ultrawide band sensors are mounted on two objects which aint visible to each other at all. How do i align them.
Dear Pallavi,
I do not know if you mean ultrawideband wireless rf sensor. If this is the case you can send a high frequency carrier which is crystal stabilized from one to the other which receives it and estimate the Doppler shift df. He then starts to adjust its speed until df is equal to zero. Which means the two have the same velocity.
Best wishes.
I guess you are talking about RF UWB, systems like Decawave that can measure distance node-to-node quite precisely.
The aliment then means learning precisely about direction to other node (the distance is already known via UWB return propagation).
What can be employed for that, considering there is always some small or not very small random movement of nodes in various directions (one can also stimulate it intentionally)-
If you take signals from gyroscopes and accelerometers for short time ( e.g. 1..2sec) and correlate them (as vectors) with node-to-node distance variations, assuming the other node is either rather static or does uncorrelated movement one can derive the direction pointing to the other node.
Such systems can be employed as smart watch pairing small children with their parents ( the parent can just swing the hand a bit to induce movement).
All required hardware (low cost , low power, ultra compact) is already available.
Nikolay Pavlov
Yes , i am using UWB device along with IMU. Cna you please elaborate on how to have node-to node distance variation.
Yes in our case, 1 device is stationary and the other device is mobile, the mobile device needs to be alignedto stationary device.Both the devices have UWB sensors as tags and anchors
The distance between UWB sensors is estimated via precise measurement of round trip signal delay (the other transceiver must be capable to respond with exact reproducible delay). To some degrees, one can also use RSSI (signal strength) but it will be affected by propagation conditions.
You use the fact that double integration of acceleration gives you distance travelled, however, the complication is that any offset error also gets integrated so it grows with time quadratically. However, you still correlate short strokes of movement with changes of node-to-node distance.
All of the above has to be done in 3 coordinate XYZ vector form, the direction being derived from gyroscopes (there is also some way to extract orientation using combination of gravity and Earth magnetic field).
There are might be some smart algorithms suppressing the double integration error via e.g. resetting the speed when the system is stationary, flipping the accelerometer to nullify the offset on g, etc.
Unfortunately, such technologies are too similar to 'dead reckoning' used by missiles, submarines, etc. so you could not find much information on it in open sources.
Nikolay Pavlov Can i have your mail id so that i can explain you my problem in detail.
- Badges
- Science topic
- Similar topics
- Methodology
- Sensors