Consider the following scenario (see figure below):
At beginning, time 0 (t0) is recorded, [A] (left) produces an acoustic signal, detected by [B] (right). If the signal can be discriminated from background noise (see below, e.g. by frequency, i.e. any ultrasound signal as opposed to human voice spectrum), [B] sends a 2nd signal, which is detected by [A], and the time (t1) is recorded (to RAM for processing or to SD card for processing later).

t1 - t0 will

  • indicate the distance
  • comprise both the "time of flight" forth/back (blue) and the electronic switch/shifting times (red).

(Background noise is from normal speaking, and signal should be non-disturbing - in the easiest(?) case signal is ultrasound, and noise is normal human voice.)

Question: what are "normal" switch/shifting times (the "red" parts) in terms of absolute time and (more impotrant for calculating accurate distances) variance? "Normal" refers to "achievable by commercially available electronic compounds", with [A] and [B] no larger than few cubic centimeters each (say max. 15 cm3), and preferably << 100 $ alltogether. If necessary, a seperate controller unit (e.g. Arduino-based) can be attached to sender/receiver units [A] and [B] by wires. All calculations (filtering, calculating of distance from ToF times) can be done afterwards.

Accuracy of the whole thing needs to allow measurement of distances in the centimeter to meter range, with precision of few centimeters. Actually the purpose of my question is, if this is achieveable by acustic time of flight. I am more interested in the increase/decrease of the distance (within ~20 min) than the exact absolute distance.

Bonus points:

  • Any suggestion that facilitates the project, e.g. what components I should look for
  • If not feasable, I would be happy for other suggestions to measure the distance between 2 objects that do not necessarily face each other in really constant angles in 3D space (therefore I thought of sound in a broad beam, ~50° would suffice).

enter image description here

  • \$\begingroup\$ Is there any reason you cant use an off the shelf ultrasonic distance sensor? It depends on reflections from whatever object is being measured, so it might have too narrow of a beam width: amazon.com/dp/B004U8TOE6/ref=cm_sw_r_cp_apa_i_gnYWDb3843A95 \$\endgroup\$
    – Ben Watson
    Nov 6, 2019 at 19:50
  • \$\begingroup\$ I'd love to. Could you please eleborate on your point regarding beam width (I simply do not understand that sentence). There was concern that beam widths of usual ultrasonic distance sensors are too narrow to allow for passive reflector objects moving around in 3D space (e.g .hands or legs - in short, I am actually interested in accuracy of movements, i.e. how close they come in cyclic movements). I am happy for any kind of solution (not necessarily limited to (ultra)sound. \$\endgroup\$
    – Martin
    Nov 6, 2019 at 19:55
  • \$\begingroup\$ I made an ultrasound location system once, but TDOA instead of ranging. I used ultrasonic flea collars for the transmitters and bat detectors for the receivers. \$\endgroup\$ Nov 7, 2019 at 16:23
  • \$\begingroup\$ @CristobolPolychronopolis: Great - what detection range did you cover? If it works, the "switch times" (in relation to the "travel times") seem ok, this would be an indirect answer, but I guess it was rather ~100 m for dogs running around than 100 cm as I would need (cm to m range)? \$\endgroup\$
    – Martin
    Nov 9, 2019 at 10:28
  • \$\begingroup\$ We used it in a conference room about 4m by 6m. We didn't try it over longer distances, but we never faced any problems we attributed to sensitivity or SNR, so I suspect we had a margin of at least 10-20dB. We could have used far narrower filters as well, but there was no need. \$\endgroup\$ Nov 11, 2019 at 16:20


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.