium. First a presentation about lasercutting was given followed by the actual 3D printing. We were to test the different amount of speeds and power on a small piece of wood. This could tell us what kind of settings should be used when you want to cut through wood or when you just want to engrave it. After this, I started working with the felt. The workshop teacher already did some felt experiments, so I could use some of her knowledge to estimate the power and speed of the laser. I imported the file of the necklace and hoped that I made it the right size, so it would be wearable. It seemed that that right speed and power were chosen, so I immediately cut the whole necklace. After this, I tried the thinner felt, but this wasn’t fantastic, since the felt was too thin which made it very vulnerable. The actual speed I used for the thicker red felt was 20 and the power was 80. This means that it went quite slowly with a high power.

Time passed quickly: I was still busy lasercutting when we had to stop as it was 9 o’clock. Therefore I only printed two of the same patterns. Hopefully we can come back to print some more.


The design is still a little problem. I talked to the workshop teacher and she found that felt is to be worn by old ladies. I obviously didn’t like this, but this is true as felt can easily become corny. That is why we have a challenge to make it attractive for all ages. The workshop teacher gave me a better felt shop which is called vilta.

Since there are no designers in our team, we are looking for one. In a few days, we will meet Birgit Daams, who is very enthusiastic about our project and makes felt  which are spatial. She might work together with us, but this will be found out during the meeting. Next to that, there are some AMFI (Amsterdam Fashion Institute) students that are willing to work with us.