Detection of 3 Hz doppler shift in 1.2GHz signal
Actually It's my final year project to make a system that will detect the alive human bodies under the debris in case of catastrophic situations. Well, I've got two methods to accomplish this task.
First one is, send 1150MHz wave and after receiving through receiving antenna subtract it with the reference signal ( signal that was sent through transmitting antenna) by manipulating the reference signal such that its phase and amplitude cancel out with the 1150MHz signal while the doppler shift of 3Hz (due to heartbeat 1.5Hz in forward while 1.5Hz in backword direction) will remain there and then by using FFT, we can detect the peak of that doppler shift frequency.
I think this method is a bit complicated. We would need to match the phase and amplitude exactly.
Second one is, send 1150MHz frequency. Receiving signal would be the part of 1150MHz and doppler shifted signal. We can take the signal to smaller frequency. Let's say 50Hz so that it is comparable with the 3Hz signal. then we can use the FM demodulation to extract the signal (3Hz).
I think this method is relatively easier than the first one.
I wanna ask, Is the second method is feasible to detect very small amount of doppler shift? or which method do you think can be easy and feasible?
PS: Due to range and penetration, we have decided to use 1150MHz frequency.
Any help would be appreciated.
Interesting project! I believe what you get isn't a 3 Hz doppler shift rather than the spectrum of a 1.5 Hz phase modulation, with sidebands at multiples of 1.5 Hz and a distribution depending on the (probably low) modulation index.
Low oscillator phase noise is surely a prerequisite to detect the backscattered modulated signal.
A few thoughts it looks like an interesting challenge for a final year project from which you should learn a lot.
The problem to be solved is the small phase shift that you need to detect not so much the frequency of the heart beat.
You will be transmitting a signal to the static target, the retrun from that target will be a a phase angle that depends on the distance, and superimposed on tha return will be a small periodic phase change that will depend on small body movements. You will need to find out how much movement to expect and rfom that you can determine the phase shifts. this will enable you to determine the level of the doppler shift you are looking for. I would guess that breathing will give the greatest phase shift, movement from heart beat will be very small. At a guess you are looking for a periodic phase shift of a few degrees on the returned signal. Knowing hte expected path loss and using some trigonometry will tell you the magnitude of the signal from your detector.
For the radar equations including doppler effects look at the standard work on radar (Radar Handbook) by Skolnik, you should be able to find most of what you need there.
You will also need to consider the mechanics of the system. You arelooking for small movemnts of the target, you can not afford to have similar movement in the radar.
As to whether phasing out the transmitted signal will work, it has been done. research cw radar or fmcw radar. some small boat radars use about 100mW and can detect targets several km away. FMCW may complicate thing but would give you an indication of distance from to target, useful if you need to dig through rubble.
Peter
Thanks FvM.
Yes. It would be phase or frequency modulated signal we will receive.
It would be something like the image.
Actually, Could that be possible to extract the heart beat component from the modulated signal? I am afraid by taking the signal to lower frequency oscillator frequency should be very accurate and with much higher precision. Isn"t it?
Or should I use the first method?
Thanks G4BCH.
Can we not call it as frequency modulation? Phase shift would be there in any case whether there is alive human or not under the debris. So, how could we confirm the detection of alive human body?
We thought to use the FMCW mechanism as for the extra advantage of distance. But it's a bit complicated and but after basic step is done, we can move further.
Thanks for suggesting the great book.
You could call it frequency modulation, but the doppler frequency shift is so small it is easier to work in phase.
To get a practical handle on the shifts involved, pretend to be the trapped victim for a while. Measure or estimate the movement of your body in time with your heart beat, not much I would guess. Then measure your movement in relation to your breathing, again not much but considerably larger than for the heart beat. Translate that movement into wavelengths to find the phase shift. That will lbe the modulaton of the returned signal at what ever rate it is changing.
As FvM pointed out a very low noise oscillator is needed to prevent the returned signal being lost in the noise.
I think Stow away prevention and detection systems use heartbeat detection though I'm not sure if they are radar or acoustic systems.
Before getting in any phase shift issue, my opinion is that would be very hard (near impossible), to get a 1.2GHz signal reflected back after crossed few meters of reinforced concrete (metal inside).
Through Wall Radars which are on the market have hard time to receive the reflected waves through a simple brick wall.
I played once with a device like this, and didn't work at all if the wall was a thick reinforced concrete.
And the ceilings of multi-storey buildings are 99% built with reinforced concrete. In case of collapse these buildings looks like a sandwich of ceilings.
In order to resolve 3Hz in 1.2GHz, your VCO will need to be extremely low Phase Noise and Frequency drift less than the capture range of the receiver and temperature drift very low or using VC-OCXO <0.1ppm over environmental range.
That wont be ordinary AT cut crystal and SC cut are too expensive. So the RX must always stay in lock with TX with bursts and suitable PLL tracking filter. 1ppm VCXO's are around S50 or less if you shop around. 0.0001 ppm OCXO's are around S300. 2 PPM TCXO's might be S10 or less if lucky, but no guarantee on phase noise or f corrections drift withing the 2PPM.
YOu need to resolve 2e-9 resolution in frequency.
So what is your error budget for f and Φ ?
Thanks vfone and SunnySkyguy.
Actually, the error could be tolerated. We want not to extract the 3Hz signal completely. What we want is, to detect the alive human body whether someone is there. Are you talking about first method or second?
there is a system made already that uses a lower frequency UWB pulse, and can detect heartbeats 10 feet under debris.
you might be wasting your time with 1.2GHz,
I believe the successful implementations are at least 60x higher frequency to go thru thick concrete walls.
But at least, it will make a good GPS/GSM jammer.
HI, SunnySkyguy,
How about delay? That can make up for poor phase noise.
And 1.2GHz should have better performance than 72GHz, but 72GHz may have wider bandwidth for better performance.