DCA1000EVM: Strange output when processing raw data from AWR1642

HI, Charlie:  

Are you using radar studio GUI for data capture?  If possible, please send me the range-Doppler heatmap plot in the radar studio postProc display.   

Best,

Zigang

Hi Zigang ,

Thanks for your response. No I am not using the radar studio or mmwave studio I am using the CLI software in Linux and then collecting the frames from the ethernet port.

Charlie

1) are you using OOB demo binary for LVDS raw data capture?

2) Is this a new problem?  or it is your first time to check the data?

3) Can you check the ADC samples cross different chirps for a particular antenna are on top of each other.  To prove you have stable connection and also you have catch the frame start. 

4) Can you send me a plot of you range FFT output or range-doppler output (amplitude in dB) for a particular antenna?  It is very possible that your data parsing has some problem

Best,

Zigang   

Hi Zigang,

No I am not using the OOB demo- where would I find it?
I will put those graphs together for you in a sec.

Thanks,

Charlie

Here is the plot of the absolute value of the Range after a single 1D fft across the samples within the chirp.

Charlie

This is the same process, but with a hamming window applied

there shouldn't be targets at these positions

HI, 

It seems that your data parsing script has some problem.  Can you let me know what binary you are loading/flashing to the EVM and what data format you are assuming for your post processing?

Best,

Zigang

I have not chnaged the binary- I assume I am using a standard one from the demo visualiser/mmwave studio. The contents of the CFG file that is sent to the sensor is  below. I assume that it is in two stream LVDS format with complex numbers, 4 rx 256 samples per chirp 16 chirps per frame. I also assume that the chirps alternate between the two tx antennas. I have also inserted my python script to receive the output data below. I also read that you receive two real numbers and then the corresponding two imaginary numbers.

Thank you again for your help,

Charlie

CFG

sensorStop
flushCfg
dfeDataOutputMode 1
channelCfg 15 3 0
adcCfg 2 1
adcbufCfg -1 0 1 1 1
profileCfg 0 77 385 7 49.03 0 0 18.75 1 256 6250 0 0 30
chirpCfg 0 0 0 0 0 0 0 1
chirpCfg 1 1 0 0 0 0 0 2
frameCfg 0 1 16 0 1000 1 0
lowPower 0 1
guiMonitor -1 1 1 0 0 0 1
cfarCfg -1 0 2 8 4 3 0 15 1
cfarCfg -1 1 0 4 2 3 1 15 1
multiObjBeamForming -1 1 0.5
clutterRemoval -1 0
calibDcRangeSig -1 0 -5 8 256
extendedMaxVelocity -1 0
bpmCfg -1 0 0 1
lvdsStreamCfg -1 1 1 1
compRangeBiasAndRxChanPhase 0.0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0
measureRangeBiasAndRxChanPhase 0 1.5 0.2
CQRxSatMonitor 0 3 14 123 0
CQSigImgMonitor 0 123 14
analogMonitor 0 0
aoaFovCfg -1 -90 90 -90 90
cfarFovCfg -1 0 0 34.99
cfarFovCfg -1 1 -1 1.00
calibData 0 0 0
sensorStart

Python script

import os,sys,socket,time,numpy as np,struct, matplotlib.pyplot as plt,math


def prep():
    os.system('sudo ifconfig eth0 192.168.33.30 netmask 255.255.255.0')
    s=socket.socket(socket.AF_INET, socket.SOCK_DGRAM, socket.IPPROTO_UDP)
    s.bind(('192.168.33.30',4098))
    return(s)

def readpacket(s):
    
    r=s.recv(1466)

    len_data=(1456//2)
    
    data=[0]*(len_data)
    
    packet_number=struct.unpack('<1l',r[:4])[0]#[42:46])
    byte_count=struct.unpack('>Q',b'\x00\x00' + r[4:10])[0]#[46:52])
    for k in range(len_data):
        data[k]=struct.unpack('<i',r[10+k:14+k])[0]
    print(packet_number)
    return(data,packet_number,byte_count)
    
    
    #plt.plot(q,bc[0:100],'.')
    #plt.show()
def readframe(s):
    packets=[]
    packets_read=0
    
    
    
    ChirpsPerFrame=16
    SamplesPerChirp=256
    RealImag=2
    rxChannels=4
    txChannels=2
    VirtualA=rxChannels*txChannels
    Bytes=2
    FrameLen=(ChirpsPerFrame*SamplesPerChirp*RealImag*rxChannels*Bytes*txChannels)
    BytesInPack=1456
    BytesInFrame=(FrameLen//BytesInPack)*BytesInPack
    PackInFrame=BytesInFrame//BytesInPack
    print(PackInFrame)
    
    
    
    #Find start of frame
    while True:
        
        data,packNum,ByteCount=readpacket(s)
        
        
        if ByteCount%(32768*2)==0:
            print('Start of frame')
            
            packets=packets+data
            
            startPack=packNum
            break
        else:
            print('Waiting for start of frame')
    while packNum-startPack<=91:
        #while data==None or packNum==None or ByteCount==None:
        data,packNum,BytesCount=readpacket(s)
        print(packNum-startPack)
        packets=packets+data
        #print(len(packets))
            
    Frame=np.empty((ChirpsPerFrame,8,SamplesPerChirp),dtype=complex)        
    ComplexPack=np.empty(len(packets)//2,dtype=complex)
    i=0
    j=0
    x=0
    TotChirp=1024
    for x in range(len(packets)):
        if x%4==0 or x%4 == 1:
            
            ComplexPack.real[i]=packets[x]
            i=i+1
        elif x%4==2 or x%4==3:
            ComplexPack.imag[j]=packets[x]
            j=j+1
        else:print('error')
    #print(ComplexPack)
        #print(len(packets), FrameLen//4)
    #Makes an array where each row corresponds to a virtual rx channel     
    for x in range(16):
  
        Frame[x,0,np.arange(SamplesPerChirp)]=ComplexPack[x*TotChirp:(SamplesPerChirp+x*TotChirp)]
        Frame[x,1,np.arange(SamplesPerChirp)]=ComplexPack[(SamplesPerChirp+x*TotChirp):(2*SamplesPerChirp+x*TotChirp)]
        Frame[x,2,np.arange(SamplesPerChirp)]=ComplexPack[(2*SamplesPerChirp+x*TotChirp):(3*SamplesPerChirp+x*TotChirp)]
        Frame[x,3,np.arange(SamplesPerChirp)]=ComplexPack[(3*SamplesPerChirp+x*TotChirp):(4*SamplesPerChirp+x*TotChirp)]
        Frame[x,4,np.arange(SamplesPerChirp)]=ComplexPack[(4*SamplesPerChirp+x*TotChirp):(5*SamplesPerChirp+x*TotChirp)]
        Frame[x,5,np.arange(SamplesPerChirp)]=ComplexPack[(5*SamplesPerChirp+x*TotChirp):(6*SamplesPerChirp+x*TotChirp)]
        Frame[x,6,np.arange(SamplesPerChirp)]=ComplexPack[(6*SamplesPerChirp+x*TotChirp):(7*SamplesPerChirp+x*TotChirp)]
        Frame[x,7,np.arange(SamplesPerChirp)]=ComplexPack[(7*SamplesPerChirp+x*TotChirp):(8*SamplesPerChirp+x*TotChirp)]

    return(Frame)#This returns the datacube

Hi Zigang,

I think you might be right that that is where the problem is. I have posted my json file below. i am confused what the bit at the end means after "dataportConfig"- could this have an effect?

Thank you,

Charlie

{"DCA1000Config":
{"dataLoggingMode": "raw",
"dataTransferMode": "LVDSCapture",
"dataCaptureMode": "ethernetStream",
"lvdsMode": 2,
"dataFormatMode": 3,
"packetDelay_us": 25,
"ethernetConfig": {
    "DCA1000IPAddress": "192.168.33.180",
    "DCA1000ConfigPort": 4096,
    "DCA1000DataPort": 4098},
"ethernetConfigUpdate": {
    "systemIPAddress": "192.168.33.30",
    "DCA1000IPAddress": "192.168.33.180",
    "DCA1000MACAddress": "12.34.56.78.90.12",
    "DCA1000ConfigPort": 4096,
    "DCA1000DataPort": 4098},
"captureConfig": {
    "fileBasePath": "/home/charlie/SourceCode/Release/data",
    "filePrefix": "Capture",
    "maxRecFileSize_MB": 1024,
    "sequenceNumberEnable": 1,
    "captureStopMode": "infinite",
    "bytesToCapture": 5000,
    "durationToCapture_ms": 5000,
    "framesToCapture":1},
"dataFormatConfig": {
    "MSBToggle": 0,
    "reorderEnable": 1,
    "laneFmtMap": 0,
    "dataPortConfig": [{
        "portIdx": 0,
        "dataType": "real"  },
        {"portIdx": 1,
        "dataType": "complex"  },
        {"portIdx": 2,
        "dataType": "real"  },
        {"portIdx": 3,
        "dataType": "real"  },
        {"portIdx": 4,
        "dataType": "complex"  }]
    }
    }
}

Hi Zigang,

Thank you so much for the help. That seems to have made it work.

this graph is more what I expected.

Charlie

You are right.  The dataType on my json file are all set to complex.  But I did not find any information on the DCA1000 CLI user guide 

C:\ti\mmwave_studio_02_01_01_00\mmWaveStudio\ReferenceCode\DCA1000\Docs\TI_DCA1000EVM_CLI_Software_UserGuide.pdf

"dataPortConfig": [
{
"portIdx": 0,
"dataType": "complex"
},
{
"portIdx": 1,
"dataType": "complex"
},
{
"portIdx": 2,
"dataType": "complex"
},
{
"portIdx": 3,
"dataType": "complex"
},
{
"portIdx": 4,
"dataType": "complex"
}
]

Best,

Zigang

Great, I am glad to hear that.  Do you have to change the dataType to "complex"?

Best,

Zigang

I have changed them all to complex and it doesnt seem to make much difference. I'm slightly confused by the output I've got still though. The plot I sent before was only half of the fft response, if I plot the full response I get a reflection across the middle of the graph like the plots below. Is this expected? Does it show that I have something specific wrong?

Charlie

It looks like the input is a real signal, not a complex signal.   Do you use the complex signal as FFT input?

Best,

Zigang

I used the absolute value of the complex output from the radar

Sorry thats wrong i used the complex data and then plotted the absolute value.

Do you know what "dataFormatMode": 3, in the json file means?

Please can find this information in the users guide located at: C:\ti\mmwave_studio_02_01_01_00\mmWaveStudio\ReferenceCode\DCA1000\Docs

Best,

Zigang

Sorry for the slow response, I have been busy doing something else. That info is very helpful, thank you. I was wondering if you know of a reference code for this application of processing the raw data to get the range, Doppler and azimuth information.

Thanks,

Charlie 

There is some python script located at: 

C:\ti\mmwave_sdk_03_05_00_04\packages\ti\demo\parser_scripts

And for all demo code, the source target code is available. 

Best,

Zigang

Thanks for that. What would you suggest the best method of calculating the range-azimuth information is?

Charlie