From Robert-Depot
Jump to: navigation, search

Work page for dissertation project


The dissertation, A Machine for Living In stages intelligent systems in domestic space to study the home as a site of intimate life. Designed to be inhabited, this project examines how technologies are altered through inhabitation. Upending the typical utilitarian concerns of the smart home with optimization and efficiency, I focus instead on the home as a nexus of personal history, desires, relationships, episodic narrative, and bodily praxis. I deploy technologies that purport to represent and analyze the self--speech recognition and synthesis, motion capture, and facial recognition, for example--and engage them with the messy conditions of domestic life. Living within the machine for a number of weeks, I produce a durable record of the home and its inhabitants: human, animal, machine. This material is restaged for viewers as a hybrid experience of recorded media and live interaction.


  1. design/build
  2. inhabit
  3. summarize

To Do

  • indoor positioning system
    • people / inhabitant tracking
  • sensing doors / appliances / objects / actions
  • event logging
  • audio recording system
  • audio plan
  • video plan
  • robot / mechatronc plan
  • tracking plan
  • event logging (house data recorder)


Black Box Recorder

  • airplane / vessel / ship analogy
  • telemetry
  • Flight Data Recorder (FDR)
  • Cockpit Voice Recorder (CVR)

Voice Assistant

  • voice operation
  • always listening in the different rooms


agonist / antagonist

  • lightswitch robots
  • door locks
  • other ways to provide / deny access


lipstick robot

head stain wiper

garbage sensor


Spark Core with OSC

Follow these instructions Particle_Photon


build sc on pi zero

cmake -L -DCMAKE_BUILD_TYPE="Release" -DBUILD_TESTING=OFF -DSSE=OFF -DSSE2=OFF -DSUPERNOVA=OFF -DNOVA_SIMD=ON -DNATIVE=OFF -DSC_ED=OFF -DSC_WII=OFF -DSC_IDE=OFF -DSC_QT=OFF -DSC_EL=OFF -DSC_VIM=OFF -DCMAKE_C_FLAGS="-mfpu=vfp -mfloat-abi=hard -march=armv6 -mtune=arm1176jzf-s" -DCMAKE_CXX_FLAGS="-mfpu=vfp -mfloat-abi=hard -march=armv6 -mtune=arm1176jzf-s" ..

ics43432 I2S audio with raspberry pi

  • download source
  • download loader code
  • makefile:
obj-m := my_loader.o
	make -C /lib/modules/$(shell uname -r)/build/ M=$(PWD) modules
	make -C /lib/modules/$(shell uname -r)/build/ M=$(PWD) clean
  • load kernel module:
sudo insmod my_loader.ko
  • see what modules are loaded
lsmod | grep snd

Electret Capsule with USB Soundcard


Electret tip.jpg

External biased electret mic with stereo 3.5mm jack. This will also work directly with cheap USB audio cards, for raspberry pi, etc.


Test throughput raspberry pi (mic to headphone):

arecord -f DAT | aplay -f DAT

Record and playback on Raspberry Pi:

arecord -f DAT -D plughw:0 | aplay -f DAT -D plughw:1
  • mics in each room

Preamp to Line Level


Play KUOW radio stream

<syntahighlight lang="bash">mpg123 -f 6000 -@</syntaxhighlight>


Start picam

picam -w 1296 -h 972 --alsadev hw:1,0
touch hooks/start_record
touch hooks/stop_record

9:8 with raspivid

raspivid -f -t 30000 -b 25000000 -md 2 -fps 15 -w 1530 -h 1360 -e -o ~/housemachine/data/local/test5.h264

time lapse from FLIR

ffmpeg -framerate 5 -f image2 -i %04d.jpg -c:v libx264 -r 30 -pix_fmt yuv420p time-lapse.mp4

Video playback service on raspbian jessie

  • Create a service file in
  • flir.service:
Description=Plays FLIR footage on loop

ExecStart=/usr/bin/screen -dmS video omxplayer -b --loop --win 0,0,1280,1024 /home/pi/housemachine/data/flir/bed4.mp4

  • enable:
sudo systemctl enable flir.service
  • start:
sudo systemctl start flir.service
  • connect:
sudo screen -r video


cat ~/.ssh/ | ssh pi@audionode.local "mkdir ~/.ssh; cat >> ~/.ssh/authorized_keys"


ssh pi@videonode1.local "sudo poweroff; exit"