AWR1843AOP: AOA estimation

Hello, 

Sorry, I have addtional question. 

3) It seems that it isn't indicated the coordinates of TX antenna and RX anttenna in "aoaprocdsp.c" and "aoaprochwa.c". 

 For example, if the antennas coordinates change (in besides AOP device), does it have any affect AOA calculation?

Please advise. 

Thanks, 

Shoko 

Hi Sakaguchi-san,

Thank you for your patience. Please find my responses as follows:

1. The mmWave demo in the SDK uses angle FFT for estimating AoA. The algorithms you mentioned to correspond to the area scanner application which is a part of TI resource explorer.
2. The "aoaprocdsp.c" runs the AoA processing on the DSP core only whereas in the "aoaprochwa.c", the AoA processing is split between the DSP and the radar HWA module on the AWR1843AOP device.
3. Yes. The angle estimation changes completely if there is any change in the position or order or sequence of the Tx and Rx antenna.

Regards,

Kaushik

Hello Kaushik, 

Thank you for your response. 

I understand that the mmWave demo in the SDK uses "angle FFT". 

Can I ask more question about estimating AOA? 

> The angle estimation changes completely if there is any change in the position or order or sequence of the Tx and Rx antenna.

1) For example,

 A) Send chirp signals in the order TX1, TX2, TX3

 B) Send chirp signals in the order TX3, TX2, TX1 

 C) Send chirp signals in the order TX2, TX3, TX1 

 are there any differences for estimating AOA between these seaquence? 

2) For example, 

 A) Use AWR1843AOP in the same direction as in the below image 

 B) Use AWR1843AOP rotated 180 degrees from the below image (the order of chirp signal transmission is the same as A))

 are there any differences for estimating AOA between A) and B)? 

Please advise. 

Thank you. 

Shoko 

Hello Kaushik, 

When can you give me an answer? 

Thanks, 

Shoko 

Hello Shoko-san,

Please find my responses as follows:

1) Are there any differences for estimating AOA between these sequences? 
Yes. You would need to re-order the data before performing any angle processing based on the appropriate virtual antenna sequence following the physical placement of the antennas. If not, it does not have any meaningful information.

2) Are there any differences for estimating AOA between A) and B)? 
For a 180-degree shift along the depth-axis, the effective azimuth angle measured will incur a sign flip provided the processing algorithm remains exact. For example: 30 deg will be come -30 deg if the sensor facing forward is rotated by 180 deg along the boresight axis.

Based on your questions, please understand that the order of the antenna physically and virtually is fundamental information required to estimate angle from MIMO radars.

Let me know if there are further questions.

Regards,

Kaushik

Hello Kaushik, 

Thank you for your response. Could I ask you more? 

Regarding 1), if I want to estimate AOA of target using the antenna array like AWR1843AOP, 

should I use the following seaquence A) or B) rather than seaquence C)? 

 A) Send chirp signals in the order TX1, TX2, TX3 

 B) Send chirp signals in the order TX3, TX2, TX1 

 C) Send chirp signals in the order TX2, TX3, TX1 

Also, how much does the difference of seaquence affect the accuracy of the angle estimation? 

Regarding 2), understood. 

Please advise. 

Thanks, 

Shoko 

Hi Shoko-san,

The order in all cases (A, B and C) needs to be re-ordered before performing angle processing. If it is not in the right order, then there is no angle estimate that would be right. After re-ordering, the AoA estimate from all cases are theoretically the same.

Regards,

Kaushik

Hello Kaushik, 

Thanks for quick reply! 

If I re-order under all cases (A, B and C) before performing angle processing,

is there no difference in the accuracy of the angle estimation under all cased?  

Please advise. 

Regards, 

Shoko