Beamforming directivity [Part 1, Fundamentals of mmWave communication]
HTML-код
- Опубликовано: 11 сен 2024
- Antenna arrays are essential in mmWave communication to achieve a directional transmission that focuses the wireless signal on the receiver. In this video, the first in a series, Prof. Emil Björnson demonstrates how beamforming directivity is achieved in practice by transmitting the same signal from multiple antenna elements - the more antennas, the higher directivity. He both explains the underlying theory and demonstrates the key principles experimentally. Beamforming is essential at both base stations and user devices when using the 5G mmWave bands, but it has long been utilized at base stations in the lower frequency bands.
The experiments are carried out using the TMYTEK Developer Kit www.tmytek.com... which the company kindly lent us. This is not a sponsored video.
Emil Björnson is an IEEE Fellow and a Professor of Wireless Communication at the KTH Royal Institute of Technology, Sweden. Website: ebjornson.com/r...
#5g #6g #mmwave #beamforming #wireless
Fantastic video!🥳👍 Your presentation was very clear and easy to follow. The way you connect theory to practice is impressive and highly educational. Thank you, Professor, for sharing your expertise with us in such an engaging manner. Your video has been an invaluable resource for me.
Excellent video and easy to understand. Please keep posting time and again
Great video, short but comprehensive! Looking forward to the next videos on this topic 😊
Very nice demo. I am Looking forward to see (in coming videos hopefully) that how "Gain" of the (say) receiving antenna will be measured in this setup, where total power is received separately at the receiver side.
This video is an excellent and easy-to-understand presentation of concepts. Thank you.
there will not be better visualisation for beamfoarming than this
❤
Very nice experiment with a fantastic presentation!
You are very grate Professor.
Your explanation is very clear and easy for understanding.
Amazing explanation
Thanks, when we turn on just one antenna and get 9 dBi - we can assume with all 4 antennas on, we get 9 + 9 + 9 + o = 32 dBi with all antennas? Is that right?
But here you obsereved, that isn't the case here. Correct?
No, one can only add up gains like that in linear scale. 9 dBi = 8 times gain compared to isotropic antenna. 8*4 = 32 times gain = 15 dBi.
One can also compute it directly as 9 + 10*log10(4) = 15 dBi.
Looking forward to more of these.
Quite insightful professor Emil. Was waiting for it :)
looking forward for the other parts of the experiment
Great video. Very easy to follow.
excellent description
Thanks a lot for the video. One quick question, there are 4 channels. Is a channel the same as one RF chain? and are there 4 RF chains? as far as I know each column has one amplitude and phase controller that's why I think (perhaps I am wrong) each column is corresponding to one RF chain. Thanks
The analog beamforming architecture challenges the notion of “RF chain”. We generate one RF signal, which is then divided into 4 branches where it is amplified and phase-shifted independently and then sent over an “antenna” (consisting of 4 vertically stacked radiating elements).
A traditional RF chain would connect the baseband to an antenna. There are not 4 full RF chains here. The setup here is also known as a phased array
Thanks, the sheets values differs from your practical observation. Are there any information about the gain of a signal element (one of the 16). You said with one antenna switched on, we have 9 dBi. Can i now conclude that one element has 9 dBi/4 ~ 2.25 dBi in your experiment?
According to the documentation, the entire 16-element array has a gain of 15 dBi. The gain of one column is 15-10*log10(4) = 9 dBI and the gain per element is 15 - 10*log10(16) ≈ 3 dBi.
Thank you 🙏
What will happen if you have two three users while the Base station sends a one beam?
The users will have to take turns. You can point the beam towards one user at a time. If one beam is wide enough to cover multiple users, these users can be served at the same time but using different frequencies.
On the two channels you found a comparative gain of 4.5(=3+1.5)dB instead of 6(=3+3)dB. However, they made an inaccurate conclusion about the reason for the decrease in gain. More correctly, for example 4.5(=2.7+2.8).
This is due to the non-identity of the directivity characteristics of the antenna elements for the selected wave polarization, the inaccurate correspondence of the horizontal distance between the antenna elements to the selected wavelength, and also to the "near zone" effect.
By the way, as you may have noticed, the real maximum of the antenna element directivity in the horizontal and vertical planes differs from the normal to the antenna plane.
There are indeed other RF effects that contribute to the observed gain variations, such as the ones that you mention, mutual coupling, and the fact that the measurements weren't made in an anechoic chamber. However, I obtained roughly the same gain variations when replacing the antennas with cables, so I'm convinced that my conclusions are correct.
By the way, the Fraunhofer distance is less than 20 cm in this setup, and all measurements that you will see in this video series were done beyond 30 cm. So I believe the near-field effects are negligible.
@@WirelessFuture Yes, of course, reflections from surrounding objects affect the resulting received power. But they can affect both "plus" and "minus". By the way, it would be interesting if you set up such an experiment.
However, regarding the interconnection effect you mentioned, I disagree. This connection is only caused by the fact that there is another antenna nearby. And it doesn't matter if this second antenna radiates or not. The main thing is that in both cases it should be impedance matched. So, for example, in real antenna arrays, to ensure maximum identity of the directional characteristics of the peripheral and internal antenna elements, additional matched "passive" antenna elements are used that "surround" the main antenna array.
And yet, did you find the deviation of the maximum directivity of the antenna from the normal to the surface of the emitting antenna?
I have made plenty of other measurements, and the most informative ones (from a teaching perspective) will be presented in the next four parts of this video series.
Yes, I have characterized the radiation pattern with one, two, and four antennas. In the single-antenna case, the maximum directivity was indeed observed a few degrees from the normal, but the difference was only ~0.2 dB.
@Andrey Sheleg the real maximum of the antenna element directivity in the horizontal and vertical planes differs from the normal to the antenna plane, I think it is due to series-fed array topology. What do you say about it?
@@AbdulUofG The effect of the asymmetry of the directivity characteristic (and, as a result, the deviation of the radiation maximum from the normal) is easier to understand on a simple example of a single rectangular patch element. Since it is impossible to avoid the asymmetry of the power supply of this element (at least in the E-plane), there is a difference in the amplitudes of the electric field at the opposite edges of the patch. The reason for this is the losses in the dielectric and in the metal (for heating and for scattering in inhomogeneities and roughness).
A single element within a column is directed mechanically toward the receive antenna, to improve directivity toward the receiving antenna by 3 db better than an isotropic antenna. Correct?
Yes!
thank you prof for all these great videos that you share with us for free
do you recommend any course any online training for someone who wants to be RF engineer ?
I'm not an RF engineer myself, but work with communication theory and signal processing algorithms, so I'm unfortunately not familar with what online courses there are.
After watching the video, I want to say thank you for sharing this informative content. I do have a question, though. Does the device only measure received power? While it's true that the received power increases as the number of channels increases from 1 to 4, I wonder if this also leads to an increase in the decoding complexity at the receiving end.
In this case, this can lead to an increase in the complexity of decoding and even to the impossibility of decoding at information transmission frequencies of 10 gigabits per second and higher (for one polarization). That is, when the duration of the symbol approaches the wavelength of the carrier wave.
But such short pulses physically cannot be effectively emitted and received even by single antennas.
The supported data rate is approximately “bandwidth*log2(1+signal-to-noise-ratio)” bit/s, so if you increase the received signal power, you can also transmit more data. On the one hand, you need a larger decoding complexity to successfully receive the data. On the other hand, wireless systems are generally designed to support a certain maximum data rate and then therefore have the hardware capability required for the decoding. From that viewpoint, beamforming enables us to achieve the maximum data rate more frequently.
Thank you for the video! I have two general questions about the link budget. Firstly, are we supposed to add a different combining gain on the uplink link budget for different combiners (eg. mmse or mr)? Secondly, are we supposed to add an interference margin to the link budget for a mu-mimo system? If so, how do we quantify these values(combining gains and interference margin)?
Yes, if there is interference so that MMSE≠MR, then the MMSE combiner will reduce the SNR to improve the SINR. The detailed modeling of this is described in my video:
ruclips.net/video/Ob92pRLj0c4/видео.htmlfeature=shared
The preferred way is to compute the SINR and optimize it instead of the SNR. It is only when there is unknown interference that one might need to include an “interference margin” to account for the worst case interference. Note that we must know the channels to apply MMSE combining, so whenever it is used, we can also compute the SINR.
@@WirelessFuture Thanks for your reply. Can I ask another question : When we plan a network, we calculate the link budget based on the minimum required SINR for maximum available path loss. In this context, is the sinr calculated based on signal before combining or after a specific combining scheme?Thanks!
@@Zh-vo5hc In practice, coverage is the same as being able to decode the broadcasted downlink system information from the base station (e.g., SSB), and reach the base station with a random-access request (on the PRACH). There is no interference at this point, but there might be a beamforming scheme. The problem is that the transmitter/receiver might not have channel state information, which is why beam-sweeping methods or space-time codes are sometimes used to obtain diversity or beamforming gains that improve the coverage.
Does this equipment have 5G Physical Layer+RF?
No, it just sends a sinusoid at the specified frequency and measures the signal strength at the receiver. If you want to send actual 5G-like waveforms and demodulate them at the receiver, you need a more complicated setup. However, I believe that TMYTEK has some other products for that.
this seems to be like constructive interference in physics if im correct i understand how to gain diversity with time division but how would it work with frequency division multiplexing also how is spatial multiplexing not causing more interference as theirs more transmitter than receiver so their wouldn't be parity thanks
Yes, beamforming is based on constructive interference. We will take a closer look at that in a later video, where we control the beamforming direction.
Frequency division multiplexing is not so hard. You send different signals at different frequencies and tell each user which set of frequencies it is assigned to.
In spatial multiplexing, you need to select the beamforming and number of simultaneous transmissions wisely so that you can keep the interference under control. Massive MIMO is a solution to this, where there are more antennas at the base station than the number of simultaneous signals that you want to send. For example, 64 antennas but only 8 signals. This gives sufficient controllability to deal with the interference
@@WirelessFuture ok thanks
@@WirelessFuture hi so from my understanding example if theirs 64 antennas at the transmitter and 8 at the receiver the x8 gain from mimo is actually beam forming or constructive interference gain that causes the high spectral efficiency gains from their being better SNR thanks