AWR1843: noncoherent-integration and coherent-integration

Hi Bryant,

                Please give me some time to check on this request and let me get back to you next week.

Thanks and Regards,

Akshay.

Hi Bryant,

              The reason for going to non-coherent integration in MRR lab is time and memory constraints. If we use non-coherent integration, ideally the angle FFT applied is less (after non-coherent integration). In coherent integration, there is a need to apply Angle-FFT across all the pixels of range -doppler heatmap which is difficult considering time constraints. But the use of coherent integration, does help us to achieve better target detection. 

              Thus as a tradeoff between time and performance, we have gone with non-coherent integration in MRR lab. Let me please know if you have any questions.

Thanks and Regards,

Akshay.

Hi Akshay,

Thanks for your reply.

You said "ideally the angle FFT applied is less",i can not understand it.In MRR demo,we store the 1D-FFT result in radarcube,then 2D-FFT and through non-coherent integration to execute CFAR,after it we detcet the obj's range index and doppler index then through radarcube to excute angle FFT.So i think non-coherent integration and coherent integration don't cause an increase in time or memory as you said.

BR

Bryant

Hello，Akshay,

We have the ADC data that the radar actually collects,I read in a TI document that the gain of coherent accumulation is 1.5-2dB higher than that of non-coherent accumulation.How can we draw this conclusion quantitatively from ADC data collected by radar?

BR

Bryant

Hi Bryant,

               Can you please point me to that document? I think the ideal way to quantify would be to have a strong and weak objects at some distances and once you do a range and doppler FFT (which is the 1st dimensional and 2nd dimensional FFT), then do a coherent or non-coherent accumulation and then find the peaks (where object is present) using CFAR algorithm. Ideally in the coherent accumulation the weak object should show up well due to high SNR while in non-coherent accumulation the weak object might not show up due to low SNR in this case.

               Let us know if you have any questions.

Thanks and Regards,

Akshay.

Hi Bryant,

               You could also refer to this document for more information about coherent and non-coherent integration gain: https://cpn-us-w2.wpmucdn.com/sites.gatech.edu/dist/5/462/files/2016/12/Noncoherent-Integration-Gain-Approximations.pdf .

Thanks and Regards,

Akshay.

Hi,Akshay Kumar Chandrasekaran ,

Thanks a lot for your reply.

I can not open the pdf your offered.Can you offer me the pdf directly not the webs?

BR

Bryant

Hi Bryant,

               Please find the pdf attached:Noncoherent-Integration-Gain-Approximations.pdf

Thanks and Regards,

Akshay.