Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
Loading tagline ....

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.

user:
pass:
register,


                                   

indoor maps for the blind

on top of google indoor maps, set it "for blind user" and shazam
  (+1)
(+1)
  [vote for,
against]

indoor maps for the blind has the following features:

Location of human information point
Entrance and exit - and emergency route
Specifically important locations that blind people need.
Warnings about obstacles.
Voice interface for both input and output.
Orient yourself - by having a look around
Human volunteers can help you out
Human volunteers can help save you in case of an emergency.

pashute, Jun 25 2017

Tikkun Olam Makers http://tomglobal.org
[JesusHChrist, Jun 25 2017]

estimote beacons http://estimote.com/
[pashute, Jun 26 2017]

Why am I seeing drift when using Madgwick algorithm to correct for orientation with 9 DOF IMU? https://electronics...t-for-orientation-w
[Ian Tindale, Jun 27 2017]

SubPos Positioning System https://hackaday.io...-positioning-system
For [I T] [notexactly, Jun 27 2017]

SubPos Ranger https://hackaday.io.../9242-subpos-ranger
For [I T] [notexactly, Jun 27 2017]

Laser velocimeter https://www.google....JLwpvMyKovADEzHq6dA
Now, I don't propose that anyone should walk around with this on their head, but this shows the astonishing accuracy available. [Ling, Jun 28 2017]

[link]






       One of the ongoing issues with a thing I was trying to invent until recently giving up on it is a way of ascertaining where in a room I am, if I move. I've been experimenting with 9DOF and 10DOF accelerometer magnetometer gyroscope MEMS boards. Talking to other people in the VR industry about it the general agreement is that it is a big problem, probably a hard one. With such a sensor, you can ascertain movement and orientation quite easily and I can readout pitch yaw roll and XYZ translation, but the translation is only within a limited area. There seems no easy way to use such a sensor arrangement to detect precisely where I've walked over there, from here. Such attempts tend to drown the readout in correction noise resulting in massive drift over a short period of time.
Ian Tindale, Jun 26 2017
  

       The Israeli standards for public buildings now has the regulation to use a BT beacon at the information and entrance point. So, at least, that's a good "starting point".
pashute, Jun 26 2017
  

       Google Tango could be a good platform to develop this on.
notexactly, Jun 27 2017
  

       Thanks for the links. You see, most solutions to this problem involve some external positioning of more than one device to define the area within which detection occurs. There isn't a really good way of doing it from the device in question alone. Most solutions I've seen resort to things like the Kinect sensor or pairs of IR or visual sensors on stands, etc. Therefore you have to have predefined the allowed area and range and you can't walk outside that.
Ian Tindale, Jun 28 2017
  

       what's not to like?
po, Jun 28 2017
  

       I think the problem with accelerometers is that you need to integrate twice to get distance. Any small offset error is summated continuously.   

       But if you can go up one level, it gets a little more accurate. Laser speed measurement can be done perpendicularly to the object. It's quite common to shine a laser spot on a moving strip, from the side, and combine the reflection with a sample of the output to get pulse measurement of velocity. Quite accurate, too.   

       One could gymbol such a laser spot vertically to the ceiling and directly measure velocity of a pedestrian in a room. Then integrate only once to get distance. Not perfect, but getting better....
Ling, Jun 28 2017
  

       My guess at this stage is to have the IMU combine the methods such as Madgwick for orientation, etc, with a much more primitive dead reckoning, or more precisely, pedestrian dead reckoning (PDR). Only sense the larger and more significant changes, and inbetween, in effect 'switch off' the sensing (rather than simply accumulating error) until it is sensed that another major change has occurred. Sensing the difference between a 'settled' step and the transition of striding to another step point, may allow the measuring to 'switch on / switch off' accordingly.
Ian Tindale, Jun 29 2017
  

       //Such attempts tend to drown the readout in correction noise resulting in massive drift over a short period of time.//   

       Sounds like your programming is really kicking in. Don't forget the basics of unit testing. If you can't track a single sensor well, then you can't take data from 12 sensors. Try tracking one dimension and then more. Also, some $20 boards work 100 times faster than others.   

       Maths ability also kicks in. Imagine X being (X + e). Just substitute X for (X + e) and work it through. Then work out how big e/X needs to get in order for things to disintegrate.   

       //Sensing the difference between a 'settled' step//   

       That is 1980's ADC talk. Most sensors these days should present a 'sensible' reading, if not, you need to look at accuracy vs settle time. Again, try a single dimension. Stick the motion sensor on a bass speaker and sum the movement.
bigsleep, Jun 29 2017
  

       // You see, most solutions to this problem involve some external positioning of more than one device to define the area within which detection occurs. There isn't a really good way of doing it from the device in question alone. Most solutions I've seen resort to things like the Kinect sensor or pairs of IR or visual sensors on stands, etc. Therefore you have to have predefined the allowed area and range and you can't walk outside that. //   

       Indeed. Perhaps a lidar+SLAM or visual odometry solution would be better for your purposes.   

       Also, [Ling], I have to look into that laser velocimeter technology some more. How does it measure transverse velocity?
notexactly, Jun 29 2017
  

       As I understand, they combine the reflection with the original source and look at the interference.   

       It was a surprise to me to find out that the perpendicular movement also changes the reflection frequency, but the instrument test engineer gave me this answer during a witness test that I did, once, for a steel rolling company.
Ling, Jun 30 2017
  

       Surely what we're missing here is the need to restore these people's sight?
MaxwellBuchanan, Jun 30 2017
  
      
[annotate]
  


 

back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle