Right now, the goal is to have a camera to detect if the orange ball is still on the plane- if yes, timer runs/ if not, timer stops.
I have tried a few methods to achieve this but every one of it doesn’t work.
Processing + Macbook Facetime HD Camera
Used the ‘Video’ library of Processing
Problem as per screenshot:
Unfortunately, I am on a macOS Catalina ver 10.15.5https://github.com/processing/processing-video/issues/134
Tried the solution as suggested by ihaveaccount on this above forum but the string doesn’t return as they said it wouldwithout the video library, this eliminates the use of external webcams also (i tried with my DSLR as a webcam)
Processing + IP Camera Since I was hoping for the webcam to be portable anyway, I tried to use an IP Camera. Problem: On my processing sketch, nothing shows up, no error, no video feed show- nothing. IP camera doesn’t show it is being connected to anything either.
Not sure if its because im accessing processing via a Macbook and im using an android IP Camera app?
TouchDesigner + IP Camera
Used the Video Stream in, and with RSTP network protocol, while yes there is a video feed, it is SOOO choppy and pretty much only refreshes when I “pulse” the node
TouchDesigner + Web Camera https://forum.derivative.ca/t/video-device-in-webcam-doesnt-work/12208/13
Once again seems like everyone else with Mac ver after 10.14, webcam doesnt seem to work
I did try with the latest beta ver of Touchdesigner however, it only works when I first start a timeline, once I save the file and close it and then reopen, its just a black screen again.
So now Im stuck. Im thinking maybe a Raspberry PI(since it runs Windows) with a webcam, or a camera module and maybe something along the line of python w OpenCV library, or it will be great if i could have processing on Raspberry PI. But this is a whole new set of problems because first I don’t have a Raspberry PI and just learning to work it is like 😵🤯😩.
Had to lift my whole hand to change the angle of the servo
Couldn’t have my wrist in place and just adjust my hand
After inserting both motors to track both front-and-back and also 360 (in which this case still 180 because using servo), the motors are going CRAZZYYY until I use just 1 finger. Likely because its taking coordinates of all 5 fingers. But, still an issue was even though just 1 finger was stuck out, the leap motion is still trying to its best to get readings of all fingers.
Looks very rigid and tried using a board and a bead to simulate the game for now,
realised UX might be abit confusing. because the bead starts off being still on the board so no point in moving your hand if the goal of the game is to balance the ball on the plane for as long as possible
DC motor might not be the most practical because cant control angle of rotation, only controls the count of rotation- in which needs sensors to send feedback also
Build a more stable structure, have the motors screwed on!
Adjust the code such that if the movement is too small, the values wont be sent over to arduino- to clean up the jerky-ness
Webcam to detect where the ball is , to start and stop timer
Previous pitch: https://oss.adm.ntu.edu.sg/ho0011an/final-project-pitch-build-me-up/ Short recap Basically, I was too fixated on the idea of using weighing scale as the input just because I wanted that specific aspect to tie the project to the theme ‘Distant Body‘. But, overall, the initial pitch didn’t really make sense and I was left off with the question- what more can you do with ‘reflection’? & advised to do experiments first.
So- that’s what I did.
Using TouchDesigner, I tried to move a red line to where there is motion.
Trying to replicate Daniel Rozin’s trash mirror but on a digital screen. This led to a couple of random ideas.
Brief explanations on some ideas Idea 2 Thinking along the lines of trying to create a tangible user interface perhaps where the physical ball is able to move according to the projected graphics or vice versa where the projected graphics changes according to the placement of the ball. Idea 3 Maybe a screen based device and user can manipulate the graphics in ‘3D’ space?
Could be in a form of 2 screens or maybe 1 screen but with acrylic to give a hologram look
What Harvey Moon said: “you don't need disguise or unreal or stype to do the virtual production stuff. you can do it…
Idea 4 Individual modules will move up to replicate the shapes of what is detect on the webcam. EG pin needle toy
With all these ideas, I was further challenged to do something with the ball. To perhaps make it into a game.
So after deliberating more, my final sketch idea is an expansion of Idea 4.
It going to be a game and the only goal is: control the plane, remotely with your palm, such that the ball stays on the plane for as long as possible. Time starts when the hand is detected and stops once the ball drops off the plane. Longest time = Highest on the score board.
L298n driver + Dual/Double Shaft DC Motor + Single Shaft DC Motor
Initially, I figured in my head that I would need to 2 arduinos (1 for the user- stagnant and 1 for the potential love- portable). But it was only half-way through building the project that I realised even if I was to have 2 arduinos, both of them still need to be bounded together (ref to bottom) so there’s no difference in UX even if just used 1 arduino. Which was when my project reduced quite significantly and became hardware heavy instead- led to my problems stemming from hardware management.
First lesson of the day, don’t be rough. In attempt to put the “necklace” over my head, I tore the vibration motor. Trying to put the devices around me also meant I couldn’t conceal a huge bulk of wire and I needed to make a lot of “long” wire to be able to reach from the controller to the user’s neck. Mess.
Decided to make an accessory box instead.
Bringing it forward, an aspect i can look into is, after realising that something doesn’t go according to plan, is to be flexible enough to trial it as something else. In this case, the necklace into maybe perhaps a belt. Working with what I have and improvising along the way.
There needs to be a way for the user to know what the different output means. So say if someone likes your photo, the ring would light up but when the user sees it- what is going to hint them to make that association? (something very simple i can think of is maybe-literally a heart pendant/gem that likes up for a ‘like’ or since the ‘super like’ icon is a star- maybe a motor could move in the direction of how you would draw a star)
Ball is a smart fire extinguisher. Ball was created with the blind in mind. The whole product comes with a sensor, the spray and a refill capsule.
Here’s how it works:
In case of fire, the sensors placed around the house would detect heat and smoke.
It will alert the user via the speaker on the fire extinguisher. It will also guide the user on how to use it.
When the fire extinguisher is activated, the nozzle is able to rotate and automatically find the source of fire using the heat detection camera.
The very obvious advantage of this product is how is it able to make up for the user’s visual capabilities. Through audio and heat detection, Ball is still able to give the user control but at the same time aids the user with this control they have. It could very well have been a vibration/smell output instead of audio, they would all be able to alert the user anyway. But in the context of a blind person who still has to get to the fire extinguisher itself, I felt audio was the smart and better mode because blind users are usually more sensitive to auditory information and use it to locate objects faster.
The mechanics is also easy. Ball sprays with a simple pull of a loop.
Considering that this is only aimed at small initial fires, I think there should be other forms of outputs to alert the user when the fire is too intense/ too close and when they should ditch the fight against the fire. This way Ball is able to maximise the potential of the information it receives to not just become a defence mechanism but also a guide to get out of the situation safely.
I am also not too sure what is going to happen if the fire is behind the user.
In conclusion, I chose to look at this device from a multimodal point of view because I believe it has potential to grow, to collect more data and transmitting them into other various forms of output. Maybe it’s able to somehow tap into the user’s mental perception of the space using audio beacons? Maybe it’s able to give hepatic feedback based on intensity of fire? Nonetheless, I can deny this is an innovative piece of assistive tech, tapping into both “low” and “high-tech”.