IWR1642BOOST: False values while reading object data over UART

Shantanu Sampath

Part Number: IWR1642BOOST
Other Parts Discussed in Thread: IWR1642

Hi all

I have written a python script to read the data stream over USB, convert the object data into (x,y) coordinates and to display it on the console, using the IWR1642 EVM. I tested it in an empty room and got some bizarre results (Ex. range of 355m, etc). I have attached my python script and sample output. Please tell me where I went wrong.

Fullscreen output.txt Download

Read Data:
 [2, 1, 4, 3, 6, 5, 8, 7, 5, 0, 0, 1, 192, 2, 0, 0, 66, 22, 10, 0, 215, 151, 0, 0, 190, 102, 4, 161, 8, 0, 0, 0, 3, 0, 0, 0, 1, 0, 0, 0, 100, 0, 0, 0, 8, 0, 7, 0, 1, 0, 0, 0, 37, 0, 228, 255, 13, 0, 0, 0, 27, 0, 0, 0, 15, 0, 185, 0, 57, 3, 0, 0, 51, 0, 0, 0, 21, 0, 187, 2, 156, 5, 0, 0, 59, 0, 0, 0, 24, 0, 181, 2, 177, 6, 0, 0, 95, 0, 0, 0, 2, 0, 162, 3, 10, 11, 0, 0, 255, 0, 0, 0, 15, 0, 186, 227, 48, 13, 0, 0, 27, 0, 0, 0, 11, 0, 246, 1, 168, 2, 0, 0, 51, 0, 0, 0, 15, 0, 100, 0, 58, 6, 0, 0, 2, 0, 0, 0, 0, 2, 0, 0, 216, 123, 160, 122, 62, 108, 58, 80, 98, 89, 101, 94, 250, 101, 104, 100, 196, 93, 73, 100, 88, 102, 244, 101, 182, 102, 99, 100, 49, 96, 111, 92, 126, 95, 56, 90, 164, 90, 170, 91, 80, 93, 242, 88, 36, 96, 235, 100, 11, 107, 59, 109, 107, 108, 62, 113, 37, 113, 13, 107, 24, 94, 80, 91, 73, 97, 205, 94, 175, 86, 166, 84, 87, 81, 59, 82, 191, 80, 115, 85, 65, 84, 70, 80, 70, 80, 57, 82, 246, 83, 116, 75, 185, 66, 104, 73, 44, 76, 150, 95, 246, 112, 42, 117, 224, 108, 143, 84, 159, 76, 164, 71, 134, 74, 251, 86, 29, 108, 81, 117, 1, 115, 85, 100, 165, 93, 150, 92, 220, 83, 0, 78, 52, 74, 229, 74, 163, 77, 32, 80, 126, 77, 123, 69, 181, 74, 154, 72, 43, 69, 87, 68, 28, 71, 113, 72, 3, 71, 250, 73, 9, 70, 15, 69, 144, 67, 191, 70, 7, 75, 51, 74, 112, 70, 148, 71, 232, 69, 218, 74, 227, 72, 234, 67, 99, 73, 232, 73, 73, 87, 109, 91, 143, 86, 246, 76, 117, 73, 72, 69, 166, 71, 175, 69, 159, 66, 126, 67, 176, 67, 176, 68, 22, 70, 219, 66, 186, 66, 90, 65, 163, 64, 23, 64, 131, 70, 146, 66, 219, 66, 119, 70, 83, 70, 39, 72, 150, 70, 244, 70, 66, 67, 39, 66, 182, 67, 202, 62, 22, 70, 183, 65, 52, 66, 115, 69, 250, 67, 45, 67, 32, 69, 194, 71, 141, 69, 11, 65, 133, 64, 218, 65, 141, 59, 232, 66, 147, 69, 180, 69, 78, 63, 35, 68, 194, 65, 205, 60, 250, 65, 248, 67, 99, 66, 53, 71, 210, 70, 106, 63, 190, 64, 223, 67, 182, 67, 246, 65, 111, 68, 23, 70, 147, 66, 181, 64, 48, 65, 146, 66, 42, 65, 109, 68, 39, 68, 225, 64, 70, 63, 226, 63, 105, 65, 84, 59, 139, 69, 78, 66, 161, 68, 236, 64, 201, 61, 112, 67, 246, 67, 251, 65, 239, 64, 119, 67, 164, 67, 20, 67, 62, 66, 93, 63, 163, 67, 125, 69] 

Detected Object data: 
 [1, 0, 0, 0, 37, 0, 228, 255, 13, 0, 0, 0, 27, 0, 0, 0, 15, 0, 185, 0, 57, 3, 0, 0, 51, 0, 0, 0, 21, 0, 187, 2, 156, 5, 0, 0, 59, 0, 0, 0, 24, 0, 181, 2, 177, 6, 0, 0, 95, 0, 0, 0, 2, 0, 162, 3, 10, 11, 0, 0, 255, 0, 0, 0, 15, 0, 186, 227, 48, 13, 0, 0, 27, 0, 0, 0, 11, 0, 246, 1, 168, 2, 0, 0, 51, 0, 0, 0, 15, 0, 100, 0, 58, 6, 0, 0] 

Range 		 Velocity		 X_coor 		 Y_coor

[[  7.81250000e+00   0.00000000e+00  -1.08527132e-01   5.03875969e-02]
 [  2.10937500e+02   0.00000000e+00   7.17054264e-01   3.19767442e+00]
 [  3.98437500e+02   0.00000000e+00   2.70930233e+00   5.56589147e+00]
 [  4.60937500e+02   0.00000000e+00   2.68604651e+00   6.63953488e+00]
 [  7.42187500e+02   0.00000000e+00   3.60465116e+00   1.09534884e+01]
 [  1.99218750e+03   0.00000000e+00  -2.80542636e+01   1.30852713e+01]
 [  2.10937500e+02   0.00000000e+00   1.94573643e+00   2.63565891e+00]
 [  3.98437500e+02   0.00000000e+00   3.87596899e-01   6.17829457e+00]]

import serial
import sys
import string
import struct
import time
import numpy as np
import math
from decimal import *
# # # # # # # # # # # # # # # # # # # # # #
# GLOBAL VARIABLES
OBJ_STRUCT_SIZE_BYTES = 12;
magicWord = [2,1,4,3,6,5,8,7]
word = [1,256,65536,16777216]
#word = [1, 16, 256, 65536]
word1 = [1,256]
# # # # # # # # # # # # # # # # # # # # # #
#####CONFIGURATION PARAMETERS FROM GUI DEMO #########
txChannelEn 		= 3 #channelCfg #2
numTxAnt 			= 2
digOutSampleRate 	= 5209 # profileCfg #11
freqSlopeConst 		= 70 #profileCfg #8
numAdcSamples 		= 256 #profileCfg #10
numRangeBins 		= 8	#pow2roundup(numAdcSamples) 
startFreq 			= 77 #profileCfg #2
idleTime 			= 429 #profileCfg #3
rampEndTime 		= 57.14 #profileCfg #5
chirpEndIdx 		= 1 #frameCfg #2
chirpStartIdx 		= 0 #frameCfg #1
numLoops 			= 16 #frameCfg #3
numChirpsPerFrame 	= (chirpEndIdx -chirpStartIdx + 1) *numLoops
numDopplerBins 		=  numChirpsPerFrame/numTxAnt
########################################################

rangeIdxToMeters =  3e8 * digOutSampleRate * 1e3 /(2 * freqSlopeConst * 1e12 * numRangeBins)
dopplerResolutionMps = 3e8 / (2*startFreq*1e9 *(idleTime + rampEndTime) * 1e-6 * numDopplerBins * numTxAnt)

ser = serial.Serial(port ='COM3', baudrate = 921600, timeout = 0.2)

while True:
#if 1>0:
	#byteVec = ser.read(ser.inWaiting())
	byteVec = ser.read(520)
	x = list(byteVec)
	startIdx = []
	#print('Read Data:\n',x, '\n')
	
	for ind in (i for i,e in enumerate(x) if e==magicWord[0]):	#To obtain the start indices of all magic words
		if x[ind:ind+8]==magicWord:
			startIdx.append(ind)
	
	for j in startIdx: 			#For each packet, do the needful
		j = j + 8 				#Skip magic word
		j = j + 20 				#get to number of detected objects. Only TLV selected is Detected Points
		numDetObj = 0
		k = 1
		for k in range(4):
			if j+k >= len(x):
				break
			numDetObj = numDetObj+ x[j+k]*word[k] #Refer line 1220 of mmw_demo.m
		#print('NumDetObj= '+str(numDetObj)+'\n')
		j = j + 8 				#Exit the header part
		
		j = j+8					 # SKIP TLV INFORMATION
		if j +4>= len(x):		 # Account for exceeding the length of list
			break
		
		#detObj part Refer line 1245 of mmw_demo.m
		
		numObj = 0
		detObj = np.zeros([numDetObj, 5])
		
		numObj = x[j+1] + x[j+2]*256
		j = j+2
		xyzQFormat = 2^(x[j+1] + x[j+2]*256)
		j = j+2
		bytes = x[j: j+numDetObj*OBJ_STRUCT_SIZE_BYTES]
		#print('Detected Object data: \n', bytes, '\n')
		if len(bytes) >= numDetObj*OBJ_STRUCT_SIZE_BYTES: 
			j = j+numDetObj*OBJ_STRUCT_SIZE_BYTES
			#print(len(bytes)/numDetObj)
			bytes = np.array(bytes).reshape(numDetObj, OBJ_STRUCT_SIZE_BYTES)
			#print(bytes, '\n')
			#print('rangeIdxToMeters = ', rangeIdxToMeters, '\n')
			#print('xyzQFormat = ', xyzQFormat, '\n')
			detObj[:, 0] = (bytes[:, 0] + bytes[:, 1] * 256)* rangeIdxToMeters
			detObj[:, 1] = bytes[:, 2] + bytes[:, 3]*256
			for r in range(numDetObj):
				if detObj[r,1] > numDopplerBins/2-1:
					detObj[r,1] -= numDopplerBins
			#detObj[:,1,detObj[:,1] > numDopplerBins/2-1] =detObj[:,1,detObj[:,1] > numDopplerBins/2-1] - numDopplerBins; # Refer mmw_demo.m line 1265
			detObj[:, 1] *= dopplerResolutionMps   		
			detObj[:, 2] = (bytes[:, 6] + bytes[:, 7]*256)
			detObj[:, 3] = (bytes[:, 8] + bytes[:, 9]*256)
			for r in range(numDetObj):
				if detObj[r,2] > 32767:
					detObj[r,2] -= 65536
			#for r in range(numDetObj):
			#	if detObj[r,3] > 32767:
			#		detObj[r,3] -= 65536
			#detObj[:,2,detObj[:,2] > 32767] =  detObj[:,2, detObj[:,2] > 32767] - 65536
			detObj[:,3][detObj[:,3] > 32767] =  detObj[:,3][ detObj[:,3] > 32767] - 65536
			detObj[:,2] = detObj[:, 2] / xyzQFormat
			detObj[:,3] = detObj[:, 3] / xyzQFormat 
			
			for r in range(numDetObj):
				if detObj[r,1] == 0:
					detObj[r, 4] = -1
					continue
				detObj[r,4] = np.sqrt(detObj[r,2]**2 + detObj[r,3]**2)/abs(detObj[r,1])
			
			print("Range \t\t Velocity  \t\t X_coor \t\t  Y_coor  \tETA\n", np.around(detObj, decimals = 2), '\n')
		
		#break
	#break
ser.close()

over 9 years ago

0 db_woodall over 9 years ago

TI__Mastermind 21025 points

Hello,

I would recommend editing/browsing the Demo Visualizer code to see how it parses the data. Some tips:

Hit F12 on the browser, go to sources tab and then go to mmWave.js in the list of files. In this file you can look for process1 function.
To browse the code on your desktop, you can download the desktop version of the App from the gallery and then navigate to <users>\guicomposer\runtime\gcruntime.v4\mmWave_Demo_Visualizer\app directory.
You can go to the gallery (https://dev.ti.com/gallery/), locate mmWave Demo Visualizer and click on to import the project in your workspace. Now you can customize the GUI as you desire and/or change the backend javascript code as per your needs. Some helpful links:

0 db_woodall over 8 years ago

TI__Mastermind 21025 points

Hi, I'm closing this due to inactivity. If you have further questions, please start a new thread.

0 Marco Mascorro over 8 years ago

Prodigy 10 points

Hi Shantanu,

Were you able to see what's going on? I'm learning how to read the packets as well from serial, so any learnings from your side would be very helpful.

Thanks.

Sensors

Sensors forum

IWR1642BOOST: False values while reading object data over UART