#pcomp2018

Code

The code I was using with IFFTT was running really slowly and there was a big delay from when I pushed the help button to when I received a text. I had also tried using Twilio, so went back and looked at this code to see if it might be faster. I set up a very simple breadboard version of my circuit so that I could debug faster and see what was going on (taking apart the bracelet was not an option at this point…) I pieced together first the text code using Twilio, then added in the button pressed if statements and finally the accelerometer trigger. This worked much faster!

It was hard to get a video of all of this in action, but here is a video showing the light and text being sent after an activation from the accelerometer and then the timer button being tested:

The biggest challenge here was connecting to the itpsandbox network. I had to register my device with NYU first, but to do so I had to put in the MAC address, but to get the MAC address I had to be connected to the internet. So I did this at home using a script I found online to run with my feather. My friend Tushar also showed my a program called Herbivore that is a packet sniffing software.

Both these methods worked and I went back to ITP and registered my device. But I still couldn’t connect to the network! After talking to Marlon and confirming that my device was in fact registered I went back to my code and realize I had typed in the password incorrectly!

User Test

The bracelet worked as expected during the user test (except when Cameron ran with it on and it triggered a fall, but actually this was also kind of expected…). The users were able to:

Press the help button and see the middle button light up and buzz and receive a text

Press the right button to set a timer

“Fall” and the middle button would light up and buzz and they would receive a text

Based on the feedback I received, the next steps before the final presentation are:

User Experience: Refine/add the communication with the wearer using different LED notifications. For example, when the timer is first pressed, flash a green light  once and when it goes off after 5 minutes flash an orange light quickly for a longer period of time.

Fabrication: 3D print molds and cast buttons. Also buy velcro to make a better attachment mechanism!

Code: Clean up the code to put the longer pieces (such as connecting to Twilio and sending a text) into functions. Also add code for the new user experiences and  add a double pressed timer function.

User testing was coming up this week so it was time to move everything off of the breadboards and start soldering.

Fabrication

I started by making my circuits as tiny as possible on the breadboards to figure out how I could position them on the. I also knew I wanted them all to be roughly the same size - or rather, I knew that the buttons had to be as big as the largest one, so it if one was really big it didn’t matter that the other two were really tiny.

Once I had a design that I thought was as efficient as possible, I placed all of the parts on a perfboard I got from Tinkersphere and drew markings where I wanted to cut. I used an x-acto knife to score the board and then used my wire cutters to snap it apart:

Then I soldered on the parts and tested the circuit to make sure all the parts still worked. The LED had trouble lighting up, so I had to re-solder that one, but otherwise it worked.

I eventually want to 3D print buttons and mold silicone, but to make sure I got the design right I started by making prototypes with cardboard and wood. I first measured my circuits and made sketches of how I could stack and cut rings of cardboard to make the button base:

Then I drew this in illustrator:

At first I got some of the measurements slightly off, so I had to do this a few times.

Cardboard test:

Wood version:

Finally, I soldered all of the circuits together in one string and started gluing on the wood pieces together. The challenging part was lining up the wires. Because my measurements were still a little too tight (I knew this when I cut them…) some of the wood pieces weren’t centered in the end:

My friend Arnab had some left over silicone from his midterm project and let me experiment with it to make the button tops. I cut it using an x-acto knife and glued the pieces together with super glue:

I love how squishy they are!

Finally, I used a hot glue gun to glue on an elastic strip I found in the soft lab to make the bracelet. I put the feather and the accelerometer on the mini breadboards and glued them to a block of wood I found in the scrap bin and then glued on more elastic to make an armband:

Circuits + Design

First, after the playtest I updated and simplified my design. I talked to my grandpa and the feature he would like is a button he can push to set a 5, 10 or 15 minute timer. For my final I am going to focus on this feature and the fall detector and to demonstrate the larger concept of a modular bracelet, in my project presentation/documentation I will plan to show other options of other "beads" that someone could order if they want different functions to make the bracelet personalized. This would come with a charging station and box to change the settings on the bracelet without the need of a computer or touch screen. See my previous post (crossover with fabrication) for more on this! Below are my sketches:

This week and last week I met with Lola to get advice on making buttons and specifically squishy buttons for my bracelet beads. She suggested as a first step that I make my circuit using mini breadboards to figure out the wiring and how small I can make the components. While I was in Seattle over Thanksgiving I went to a huge electronics store and found some mini breadboards to try this with. I also found some very sensitive push buttons that I thought would work well inside the silicone buttons.

She also gave me some silicone to play with!

And showed me some other squish circuits for inspiration/research:

My feather and accelerometer hadn’t come in the mail in time before I left for break so I did this using my Arduino. I wrote some code to get the basic functions working to make the light/buzzer button go off when the other buttons are pressed or with a delay for the timer. I also need to figure out how to program the button on the right to function as a timer that can be pressed once for 5 minutes, twice for 10 minutes and three times for 15 minutes.

I also consulted Arnab who had made a squishy button for his midterm and based off of what he told me I started making designs in SketchUp. I showed these to Lola and she had some suggestions of how to refine them, but before I do that, she suggested that I make my first prototype out of cardboard to make sure I have the sizes right, so that will be my next step this coming week.

Feather

Using IFFTT and following a tutorial online, I got the Feather to send me a text. I modified code from the tutorial and at first I thought it wasn’t working, but I then realized I had changed the wrong delay function when I got a text five minutes later! I fixed this and it worked the way I had expected:

I found out through this process that IFFTT can’t make calls from your own phone, so I’ll eventually need to use Twilio which will be my next step to figure out this week. Then I’ll need to combine it with my accelerometer.

Accelerometer

My Adafruit IMU came in the mail and I finally hooked it up and started looking at the library and data I can get from it. It has an accelerometer and gyroscope (and thermometer!) so there is a library to get absolute orientation, but I found it easier to use the raw data from the accelerometer. After consulting some friends with more experience I started by looking at the raw data for the accelerometer velocity on the serial plot to understand the general patterns.

Then I averaged the X, Y and Z variables and looked at the patterns when I moved it normally and when I dropped it on my lap (I couldn't drop it all the way to the floor because my usb cable isn’t long enough…).

I could see larger spikes when I dropped it, so I edited the code to print out “fall!” when one of these spikes happened. I made the threshold -4 and 4 to start. This worked most of the time when I did the drop test, but there were times when it didn’t detect the fall.

For my enclosure I wanted to make an enclosure to go with my PComp final. For that project, I am making a “smart” bracelet for my grandpa with beads that are designed to have modular functions. At first I was thinking of making a program on the computer that would allow him to reprogram the beads (for example change the “reminder” button from “remind me to do my PT at 11am” to “remind me to water the plants at 5pm”).

But then talking to my grandpa and my PComp professor I started to think that a computer interface wouldn’t solve the challenges that my grandpa currently has with smart devices. Touch screens are hard for him to use and plugging something into a computer and opening the right program can be challenging to navigate.

Designing

I started by sketching some ideas:

I had previously seen boxes on the internet that use the living hinge and I knew I wanted to try to incorporate this, so I looked around for some options that might work and found this box that I liked, had a file I could download and seemed simple enough that I could edit it:

I downloaded the file and in illustrator added the features that I wanted for my grandpa to be able to control the button settings:

I made holes for two dials that can be changed to different settings.

I used icons from the noun project for the light (created by Numero Uno), button (created by emma mitchell) and charging symbols (created by faisalovers).

I’m still not sure about a few of the final design choices, so I made large holes on the sides for potential cables that I will need to charge and program the electronics inside.

I also made three holes to put LEDs through to show how charged the bracelet is (the electronics store I went to in Seattle didn’t have right kind of LED holders I was looking for and I didn’t have time to find others).

Finally, I made a “cradle” on top where my grandpa can lay the bracelet while it is charging.

I was almost ready to laser cut, but then I realized that I needed to resize all of the “puzzle” edge pieces because I had scaled the box and they would no longer fit 3mm wood. Luckily, Emily is very good at illustrator and showed me how to do this relatively quickly

Laser Cutting

I started by trying to cut cardboard, but there was something wrong with the settings and it ended up etching the whole thing. I was running out of time (and patience...), so I decided to go right to cutting it on the 3mm plywood I bought. This actually worked well! I had to run the laser cutter serval times, but it pretty much worked!

I was very nervous that it would break at this point...

When I tried to put it together I realized that some of it didn’t fit together. I had resized the box from the template I downloaded and then adjusted the parts that needed to be the thickness of the wood, but in doing so I had adjusted a few of the pieces the wrong way. 

These are supposed to be the same length!

The next day I went back into illustrator and adjusted a few of the pieces. Then I ran the laser cutter again. This time the pieces fit together, so I found some wood glue in the shop, glued the pieces together and left it to dry overnight:

After it dried, I added the knobs I had found in Seattle and the LEDs to indicate charging. I also found my friend Matt’s LEDs and switch in the junk shelf, so I put these inside to simulate the LEDs lighting up. The last challenge was to find a was to find a way to keep the box closed. I wanted to use magnets, but ran out of time to pick them up. Luckily I ran into my friend Elizabeth who had some velcro and used this with a bracket to create a “latch.” This isn’t very secure, so I think I’ll have to update it for the final version:

A progress update on my final project...

Technical

This week I tested out the core functions of the bracelet, the accelerometer and the wifi mini microcontroller.

Before buying anything, I checked out an 3-axis accelerometer (Adafruit LIS3DH) and Adafruit Feather Huzzah from the shop.

I first worked on getting the Feather to connect to my computer. I followed the instructions on the Adafruit website to get the Feather blink test working:

Then I worked on connecting it to wifi. I had trouble getting it to connect at school, so I brought it home and it worked. I think this may have something to do with the username that is required at school.

I also connected the accelerometer and followed the Adafruit setup instructions. I got it to show data in the serial monitor:

I also found out that it can sense taps!

The next steps will be connecting these two and seeing if I can send the data to the computer through wifi.

I have some questions to answer in the next week:

How do I connect the Feather to NYU wifi?

How do I send signals across the wifi?

What functions does the accelerometer library have that will be useful to me?

Design/Concept

To get ready for playtesting, I spoke to my grandpa about the fall sensor he has and other wearable “smart” devices he has used. He is an engineer and used to work with gyroscopes so I was especially interested to hear his thoughts. I found hearing his perspective very helpful and am planning to share more updates as I’m working on it. Here are the main takeaways from our conversation:

Current fall pendant model

Wears around neck

If you fall it automatically calls a help center

It also has a help button

Dislikes

Bad battery life

Because of this he always leaves it on the charging station and doesn’t wear it

If it has low battery the call won’t go through on the pendant, so the call center calls you on the phone, but if you’ve fallen you won’t be able to get to the phone.

The pendant senses false falls - sometimes it goes off when he is sitting

Thoughts on my project

Seems harder to automatically sense on wrist

Likes the idea of having other functions and indicators

Need a speaker inside

His says: “A fall has been detected an operator will speak with you soon”

Then: “Are you ok?” and he says “Yes, this is a false alarm”

A help button needs to be big and can’t be too hard to push

Also needs to be raised so that you can feel → if an older person falls and loses their glasses they won’t be able to see

Thoughts on technology and wearables generally

Doesn’t like touch screen because it requires two hands to tell the time - you have to lift up your arm and tap it with your other hand

Has another bracelet that counts steps, has the time, date and heart rate

Not wearing it anymore because it didn’t update for daylight savings time automatically and he needs to plug it into the computer

To charge it you have to take off the leather strap - you pull on one end real hard and it comes off of main body and you can plug it into computer to charge

Description:

This is a wearable technology designed with elderly people in mind. It is a “charm” bracelet made of modular silicone balls that can be programmed to have different “smart” functions. The main motivation for this device was to redesign and improve the current fall pendants on the market, but this device aims to go beyond traditional fall pendants - the goal of this technology is to give the user a feeling of both safety and power. Is grounded in Kenyatta Cheese’s Life Affirming technology principals:

Listen - The device will listen to the user  by sensing falls. The device will also have a button as a sensor that the user can push to communicate intentionally with the device.

Amplify - The device will have the capability to call for help if the user falls. It will also have the ability to send other signals out to a computer.

Support - The device will have a website or app system paired with it to support the wellbeing of the wearer. The user will be able to see and track their data and also program the modular component of the device to suit their needs.

Organize - This may be beyond the scope of this project, but I would like to make the devices networked so that they could “talk” and respond to each other.

System plan:

Each sensor or hardware component will be held inside the silicone balls that make up the bracelet. Wiring will be hidden in the connectors between balls.

Because of the small size of the balls, I will need to use small parts. One challenge I foresee is that mini microcontrollers only have 3 or 4 pins, so depending on the number of sensors I use, I may need two of them. The sensors I hope to include are:

Accelerometer

Mini buzzer

Two light indicators

Button

The microcontrollers will communicate this information wirelessly to a computer (or phone) using a bluetooth chip. I will also design a computer a program (website?) that would have two main functions:

Allow the user to program the modular features such as the light indicators, buzzer and and button

Track data from the device and allow the wearer to interact with these data

One example of this functionality could be to have a website where family members can upload photos and when a photo is uploaded the device lights up to let the wearer know that they have a photo message.

Another use could allow the user to program the buzzer to go off at 5pm every day to remind them to take their medication (or share a daily affirmation)!

And many more!

Timeline and testing plan:

Week of:

November 11

Order parts

Learn to mold silicone

Learn to 3D print

Start building responsive circuit with components using Arduino

November 18

Finish responsive circuit with components using Arduino

Make my own silicone balls?

Playtesting

Test what people would want to program the different components to do

What data do they want to see and interact with from the device?

November 25

Try different clips for bracelet fastener 

Make 3D printed “scaffolding” to hold parts inside balls?

Solder pieces together

User test with prototype

Is it responsive in the right way?

Does it engage the user?

November 2

Prototype completed - put components in silicone balls

User testing

Does it work seamlessly with the computer program?

November 9

Over the past year or so, I have developed a manifesto of sorts for the work I am hoping to do at the intersection social science, technology and design. It is a work in progress and I am hoping to refine it at ITP by putting it into action in my larger projects. I’m posting it here to remind myself to keep this at the center of what I make as I brainstorm for my final project:

I believe that the next technological revolution needs to be in governance. Most questions people ask about technology are about how we can be more efficient, have access to more information, or get to places faster, but these are tools, not ends in themselves. Instead, I think we need to ask: how do we design systems that enable our best humanity? Technology is not capable of building just and equitable social systems on its own. That ability and responsibility is ours.

In terms of a more specific PComp final project concept, I am interested in exploring a few concepts/theories and a few different materials/techniques.

Concepts/Theories

I’ve been thinking about the following concepts and would be interested in making a project related to one of them:

1. Spimes

I was captivated by spimes when I first read about them in Design Meets Disability by Graham Pullin. A combination of “space” and “time”, spime is a word for an object that crosses both - the physical aspect can change, but it keeps it’s internal data throughout these two dimensions. To paraphrase Wikipedia, they are virtual objects that have physical incarnations. Or put another way, the concept of a spime is to “embed a story in every object and link all objects together” (from this article).  Bruce Sterling wrote a book in 2005 about this called Shaping Things. QRcodes and previously StickyBits (is this still a thing?) are current examples of technologies that can turn objects into spimes.

For this project, could I build a spime? Or spimes? What would spimes look like in 2018?

2. Invisible Infrastructure

My friend who (also) just started grad school at the Kennedy School sent me a paper on Invisible Infrastructure the other week and while I haven’t read all of it, that term has stuck with me and captured my imagination. The paper hasn’t been published yet, so I probably shouldn’t quote it here, but here is short article that gets at the gist of it. The article is focused on gender equality in developing countries and explains why simply building infrastructure and providing programs isn’t enough to reduce poverty:

[T]here are invisible elements of infrastructure that are often overlooked — a social structure that shapes how the physical infrastructure works....Infrastructure development is not just about structures and steel. It is about adapting to the changing habits and lifestyle that are shifting women’s roles.”

This doesn’t apply to gender alone and in their longer paper, the authors discuss other topics such as education and healthcare. They explain that invisible infrastructure is made up of human and social systems. I think what they are trying to do is create a term that takes into account power - something that the field of design often overlooks.

For the midterm, could I build invisible infrastructures? What could help people “realize their capabilities”? Could I explore social power through invisible infrastructures?

3. Collective Efficacy

In the field of economics/sociology, collective efficacy is a term used to describe the ability that community members have through “social cohesion” to influence or control the behaviors of other community members - for example, create an environment where crime in unlikely to happen:

Collective efficacy is defined as the process of activating or converting social ties among neighborhood residents in order to achieve collective goals, such as public order or the control of crime (Sampson, 2006a; Sampson, Raudenbush, & Earls, 1997). Empirically, collective efficacy has been represented as a combined measure of shared expectations for social control and social cohesion and trust among neighborhood residents. The theory of collective efficacy helps explain one of the most robust findings in criminological research, that crime is nonrandomly distributed across geographic space. Collective efficacy also explains why neighborhood characteristics such as concentrated poverty and high levels of residential turnover are positively related to crime. Neighborhoods vary in their capacity for efficacious action, and Robert J. Sampson, Stephen W. Raudenbush, and Felton Earls argue that this variation explains differences across neighborhoods in levels of crime and violence. -- From Sampson 2010 (Sampson 1997 is also useful to look at for the first publication of these findings).

Collective efficacy is extremely hard to accurately measure (think extensive surveys...). There is also a debate about whether it is situational - someone intervening in a crime -  or enduring - influence is exerted even when a given individual isn’t physically present.

For this project, could I explore or test whether collective efficacy is situational or enduring? Could I make tools for collective efficacy? Could I make a collective efficacy game?

Methods/Materials

In terms of methods and materials, I’d be interested to find a way to incorporate any of the following that are new but also very exciting to me:

Pneumatics/programmable air

Soft “robots”

Modular components

3D printing

Silicone/plastic molding (using 3D printing?!)

Potential Projects

1. Smart Bracelet

My most developed idea is to re-design the fall sensor pendants that many elderly people wear.

I was inspired when I saw Black Panther to turn these into “smart” bracelets with different functions. 

I am imagining a bracelet made of soft, colorful silicone balls that have different parts inside of them for different functions:

These could even be a spime of sorts and could be transferred or connect to other devices. I wonder if there could be a feature to build collective efficacy using the bracelet? Or effectively act as invisible infrastructure?

2. Soft controller

The other idea that I haven’t been able to flesh out as much is making soft controllers. When I saw Amitabh’s presentation on programmable air, I was captivated by the possibility for tactile and tangible interfaces. I’m not sure what I would want them to control yet, but I’m trying to form a more cohesive idea around this and if I can connect it to the three concepts I outlined above.