Thank you Matlab! Fantastic Pulse Doppler video. I've been telling young radar engineers for years that Pulse Doppler radar doesn't actually use the Doppler effect. Most radar literature introduce Pulse Doppler Radar by flashing the Doppler equation which is misleading.
Excellent explanation with intuitive concept telling. May we request for a video on Spectral analysis of signals comparing them to Time vs. Frequency domains?
Thanks Dr Brian for your amazing video, I am learning a lot of RADAR stuff from your videos. What is limiting the maximum range that can be detected. Is it limited by the time of transmit signal in fast time axis. If there is an echo beyond this time how will it be handled.
14:38 what happen if the object has an acceleration? In this case our sampling interval will vary in time. I think after sampling, aliasing sinus will not look like a sinusaidal, there will be a distortion, and it will spread around doppler frequency in frequency domain. in such case how will we determine the velocity?
Is it possible to measure target velocity by feeding the Rx signal to an FFT block and calculating the center frequency of the modulated echo? I would expect it to be shifted by the doppler frequency.
17:25 When can we expect "the next video on phased array antennas" ? Or is that already released, but under a different playlist. Also, great video series, thanks for the explanation.
What is the mixed filter in a receiver? Like what so the construction how does it work? Do you need a computer or could it be done with more simpler radars?
Wikipedia has a good write up for this if you search for matched filter. "In signal processing, a matched filter is obtained by correlating a known delayed signal, or template, with an unknown signal to detect the presence of the template in the unknown signal.[1][2] This is equivalent to convolving the unknown signal with a conjugated time-reversed version of the template."
How did he convert the frequency shift to velocity, i know Fd= 2V/lambda, with lambda C/F(2,8ghz), when calculating for 500hz i found about 60mph, while he found 120. Is there something i am missing
14:20 why are you assuming that reflected waves all have same freq ? if the target is getting far from you, the freq you recieve will be lower than what you sent. (dupler effect) Also 50 years ago when jets used pulse-dupler radar (F14) did they use sampling on IQ or they basically calculated the closure rate based on the time shift between pulses. As far as I know we calculate the closure rate based on the difference between Freq_sent and Freq_recieved. Otherwise, assuming that we are on a jet and getting closer to a target which is flying perpendicular to us (radar notch), since we are getting closer to the target, we will still have time shift between each pulse we recieve so we should be able to see the target again ! ... howerver this is not true. But looking into this using dupler effect since target closure rate is 0 in this case, therefore fsent=freceive and this target wont be shown on radar if we are using pulse dupler filter.
The comment around 2:00 is inaccurate. Modern radars don't avoid ambiguity, they simply use multiple pulse repetition frequencies that allow for disambiguation.
I really enjoy the way you pack this signal over time in a far more condensed version when compared to my last lecture. My hope is that this channel and one’s like it aren’t affected by the noise to signal ratio of channels that show the latest purse carried by some useless reality TV star, which is eating the brain cells of our children. Epically horrible and I hope some see that time wasted is never recoverable
The beat frequency is typically used to describe the difference in the transmitted and received signals for an FMCW radar. This difference in frequency for FMCW radar can be either induced by range and/or relative velocity of a target return. If you have a stationary target the beat frequency represents range (time delay). If you have a moving target you need another waveform slope (or constant CW tone) to solve for both range and relative velocity. If you have multiple targets you need even more slopes. There are many excellent youtube videos that explain that. In pulsed Doppler radar we observe the phase change of subsequent samples in slow-time (at the same range bin). The rate at which phase changes over your slow-time data gives you a frequency - and you can think of it as a "beat" frequency of sorts, but it only provides Doppler information. In pulsed Doppler radar instead of slopes (FMCW) we have different PRF to resolve range and velocity ambiguities should they exist. Pulsed Doppler radar are usually divided in low-medium-high PRF regimes and have different range/Doppler ambiguity implications. The MIT Lincoln Lab Radar Series I think explains this well. An important but challenging concept to understand is that a radar can receive signals from any direction, at any time, and with any frequency shift (subject to constraints). The radar's designer attempts to solve these ambiguities (or mitigate their occurrences) through various signal processing and hardware design schemes.
@@jonathananderson730 Thanks for the explanation. I understand pulsed Doppler radar uses the shift of RPF to solve the velocity, but it introduces velocity ambiguity. What I am think of is solving beat frequency on the other hand seems it can resolve velocity ambiguity. And waveform slope doesn't have to be used because pulsed Doppler radar can solve the range by solving the time delay of the receiving signal. So my question is, is my theory correct or is there any difficulty to utilize beat frequency for pulsed Doppler radar? BTW I am not a radar engineer, but a combat sims player who is so curious about how radar works.
@@YYCH2 I'm not sure I get where you are coming from but we can't typically get an accurate instantaneous frequency difference between the receive and transmitted carrier over a single pulse. Often this measurable difference has large errors that make it impractical to attempt to extract the Doppler from a single pulse.
@@YYCH2 I think I get what Jonathan means. Indeed most RADAR systems utilizing Doppler processing will use the output of a frequency mixer to pass Doppler frequency directly to the signal processing system-this lower-bandwidth signal can be more easily worked with for signal processing compared to the GHz-band raw signals. (After a fashion, the sample rate of the receiver acts as a mixer/filter anyways.) However, both systems will have a similar problem-the sampling rate of the received signal. Since the samples are being taken at the intervals of the pulse width, this puts a fundamental limit on the maximum non-aliased frequency that can be determined (Nyquist-Shannon sampling theorem), which is half that of the sampling frequency. If your Doppler beat frequency is higher than this sample rate, it will alias to a lower frequency and make for ambiguous velocity anyways. That being said, if you wanted to get a more complex spectrum from the pulse for things like non-cooperative target recognition and such, you could theoretically sample at a higher rate than the pulse width, but then you have to deal with more complex interactions between the target and receiver, determining which of the received frequencies is actually due to "real" target motion and not moving parts on the target, and so on.
I've tested and calibrated radar systems for years and I learned quite a few things watching your videos. Very interesting !
Thank you Matlab! Fantastic Pulse Doppler video. I've been telling young radar engineers for years that Pulse Doppler radar doesn't actually use the Doppler effect. Most radar literature introduce Pulse Doppler Radar by flashing the Doppler equation which is misleading.
one of the best videos and explanation I have seen in UA-cam. Great job!
Best of the animation every seen to understand the complex topics of EM and Radar Theory. Commendable..
Amazing work....so easy to understand...thanks!
Beautifully and masterfully done. Great video that makes me feel smarter than I actually am.
Thanks!
Superb Explanation! Expecting one playlist in sequence in more details and practical analysis on dummy data.
The first two Matlab below other links posted by the author would certainly allow you to take that next step.
my god, this video explains so much better and easier the Pulse Radar principle than my master class in germany
Glad it helped!
TU Darmstadt?
This series of videos is of great value. Thank you for your great offer to make this such useful, informative videos. Really thank you.
came to learn radar for DCS stayed because that was so interesting
Excellent explanation with intuitive concept telling.
May we request for a video on Spectral analysis of signals comparing them to Time vs. Frequency domains?
Yeah I Also want that topic to cover more intuitively
Thanks Dr Brian for your amazing video, I am learning a lot of RADAR stuff from your videos. What is limiting the maximum range that can be detected. Is it limited by the time of transmit signal in fast time axis. If there is an echo beyond this time how will it be handled.
Absolutely amazing explanation, thank you.
Thanks!
Thank you for making this video, it is very concise and informative.
14:38 what happen if the object has an acceleration? In this case our sampling interval will vary in time. I think after sampling, aliasing sinus will not look like a sinusaidal, there will be a distortion, and it will spread around doppler frequency in frequency domain. in such case how will we determine the velocity?
Is it possible to measure target velocity by feeding the Rx signal to an FFT block and calculating the center frequency of the modulated echo? I would expect it to be shifted by the doppler frequency.
17:25 When can we expect "the next video on phased array antennas" ? Or is that already released, but under a different playlist. Also, great video series, thanks for the explanation.
I think MIT lincoln lab has a series on phased array radars if you want even more info
Great Explanation, thanks
Thanks for the video!
What is the mixed filter in a receiver? Like what so the construction how does it work? Do you need a computer or could it be done with more simpler radars?
5:50 Is the matched filter used for correlation, right? But it seems to show us convolution.
Wikipedia has a good write up for this if you search for matched filter. "In signal processing, a matched filter is obtained by correlating a known delayed signal, or template, with an unknown signal to detect the presence of the template in the unknown signal.[1][2] This is equivalent to convolving the unknown signal with a conjugated time-reversed version of the template."
I assume the pulse interval is actually 1ms not 1us? 1ms * 55 m/s = 0.055 m
Fabulous video ❤
Do the pulses move in straight lines?
And I still dont found an answer to my guestiion, wich factor makes the SRC PD losse track of the targets?
7:54 what? Isn't the triangle wider? Even going by FWHM, the triangular pulse seems wider?
Or is it narrower in frequency space cus it's wider in time? Am I misunderstanding something?
Please provide Remix option to this Playlist.
Thank you.
Thank you for your suggestion.
How did he convert the frequency shift to velocity, i know Fd= 2V/lambda, with lambda C/F(2,8ghz), when calculating for 500hz i found about 60mph, while he found 120. Is there something i am missing
14:20 why are you assuming that reflected waves all have same freq ? if the target is getting far from you, the freq you recieve will be lower than what you sent. (dupler effect)
Also 50 years ago when jets used pulse-dupler radar (F14) did they use sampling on IQ or they basically calculated the closure rate based on the time shift between pulses.
As far as I know we calculate the closure rate based on the difference between Freq_sent and Freq_recieved. Otherwise, assuming that we are on a jet and getting closer to a target which is flying perpendicular to us (radar notch), since we are getting closer to the target, we will still have time shift between each pulse we recieve so we should be able to see the target again ! ... howerver this is not true. But looking into this using dupler effect since target closure rate is 0 in this case, therefore fsent=freceive and this target wont be shown on radar if we are using pulse dupler filter.
This was a GOOD 1
The comment around 2:00 is inaccurate. Modern radars don't avoid ambiguity, they simply use multiple pulse repetition frequencies that allow for disambiguation.
Doesn't the curve of the planet hide the object?
thanks my bro
I really enjoy the way you pack this signal over time in a far more condensed version when compared to my last lecture. My hope is that this channel and one’s like it aren’t affected by the noise to signal ratio of channels that show the latest purse carried by some useless reality TV star, which is eating the brain cells of our children. Epically horrible and I hope some see that time wasted is never recoverable
Why don't we use beat frequency to measure velocity for pules doppler radar?
The beat frequency is typically used to describe the difference in the transmitted and received signals for an FMCW radar. This difference in frequency for FMCW radar can be either induced by range and/or relative velocity of a target return. If you have a stationary target the beat frequency represents range (time delay). If you have a moving target you need another waveform slope (or constant CW tone) to solve for both range and relative velocity. If you have multiple targets you need even more slopes. There are many excellent youtube videos that explain that.
In pulsed Doppler radar we observe the phase change of subsequent samples in slow-time (at the same range bin). The rate at which phase changes over your slow-time data gives you a frequency - and you can think of it as a "beat" frequency of sorts, but it only provides Doppler information. In pulsed Doppler radar instead of slopes (FMCW) we have different PRF to resolve range and velocity ambiguities should they exist. Pulsed Doppler radar are usually divided in low-medium-high PRF regimes and have different range/Doppler ambiguity implications. The MIT Lincoln Lab Radar Series I think explains this well.
An important but challenging concept to understand is that a radar can receive signals from any direction, at any time, and with any frequency shift (subject to constraints). The radar's designer attempts to solve these ambiguities (or mitigate their occurrences) through various signal processing and hardware design schemes.
@@jonathananderson730 Thanks for the explanation. I understand pulsed Doppler radar uses the shift of RPF to solve the velocity, but it introduces velocity ambiguity. What I am think of is solving beat frequency on the other hand seems it can resolve velocity ambiguity. And waveform slope doesn't have to be used because pulsed Doppler radar can solve the range by solving the time delay of the receiving signal. So my question is, is my theory correct or is there any difficulty to utilize beat frequency for pulsed Doppler radar?
BTW I am not a radar engineer, but a combat sims player who is so curious about how radar works.
@@YYCH2 I'm not sure I get where you are coming from but we can't typically get an accurate instantaneous frequency difference between the receive and transmitted carrier over a single pulse. Often this measurable difference has large errors that make it impractical to attempt to extract the Doppler from a single pulse.
@@YYCH2 I think I get what Jonathan means. Indeed most RADAR systems utilizing Doppler processing will use the output of a frequency mixer to pass Doppler frequency directly to the signal processing system-this lower-bandwidth signal can be more easily worked with for signal processing compared to the GHz-band raw signals. (After a fashion, the sample rate of the receiver acts as a mixer/filter anyways.) However, both systems will have a similar problem-the sampling rate of the received signal. Since the samples are being taken at the intervals of the pulse width, this puts a fundamental limit on the maximum non-aliased frequency that can be determined (Nyquist-Shannon sampling theorem), which is half that of the sampling frequency. If your Doppler beat frequency is higher than this sample rate, it will alias to a lower frequency and make for ambiguous velocity anyways.
That being said, if you wanted to get a more complex spectrum from the pulse for things like non-cooperative target recognition and such, you could theoretically sample at a higher rate than the pulse width, but then you have to deal with more complex interactions between the target and receiver, determining which of the received frequencies is actually due to "real" target motion and not moving parts on the target, and so on.
@@eddievhfan1984 Thank you so much for the detailed break down.
Excellent...too good 👍
I am a war thunder player and I am watching this shit at 5am😂
Same. Same
"just notch bro"
Ah yes radar good
100 nano seconds is actually 0.1 micro seconds.
Do the links to the files also work in Octave? Asking for a friend. 😬
thanks