Range, velocity, angle accuracy

Predrag Bratic

Other Parts Discussed in Thread: IWR1443BOOST

I have a question about accuracy formulas that are presented here in webinar(17/25 slide for range, and 19/25 velocity): https://training.ti.com/epd-pro-rap-mmwaveradar-adh-tr-webinar-eu

Angle accuracy formula is not presented. Can someone give a hint for calculating this forumulas for range and velocity accuracy and if possible angle accuracy?

Regards,

Predrag

over 7 years ago

0 Nitin Sakhuja over 7 years ago

TI__Mastermind 30405 points

Hi Predrag,

Accuracy (range, velocity or angle) is a function of the corresponding resolution and SNR. The difference between resolution and accuracy is discussed in the following thread:

IWR1443BOOST: Sub-millimeter accuracy distance measurement on a short range (from 5 to 30cm)

The following threads discuss accuracy in context of range, velocity and angle.

Linux/IWR1443BOOST: Change chirp configuration (.cfg file) using ROS (Discusses range accuracy. The formula for velocity accuracy is same, just replace range resolution with velocity resolution)
IWR1443BOOST: Resolution and accuracy of elevation angle for IWR1443BOOST (Discusses elevation angle accuracy for the IWR1443BOOST EVM.)
IWR1443BOOST: The Performance for the ROS visualizer lab (Discusses range and angle accuracy in context of the IWR1443BOOST EVM with example configuration).

Please mark the thread answered if the above answered your question otherwise get back in case more support is needed.

Thanks

-Nitin

0 Predrag Bratic over 7 years ago in reply to Nitin Sakhuja

Prodigy 210 points

Hi Nitin,

Thank you for these great posts! To conclude, accuracy estimation is clear to me, just to add in some posts you include SNR in dB in formula for calculating accuracy (in one post: Make sure to convert the SNR to linear scale (10^(SNRdb/10)) and I think the linear value is needed.
One more thing, accuracy is resolution/sqrt(2SNR) in general; why in webinar 3.6 value is in denominator of formulas? I assume this 1.8 multiplication is experimentally calculated or?

Thanks again,
Predrag

0 Nitin Sakhuja over 7 years ago in reply to Predrag Bratic

TI__Mastermind 30405 points

Hi Predrag,

That's correct, linear value should be used as mentioned in my other post. By the time I realized this, the original thread was locked so I couldn't fix the calculation. I will try to get the above thread unlocked and fix the calculation. Thanks for pointing it out.

I have requested the author of the Webinar for clarification regarding the extra constant multiplier in the denominator in slide 17. Will get back to you as soon as I have the response.

Regards
-Nitin

0 Predrag Bratic over 7 years ago in reply to Nitin Sakhuja

Prodigy 210 points

Thank you Nitin!

Regards,
Predrag

0 Predrag Bratic over 7 years ago in reply to Predrag Bratic

Prodigy 210 points

Can this correction factor be from effective bandwidth of a waveform of linear chirp signal?
In Skolnik book, it is said: The effective bandwidth of a waveform with a UNIFORM frequency spectrum
of width B (our chirp bandwidth) is beta = pi*B/sqrt(3) => pi/sqrt(3) = 1.81

regards,

Predrag

0 Nitin Sakhuja over 7 years ago in reply to Predrag Bratic

TI__Mastermind 30405 points

Hi Predrag,

I'm still waiting for a response from the author but I’m not sure if the above could be the reason, because intuitively, having a multiplier greater than 1 in the denominator effectively increases the bandwidth and hence accuracy.

Regards
-Nitin

0 Nitin Sakhuja over 7 years ago in reply to Nitin Sakhuja

TI__Mastermind 30405 points

Hi Predrag,

Sorry for the delay. I got a response from the author and your understanding is correct. The 1.8 factor in the denominator is indeed due to the effective bandwidth of the chirp with bandwidth B. Please mark this thread answered otherwise get back in case you have any more questions in this regard.

Regards
-Nitin

Sensors

Sensors forum

Range, velocity, angle accuracy