2
\$\begingroup\$

I am trying to make a little wireless-connected gizmo to attach to my grandfather's walking cane. The idea is that when it senses the cane has fallen down, it would alert/call one of us so that we can rush to him if we're not nearby.

The question is: What sensor or combination of sensors could I use to reliably differentiate between these two scenarios?

NORMAL situation: The cane is being held upright (or at only a slight angle from vertical) and used normally.

versus

CRITICAL situation: The cane has been let go of, falls down and continues to lie on the floor/ground

Obviously, occasional false positives would be fine, but detection misses (false negatives) should be minimized.


I am thinking a combination of two things might help:

--sense whether my grandpa's hand is holding the cane or not (perhaps a light sensor that is blocked when the hand is on it?)

--sense whether the cane is upright or has taken a fall (perhaps an accelerometer+gyroscope+vibration sensor to determine orientation and shock events?)

\$\endgroup\$
  • 1
    \$\begingroup\$ Yep I would have said the second one and you can use the sensor to have it sleep until it's picked up to save some battery life. Maybe a cap sense style sensor around the handle to see if it's being held? Led could work too \$\endgroup\$ – Some Hardware Guy Jan 8 '15 at 1:50
  • \$\begingroup\$ Btw there are a lot of things out there doing something like this to detect falls so you might find some relevant research out there that could help. \$\endgroup\$ – Some Hardware Guy Jan 8 '15 at 1:51
  • \$\begingroup\$ Why not measure directly the person rather than the cane he might hold? What about those pendant systems that can flag a critical situation? \$\endgroup\$ – Andy aka Jan 8 '15 at 8:35
  • \$\begingroup\$ @SomeHardwareGuy: I'm going with that. As for the capacitive sensor idea: what would you suggest given that the handle of the cane seems to be made of some sort of synthetic rubber? I'm guessing some sort of thin metallic "ring" around the handle, and a wire soldered from the ring to the stripboard inside? \$\endgroup\$ – boardbite Jan 10 '15 at 19:03
2
\$\begingroup\$

I agree with one of the above replies in that you need something (probably two accelerometers) to detect the angle at which the cane stands (calibrated at the factory if you are industrializing this). This way you have a continuum of what angle the cane is at, relative to upright.

This gives a distinct advantage over a simple tilt switch for the following reasons. 1. You can calculate the angle with two accelerometers. With a tilt (on/off) switch, you cannot 2. Not only can you calculate the angle with two accelerometers, you can detect the rate at which the angle changes relative to the cane standing up straight. This is key as you can then at least have the information to characterize what the rate of fall looks like by a simple drop test and base your fall detection algorithm model on that.

\$\endgroup\$
  • 2
    \$\begingroup\$ It's barely more money for a 3-axis now a days but I think he could keep it simple with a single axis accelerometer. He could still calculate a rough angle with this (between horizontal and vertical) albeit with decreasing resolution as the sensor approaches 1g. He couldn't distinguish between pitch and roll (if there is such a thing for a cane) but I don't think it matters for this application. Just need to know if it's horizontal, vertical or somewhere in between with coarse accuracy. In one situation you'll be measuring ~1g, in the other ~0g. \$\endgroup\$ – bt2 Jan 8 '15 at 4:45
1
\$\begingroup\$

A very simple solution is a tilt sensor such as the AT-407 from Sparkfun. It consists of 2 steel balls inside a small tube and 2 leads. When the tube is tilted, the balls separate and continuity is lost between the 2 leads. It costs only $1.95.

\$\endgroup\$
1
\$\begingroup\$

The most reliable way would be a gyro+accel combo (like the mpu6050 from invensense). You can just use their DMP stuff to get a roll/pitch/yaw angle for the sensor, or do the work yourself by running the data through a kalman filter. You don't need a vibration sensor, as you can get that from the accelerometer data too (just look for impact events).

\$\endgroup\$
0
\$\begingroup\$

Seems like an accelerometer and a small microcontroller (a PIC for example) could do this quite easily. You could then set the angle at which the alarm goes off in the code.

With the light sensor approach you'd need to be careful that the area stayed clean. Otherwise if he covers the sensor with crud it will always indicate that the cane is vertical even when it is not.

With a vibration sensor you'd have to get creative about detecting the impulse event (when the cane falls) and then latch the output so it continues to sound the alarm even after the can has come to a rest.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.