Dear everyone,
I performed some experiements with a FMCW Radar Instrument emitting waves of 3.8mm. I can see some "movements" when I am looking on the relative displacements (+/- 0.15mm) over multiple stable points (corner reflectors). I assume part of this "displacements" is caused by temperature, humidity and pressure variation over the time of measurement (outdoor experiment). However, when applying Rüeger (2002, Refractive Index Formulae for Radio Waves) formula then the values are about a factor of 10x too small. There does not seem to be an error of units. The paper mentions resonance lines at about 67GHz and I am wondering if this could have an influence on the measurements with 79.5GHz and is there a publication specifically for the frequencies around 80 GHz?
Would anyone have a suggestion what else could cause this "movements". The radar and the targets are mounted on a concrete pillar and I assume that there is no relative displacements due to the similar setup.
Thank you very much and best regards,
Andreas
Ridha Ghayoula thank you. It is indeed an interesting question and it gave me quite some headache already.
The 80 GHz band is chosen because it is not close to the absorption bands, so the absorption bands are not causing your problem.
The movement of 0.15 mm is about 15 degrees of phase. However, it does not seem to be a problem with the radar because it does not track the same way with the different targets. I don't think it is due to the properties of the air either, because I would expect that it would show a common trend with all the targets.
Is it possible it is thermal expansion in sunlight?
Many materials expand by around 10 parts per million per degree. This would be 100 microns per meter for 10 degrees of warming. A concrete pillar warmed on one side might bend.
I carried out accurate phase measurements at 80 GHz indoors for years and never observed effects like this from humidity or air density, but I did see problems like this from expansion of a delay line as it warmed up and cooled down.
Malcolm White : Thank you very much for your suggestions!
A) I am a little bit confused with "it does not seem to be a problem with the radar because it does not track the same way with the different targets. I don't think it is due to the properties of the air either, because I would expect that it would show a common trend with all the targets.": In my plots, I selected three different and strong targets (2 corner reflectors and 1 rail) at various distances from the radar (20-41m). From my point of view they show the same "track" or "trend". Attached I am also sharing a better figure, where I multiplied the "phase values of the atmospheric model" by a factor of 10.
B) I was also thinking of thermal expansion of the structure, e.g. concrete pillar. However, on this day, there was almost no direct sunlight. It was mostly cloudy/foggy and the temperature of the air stayed between 0.2 and 0.7°C. Therefore, I doubt that any structure warmed up by 10°. However, I am going to repeat my experiment and also measure to an other pillar on the terrace and see if the "movements" can be observed between these two "twin" pillars or if they cancel each other out. If they cancel each other then a thermal expansion would be a possible explanation for this phenomena.
Best regards,
Andreas
I hadn't understood what the dotted curves were. I understand that now. What I meant about common factors is that there seem to be at least two contributions to the variation. The three curves follow a similar variation, but the blue curve, for example, is largely consistently above the other two, and is sometimes below them, indicating that at least part of the variation is not common to the three measurements. The yellow also varies compared to the red. You might try to subtract from the signals a common deviation that leaves the smallest RMS deviation for them all, or just subtract the mean of the three at every point. This mean may be due to phase stability in your radar - or it might not be. The remaining differences are between individual targets and will be smaller. Separation it out into a common variation and differences between targets may help with understanding it.
Malcolm White : Thank you very much for your help. I will try to subtract the mean values or the common deviation with the smallest RMSE and see what is left.
A related question: You mentioned that you carried out indoor measurements at high accuracy for years. Did you by chance publish something or could you recommend some papers related to accuracies of 80GHz radar sensors?
Thank you very much for your help!
Andreas
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Atmosphere