i have a basic question regarding MIMO OFDM. will it work for long distance communication like deep-space comm. or intercontinental communication? and if not, why so? will it be due to different length of channels?
Will MIMO OFDM work for long distance communication?
Started by ●April 23, 2009
Reply by ●April 23, 20092009-04-23
On Apr 23, 2:15�pm, "grazer139" <mimo...@gmail.com> wrote:> i have a basic question regarding MIMO OFDM. > > will it work for long distance communication like deep-space comm. or > intercontinental communication? > > and if not, why so? will it be due to different length of channels?Your question is not really well-posed, but I'll try to answer anyway. If you have a lot of path loss, then you are going to be very constrained by energy per bit. This typically means that you need very narrowband modulation in order to overcome a few factors, including: 1. signal to noise ratio, obviously, against noise. 2. peak-to-average ratio of your transmitter and receiver. 3. dynamic range and noise floor of your receiver. Considering that OFDM's weaknesses include (2), then it's not the best candidate. Julius
Reply by ●April 24, 20092009-04-24
well ..
my question really should be that channel impulse response of four
channel for 2X2 MIMO , will be of different length in case of very long
distance...
so signals at reciver will be arriving at differnt times and
hence single tap equalisation not possible...
can anyody help me on that
>Your question is not really well-posed, but I'll try to answer anyway.
>If you have a lot of path loss, then you are going to be very
>constrained
>by energy per bit. This typically means that you need very narrowband
>modulation in order to overcome a few factors, including:
> 1. signal to noise ratio, obviously, against noise.
> 2. peak-to-average ratio of your transmitter and receiver.
> 3. dynamic range and noise floor of your receiver.
>
Reply by ●April 24, 20092009-04-24
On Apr 24, 2:02�pm, "grazer139" <mimo...@gmail.com> wrote:> well .. > � � � �my question really should be that channel impulse response of four > channel for 2X2 MIMO , will be of different length in case of very long > distance... > � � � � � � � so signals at reciver will be arriving at differnt times and > hence single tap equalisation not possible... > > can anyody help me on thatIt depends on the scattering environment. When you say MIMO, I assume you are referring to spatial multiplexing of multiple data streams (rather than simple beamforming of a single data stream). In some cases (I assume) the channel for deep space communications is close to line of sight (certainly satellite communications is virtually LoS), therefore in general the MIMO channel matrix is badly conditioned (large difference between the largest and other eigenvalues), and so will not effectively support spatial multiplexing. (In other words, the MIMO detector at the receiver would produce very large BER even with very small amounts of noise). You might be able to make the channel better conditioned with clever antenna array design, but the alignment would be extremely difficult at such long range). In addition, as others have said, the SNR is likely to be very low. In these situations it is better to use multiple antennas for beamforming (which increases SNR by increasing the effective array gain) for a single data stream rather than trying to send multiple streams. In general, the various spatial techniques (beamforming, diversity with scalar coding or space-time/frequency coding, spatial multiplexing) should be chosen appropriately depending on the channel characteristics, and therefore the operating regime with respect to the Shannon capacity. Finally if the channel is line of sight, OFDM does not make much sense. You might as well use SISO single carrier techniques, in which case you don't need equalization, or guard intervals, etc. Hope that helps. -T
Reply by ●April 24, 20092009-04-24
On Apr 24, 1:59�am, Tom <tom.der...@gmail.com> wrote:> On Apr 24, 2:02�pm, "grazer139" <mimo...@gmail.com> wrote: > > > well .. > > � � � �my question really should be that channel impulse response of four > > channel for 2X2 MIMO , will be of different length in case of very long > > distance... > > � � � � � � � so signals at reciver will be arriving at differnt times and > > hence single tap equalisation not possible... > > > can anyody help me on that > > It depends on the scattering environment. When you say MIMO, I assume > you are referring to spatial multiplexing of multiple data streams > (rather than simple beamforming of a single data stream). > > In some cases (I assume) the channel for deep space communications is > close to line of sight (certainly satellite communications is > virtually LoS), therefore in general the MIMO channel matrix is badly > conditioned (large difference between the largest and other > eigenvalues), and so will not effectively support spatial > multiplexing. > (In other words, the MIMO detector at the receiver would produce very > large BER even with very small amounts of noise). > You might be able to make the channel better conditioned with clever > antenna array design, but the alignment would be extremely difficult > at such long range).How to align antennas in such a case?> > In addition, as others have said, the SNR is likely to be very low. In > these situations it is better to use multiple antennas for beamforming > (which increases SNR by increasing the effective array gain) for a > single data stream rather than trying to send multiple streams.Beamforming might be the way to go in the context of deep space communications but it requires the channel feedback.> > In general, the various spatial techniques (beamforming, diversity > with scalar coding or space-time/frequency coding, spatial > multiplexing) should be chosen appropriately depending on the channel > characteristics, and therefore the operating regime with respect to > the Shannon capacity. > > Finally if the channel is line of sight, OFDM does not make much > sense. You might as well use SISO single carrier techniques, in which > case you don't need equalization, or guard intervals, etc. >OFDM will still work and you do not need guard interval is your channel is single tap. However, you probably need do coded OFDM is you are working with low SNR.> Hope that helps. > > -T
Reply by ●April 27, 20092009-04-27
On Apr 25, 10:12�am, xsong <songxiao...@gmail.com> wrote:> Beamforming might be the way to go in the context of deep space > communications but �it requires the channel feedback.For line-of-sight case, beamforming only requires the angle of the receiver (relative to the transmitter), which presumably is known. It doesn't require any other channel feedback.> OFDM will still work and you do not need guard interval is your > channel is single tap. However, you > probably need do coded OFDM is you are working with low SNR.OFDM will still work, but there is *no point* for LoS case. OFDM works well where you have a strongly frequency selective channel, where the channel nulls are mitigated by the channel coding. The use of a guard interval (common in OFDM but also in other techniques such as SC-FDE) simplifies the receiver processing for a multi-path (=frequency-selective) channel, but again has no merit for LoS (just wastes resources). Basic single carrier (with pulse shaping if necessary to increase bandwidth use efficiency) is simple and fine. The coding should probably be optimized for burst errors. -T
Reply by ●April 30, 20092009-04-30
On Apr 27, 10:13�pm, Tom <tom.der...@gmail.com> wrote:> On Apr 25, 10:12�am, xsong <songxiao...@gmail.com> wrote: > > > Beamforming might be the way to go in the context of deep space > > communications but �it requires the channel feedback. > > For line-of-sight case, beamforming only requires the angle of the > receiver (relative to the transmitter), which presumably is known. It > doesn't require any other channel feedback.Not really. Line of sight does not mean the channel is static. If the channel changes, you have to do feedback.> > > OFDM will still work and you do not need guard interval is your > > channel is single tap. However, you > > probably need do coded OFDM is you are working with low SNR. > > OFDM will still work, but there is *no point* for LoS case. > > OFDM works well where you have a strongly frequency selective channel, > where the channel nulls are mitigated by the channel coding. > > The use of a guard interval (common in OFDM but also in other > techniques such as SC-FDE) simplifies the receiver processing for a > multi-path (=frequency-selective) channel, but again has no merit for > LoS (just wastes resources). > > Basic single carrier (with pulse shaping if necessary to increase > bandwidth use efficiency) is simple and fine. The coding should > probably be optimized for burst errors. > > -T
