Delay Spread: It is the difference of the arrival time of multipaths. If the BW of the signal is W, then all the paths that arrive within 1/W are non-resolvable, i.e., they cannot be separated and hence they result in either constructive or destructive signals due to the fact that these different phased signals are added all together. On the other hand, if W is very large, then 1/W is very small. Therefore, the multipaths can be easily resolved, i.e. separated. If we had a signal with infinite W, then 1/W would have been almost zero. Therefore, we could resolve all the multipaths and we would not have the fading phenomenon. But BW is a scarce resource and hence we don't have an infinite bandwidth and hence we are to accept fading. If the delay spread is smaller than 1/W, then all the multipaths would fall within the duration 1/W, and there will be only one tap of the channel. This is known as the flat fading channel. So, this happens when delay spread W. 1/delay spread is also known as coherence bandwidth. So, we have a flat fading channel when the coherence bandwidth of the channel is greater than the signal bandwidth. Otherwise, the channel is a frequency selective channel Doppler Spread: On the other hand , Doppler spread is a phenomenon linked to the change of the channel with time. It is proportional to carrier frequency, angular spread of the multipaths, velocity etc. Doppler spread is higher in a high scattering environment where the angular spread is larger. Doppler spread is not directly related to delay spread. Doppler spread is proportional to the range of rate of change of path-delays while delay spread is the range of delays. If two paths rotate at the same speed (same doppler shift), then the doppler spread will remain the same and the channel will not vary.
I'm very glad that Qualcomm has made these lectures publically available. Really a treasure to the wireless communications community!
Delay Spread:
It is the difference of the arrival time of multipaths. If the BW of the signal is W, then all the paths that arrive within 1/W are non-resolvable, i.e., they cannot be separated and hence they result in either constructive or destructive signals due to the fact that these different phased signals are added all together.
On the other hand, if W is very large, then 1/W is very small. Therefore, the multipaths can be easily resolved, i.e. separated. If we had a signal with infinite W, then 1/W would have been almost zero. Therefore, we could resolve all the multipaths and we would not have the fading phenomenon. But BW is a scarce resource and hence we don't have an infinite bandwidth and hence we are to accept fading.
If the delay spread is smaller than 1/W, then all the multipaths would fall within the duration 1/W, and there will be only one tap of the channel. This is known as the flat fading channel. So, this happens when delay spread W. 1/delay spread is also known as coherence bandwidth. So, we have a flat fading channel when the coherence bandwidth of the channel is greater than the signal bandwidth. Otherwise, the channel is a frequency selective channel
Doppler Spread:
On the other hand , Doppler spread is a phenomenon linked to the change of the channel with time. It is proportional to carrier frequency, angular spread of the multipaths, velocity etc. Doppler spread is higher in a high scattering environment where the angular spread is larger.
Doppler spread is not directly related to delay spread. Doppler spread is proportional to the range of rate of change of path-delays while delay spread is the range of delays. If two paths rotate at the same speed (same doppler shift), then the doppler spread will remain the same and the channel will not vary.
Great explanation..!
Legendary D. Tse! Thanks!