In this part of the lab we are going to experiment with Digital modulation and communication. Network Communication systems have layered architechture. The bottom layer is the physical which implements the modulation. Here we will use your AFSK modules that you implemented in part II. In addition, we will leverage AX.25, which is an amateur-radio data-link layer protocol. AX.25 is a packet based protocol that will help us transmit data using packets. It implements basic synchronization, addressing, data encapsulation and some error detection. In the ham world, an implementation of AFSK and AX.25 together is also called a TNC ( Terminal Node Controller ). In the past TNC's were separate boxes that hams used to attach to their radios to communicate with packet-based-communication. Today, it is easy to implement TNC's in software using the computer's soundcard.... as you will see here!
%pylab
# Import functions and libraries
import numpy as np
import matplotlib.pyplot as plt
import pyaudio
import Queue
import threading,time
import sys
from numpy import pi
from numpy import sin
from numpy import zeros
from numpy import r_
from numpy import ones
from scipy import signal
from scipy import integrate
import threading,time
import multiprocessing
from numpy import mean
from numpy import power
from numpy.fft import fft
from numpy.fft import fftshift
from numpy.fft import ifft
from numpy.fft import ifftshift
import bitarray
from scipy.io.wavfile import read as wavread
import serial
import ax25
from fractions import gcd
%matplotlib inline
For the following tasks you will need the functions:
sg_plot
included below
myspectrogram_hann_ovlp
included below
play_audio
included below
record_audio
included below
printDevNumbers
included below
text2Morse
(from Part I, to Identify yourself before transmission)
afsk1200
(from Part II)
nc_afsk1200Demod
(from Part II)
PLL
(from Part II)
# function to compute least common multipler
def lcm(numbers):
return reduce(lambda x, y: (x*y)/gcd(x,y), numbers, 1)
# function to compute average power spectrum
def avgPS( x, N=256, fs=1):
M = floor(len(x)/N)
x_ = reshape(x[:M*N],(M,N)) * np.hamming(N)[None,:]
X = np.fft.fftshift(np.fft.fft(x_,axis=1),axes=1)
return r_[-N/2.0:N/2.0]/N*fs, mean(abs(X)**2,axis=0)
# Plot an image of the spectrogram y, with the axis labeled with time tl,
# and frequency fl
#
# t_range -- time axis label, nt samples
# f_range -- frequency axis label, nf samples
# y -- spectrogram, nf by nt array
# dbf -- Dynamic range of the spect
def sg_plot( t_range, f_range, y, dbf = 60, fig = None) :
eps = 10.0**(-dbf/20.0) # minimum signal
# find maximum
y_max = abs(y).max()
# compute 20*log magnitude, scaled to the max
y_log = 20.0 * np.log10( (abs( y ) / y_max)*(1-eps) + eps )
# rescale image intensity to 256
img = 256*(y_log + dbf)/dbf - 1
fig=figure(figsize=(16,6))
plt.imshow( np.flipud( 64.0*(y_log + dbf)/dbf ), extent= t_range + f_range ,cmap=plt.cm.gray, aspect='auto')
plt.xlabel('Time, s')
plt.ylabel('Frequency, Hz')
plt.tight_layout()
return fig
def myspectrogram_hann_ovlp(x, m, fs, fc,dbf = 60):
# Plot the spectrogram of x.
# First take the original signal x and split it into blocks of length m
# This corresponds to using a rectangular window %
isreal_bool = isreal(x).all()
# pad x up to a multiple of m
lx = len(x);
nt = (lx + m - 1) // m
x = append(x,zeros(-lx+nt*m))
x = x.reshape((m/2,nt*2), order='F')
x = concatenate((x,x),axis=0)
x = x.reshape((m*nt*2,1),order='F')
x = x[r_[m//2:len(x),ones(m//2)*(len(x)-1)].astype(int)].reshape((m,nt*2),order='F')
xmw = x * hanning(m)[:,None];
# frequency index
t_range = [0.0, lx / fs]
if isreal_bool:
f_range = [ fc, fs / 2.0 + fc]
xmf = np.fft.fft(xmw,len(xmw),axis=0)
sg_plot(t_range, f_range, xmf[0:m/2,:],dbf=dbf)
print 1
else:
f_range = [-fs / 2.0 + fc, fs / 2.0 + fc]
xmf = np.fft.fftshift( np.fft.fft( xmw ,len(xmw),axis=0), axes=0 )
sg_plot(t_range, f_range, xmf,dbf = dbf)
return t_range, f_range, xmf
def play_audio( Q,ctrlQ ,p, fs , dev, ser="", keydelay=0.1):
# play_audio plays audio with sampling rate = fs
# Q - A queue object from which to play
# ctrlQ - A queue object for ending the thread
# p - pyAudio object
# fs - sampling rate
# dev - device number
# ser - pyserial device to key the radio
# keydelay - delay after keying the radio
#
#
# There are two ways to end the thread:
# 1 - send "EOT" through the control queue. This is used to terminate the thread on demand
# 2 - send "EOT" through the data queue. This is used to terminate the thread when data is done.
#
# You can also key the radio either through the data queu and the control queue
# open output stream
ostream = p.open(format=pyaudio.paFloat32, channels=1, rate=int(fs),output=True,output_device_index=dev)
# play audio
while (1):
if not ctrlQ.empty():
# control queue
ctrlmd = ctrlQ.get()
if ctrlmd is "EOT" :
ostream.stop_stream()
ostream.close()
print("Closed play thread")
return;
elif (ctrlmd is "KEYOFF" and ser!=""):
ser.setDTR(0)
#print("keyoff\n")
elif (ctrlmd is "KEYON" and ser!=""):
ser.setDTR(1) # key PTT
#print("keyon\n")
time.sleep(keydelay) # wait 200ms (default) to let the power amp to ramp up
data = Q.get()
if (data is "EOT") :
ostream.stop_stream()
ostream.close()
print("Closed play thread")
return;
elif (data is "KEYOFF" and ser!=""):
ser.setDTR(0)
#print("keyoff\n")
elif (data is "KEYON" and ser!=""):
ser.setDTR(1) # key PTT
#print("keyon\n")
time.sleep(keydelay) # wait 200ms (default) to let the power amp to ramp up
else:
try:
ostream.write( data.astype(np.float32).tostring() )
except:
print("Exception")
break
def record_audio( queue,ctrlQ, p, fs ,dev,chunk=1024):
# record_audio records audio with sampling rate = fs
# queue - output data queue
# p - pyAudio object
# fs - sampling rate
# dev - device number
# chunk - chunks of samples at a time default 1024
#
# Example:
# fs = 44100
# Q = Queue.queue()
# p = pyaudio.PyAudio() #instantiate PyAudio
# record_audio( Q, p, fs, 1) #
# p.terminate() # terminate pyAudio
istream = p.open(format=pyaudio.paFloat32, channels=1, rate=int(fs),input=True,input_device_index=dev,frames_per_buffer=chunk)
# record audio in chunks and append to frames
frames = [];
while (1):
if not ctrlQ.empty():
ctrlmd = ctrlQ.get()
if ctrlmd is "EOT" :
istream.stop_stream()
istream.close()
print("Closed record thread")
return;
try: # when the pyaudio object is distroyed stops
data_str = istream.read(chunk) # read a chunk of data
except:
break
data_flt = np.fromstring( data_str, 'float32' ) # convert string to float
queue.put( data_flt ) # append to list
def text2Morse(text,fc,fs,dt):
CODE = {'A': '.-', 'B': '-...', 'C': '-.-.',
'D': '-..', 'E': '.', 'F': '..-.',
'G': '--.', 'H': '....', 'I': '..',
'J': '.---', 'K': '-.-', 'L': '.-..',
'M': '--', 'N': '-.', 'O': '---',
'P': '.--.', 'Q': '--.-', 'R': '.-.',
'S': '...', 'T': '-', 'U': '..-',
'V': '...-', 'W': '.--', 'X': '-..-',
'Y': '-.--', 'Z': '--..',
'0': '-----', '1': '.----', '2': '..---',
'3': '...--', '4': '....-', '5': '.....',
'6': '-....', '7': '--...', '8': '---..',
'9': '----.',
' ': ' ', "'": '.----.', '(': '-.--.-', ')': '-.--.-',
',': '--..--', '-': '-....-', '.': '.-.-.-',
'/': '-..-.', ':': '---...', ';': '-.-.-.',
'?': '..--..', '_': '..--.-'
}
Ndot= 1.0*fs*dt
Ndah = 3*Ndot
sdot = sin(2*pi*fc*r_[0.0:Ndot]/fs)
sdah = sin(2*pi*fc*r_[0.0:Ndah]/fs)
# convert to dit dah
mrs = ""
for char in text:
mrs = mrs + CODE[char.upper()] + "*"
sig = zeros(1)
for char in mrs:
if char == " ":
sig = concatenate((sig,zeros(Ndot*7)))
if char == "*":
sig = concatenate((sig,zeros(Ndot*3)))
if char == ".":
sig = concatenate((sig,sdot,zeros(Ndot)))
if char == "-":
sig = concatenate((sig,sdah,zeros(Ndot)))
return sig
def printDevNumbers(p):
N = p.get_device_count()
for n in range(0,N):
name = p.get_device_info_by_index(n).get('name')
print n, name
Insert your
afsk1200
(from Part II)
nc_afsk1200Demod
(from Part II)
PLL
(from Part II)
Below
def afsk1200(bits, fs = 48000):
# the function will take a bitarray of bits and will output an AFSK1200 modulated signal of them, sampled at fs
# Inputs:
# bits - bitarray of bits
# fs - sampling rate
# Outputs:
# sig - returns afsk1200 modulated signal
return sig
def nc_afsk1200Demod(sig, fs=48000.0, TBW=2.0):
# non-coherent demodulation of afsk1200
# function returns the NRZ (without rectifying it)
#
# sig - signal
# baud - The bitrate. Default 1200
# fs - sampling rate in Hz
# TBW - TBW product of the filters
#
# Returns:
# NRZ
N = (int(fs/1200*TBW)//2)*2+1
return NRZ
def PLL(NRZa, a = 0.74 , fs = 48000, baud = 1200):
return idx.astype(int32)
Now, similarly to before, find the audio interface numbers. And intitialize the variables: dusb_in, dusb_out, din, dout
p = pyaudio.PyAudio()
printDevNumbers(p)
p.terminate()
# CHANGE!!!!
dusb_in = 4
dusb_out = 4
din = 1
dout = 2
Initialize serial port
if sys.platform == 'darwin': # Mac
s = serial.Serial(port='/dev/tty.SLAB_USBtoUART')
else: #windows
s = serial.Serial(port='COM1') # CHANGE !!!!!!
Before we go to demodulate AFSK1200 we will construct data in the form of AX.25 packets. The structure of the AX.25 packet, and in particular the flag that starts and ends a frame will help us sync to the beginning of the packet for accurate demodulation. The AX.25 protocol uses several measures for standartization and to improve detection and demodulation of packets. We will go over them now:
AX.25 does not encode NRZ '1's and '0's in the usual mark and space frequencies. Instead it uses a scheme called NRZI, or non-return to zero inverted. NRZI encodes a '0' bit as a change from mark to space, or space to mark. A '1' is encoded as no change. To encode an AX.25 packet we need to convert our bit sequence to an NRZI one. For example, an original bit stream of 11011000 should be first converted to 11000101 (initial state is 1).
Because a '1' is represented by no change, sending a long string of '1's would result in a constant signal. This may result in a receiver drifting out of sync with the transmitter. In order to circumvent this, the encoder performs bit stuffing before transmition by placing a bit '0' after every fifth '1' in the the stream. The decoder does bit unstuffing and removes the extra '0's. Bit stuffing is performed before converting to NRZI.
Bytes are sent least-significant-bit first
A Flag field starts and ends each packet. It is a unique sequence and is used to detect the beginning and the end of packets. The flag consists of the bit sequence: 0x7E or: 01111110. The flag is an exception in which no bit stuffing is performed. Therefore it is the only place in the packet in which 6 consequitive '1's appear. In NRZI it will translate to a time interval of 7 bits between zero-corssing of the non-coerent detector output. This means that we can use the flag sequence to uniquly detect a packet.
The Ax.25 protocol defines several type of frames. We will use the Un-numbered Information (UI) frame as defined in the protocol. UI frame is used for connectionless mode, where AX.25 frmaes are transmitted without expecting any response, and reception is not guaranteed. This is similar to UDP in concept (however UDP is atransport layer protocol). The UI frame that is used in the Automatic Positioning and Reporting System (APRS) protocol and has 9 fields of data:
flag | Dest. Addr. | Src. Addr. | Digipeater Addresses | Control field | ID | Information Field | FCS | Flag |
---|---|---|---|---|---|---|---|---|
1 | 7 | 7 | 56 | 1 | 1 | 256 | 2 | 1 |
Of importance are the Source address, which are your call sign, the Information Field which is the packet payload, the FCS which is a error checksum.
The FCS field is always the last two bytes in the packet. They are a checksum that can be used to determine the integrity of the packet.
APRS is a ham packet-based system for real-time tactical digital communication for local area. APRS uses the AX.25 protocol in its core. Using APRS you can report position, status, and send messages that other hams or stations with APRS capability will be able to decode, interpret and display the information. APRS also provides means of packet repeating (Digipeters) alongside with internet terminal nodes. Some radio manufacturers saw the potential and included APRS in some of their products as well. Go to this website: https://aprs.fi to see the APRS activity in the surrounding area that is aggregated from the internet nodes. You will see fixed stations, weather stations as well as mobile operators in your area. We will use the website to confirm that our transmitted packets were received.
The national APRS frequency is 144.39MHz (ch-117 on your radio) There is much activity and infrastracture transmitting and listenning to that frequency. The international space station also has an APRS digipeter on board operating at 145.825MHZ (ch-50 on your radio). You can also use AX.25 on any of the digital or experimental channels in the bandplan -- though you will have to coordinate if you want anybody to hear you!
To prevent interference with others while we experiment with APRS, I have set up an APRS station in Cory hall. It will monitor the DGTLU1 channel (ch-113), 433.550MHz on your radio. It has an internet link as well. Use this frequency first, and only if it does not work out, switch to the APRS frequency 144.39 (ch-117). We will do our best to make it work.
For APRS packets, the Destination address field is not used for routing. Instead it represents the version of the APRS software you are using. In order to be decoded by receivers in the APRS network it must start with the letters AP. We will use APCAL
just for fun.
The source address is your call sign. The digipeter addresses require some explenation but the fields 'WIDE1-1,WIDE2-1' will result in the packet being digipeted a maximum of 2 hops. In dense population areas like the bay area 'WIDE1-1' is often enough to get your packet to its destination. The Control and ID fields in APRS packets are fixed to "\x03" and "\xF0" respectively. The FCS field is the checksum field, that is used to verify the packet integrity, as defined by AX.25. The flag fields are the usual 01111110.
We have prepared for you code that generates valid bitstream of AX.25 packets from the appropriate fields as well as decode the fields from a bitstream. The code is a modification of code originally written by: Greg Albrecht W2GMD. You will have to download the code from the class website
Below is code that will download it for you:
import urllib, ssl
testfile = urllib.URLopener()
testfile.context = ssl._create_unverified_context()
testfile.retrieve("https://inst.eecs.berkeley.edu/~ee123/sp16/lab/lab5/ax25.py", 'ax25.py')
The information field of the packet are 256 Bytes payload that contain the information you want to send. We will go over some of the information that is needed to construct valid and useful information field messeges. There are several digipeters in the bay area that have internet terminal nodes. These implement several "fun" and useful services. For example, you can send a position report that would show up on a google map. You can also send a short EMAIL or an SMS test message by sending an APRS packet. In fact, we have out own on Cory Hall!
How a node or a client interprets your packet depends on the information field structure. There are three types of packets: Position, Status and Messages
Just begin your packet line with a Colon and a 9 character TOADDRESS, another colon and then text. The TOADDRESS must have 9 characters and should be filled with spaces if it is smaller.
Examples of messages:
The "-----" are blank spaces to fill the space to 9 characters.
You can report your position to people on the APRS system. If your report is picked up by a node it will show up on http://www.aprs.fi. The basic format of a position packet is:
! or = symbols | Lattitude 8 chars | / | Longitude 9 chars | icon 1 char | Comment max 43 chars |
---|---|---|---|---|---|
= | 3752.50N | / | 12215.43W | K | Shows a school symbol on Cory Hall position |
= | 3752.45N | / | 12215.98W | [ | Shows a person walking on Oxford and Hearst |
= | 2759.16N | / | 08655.30E | [ | I'm on the top of the world! (Mt. Everest) |
The latitude format is expressed as a fixed 8-character field, in degrees and decimal minutes (up to two decimal places), followed by a letter N for north and S for south. Latitude minutes are expressed as whole minutes and hundredth of a minute, separated by a decimal point. Longitude is expressed as a fixed 9-character field, in degrees and decimal minutes (to two decimal places), followed by the letter E for east or W for west. Longitude degrees are in the range 000 to 180. Longitude minutes are expressed as whole minutes and hundredths of a minute, separated by a decimal point.
In generic format:
For example Cory Hall is at N37° 52.5022', W122° 15.4395'. So the position is encoded as: 3752.50N/12215.43W
You can go to http://www.gpsvisualizer.com/geocode to find the coordinates of an address. Note: use the degree, minutes representation, not the decimal one.
a 1 character icon is provided after the coordinates. This will show an icon on the http://aprs.fi maps. Here are some useful ones:
Examples:
a school symbol
The empty squares ❏ represent space charcter
A status packet starts with '>' character. It wil show on APRS equipped radios.
Examples:
In the following section we will construct a valid (bitstuffed) bitstream from the different APRS packet fields. We will convert to an NRZI representation and modulate to generate a valid AFSK1200 APRS Packet. We will transmit it over the radio and look at http://aprs.fi to see if it was received by a node.
The following code shows you how to construct a message packet that will tell a digipeter to send you an email. Make sure you fill the correct information in the fields. The bitstream will already be bitsuffed with zeroes.
import ax25
callsign = "REPLACE THIS STRING WITH YOUR CALLSIGN"
Digi =b'WIDE1-1'
dest = "APCAL"
# Uncomment to Send Email
#info = ":EMAIL :mlustig@eecs.berkeley.edu Hi, its YOURNAME, what a great lab!"
# Uncomment to Send an SMS message to a phone number
info = ":SMSGTE :@9255873969 Hi James, this is YOURNAME. COMMENT HERE"
#uncomment to show yourself on mt everest
#info = "=2759.16N/08655.30E[I'm on the top of the world"
#uncomment to send to everyone on the APRS system near you
#info = ":ALL : CQCQCQ I would like to talk to you!"
# uncomment to report position
#info = "=3752.50N/12215.43WlIm using a laptop in Cory Hall!"
# uncomment to send a status message
#info = ">I like radios"
packet = ax25.UI(
destination=dest,
source=callsign,
info=info,
digipeaters=Digi.split(b','),
)
print(packet.unparse())
Recall that AX.25 packets are sent with NRZI encoding in which a '0' is a change and a '1' is no change.
NRZI = NRZ2NRZI(bits)
, takes a standard bitarray stream and converts it to a bitarray stream representing '0's as change and '1's as unchanged. For example, an input of 0000111100 will result in 0101111101. This assume an initial state of '1'. NRZI2NRZ(NRZI, current = True)
, takes an NRZI bitstream and converts it to an NRZ. It assumes an inital state of current
with default current=1
.def NRZ2NRZI(NRZ):
NRZI = NRZ.copy()
current = True
for n in range(0,len(NRZ)):
if NRZ[n] :
NRZI[n] = current
else:
NRZI[n] = not(current)
current = NRZI[n]
return NRZI
def NRZI2NRZ(NRZI, current = True):
NRZ = NRZI.copy()
for n in range(0,len(NRZI)):
NRZ[n] = NRZI[n] == current
current = NRZI[n]
return NRZ
Most packet radio and TNC's will pad packets with extra flag
sequences before and after the packet, to allow for synchronization. Rarely, the padding would be with zero-bit sequences (This will translate into alternating between Mark and Space which also helps with synchronization). Zero-padding is a deviation from the APRS protocol, but is allowed in most receivers. We will also pad our packets.
Here's a python trick to generate 20 flags
prefix = bitarray.bitarray(tile([0,1,1,1,1,1,1,0],(20,)).tolist())
# your code
# code for playing audio
p = pyaudio.PyAudio()
Q = Queue.Queue()
ctrlQ = Queue.Queue() # dummy, but necessary!
Q.put(msg*0.5)
Q.put('EOT')
play_audio(Q, ctrlQ, p, 48000, dout )
p.terminate()
play_audio
use the option keydelay=0.3
This will add a delay of 300ms between keying and playing the sound. If it works for you, you can later reduce it, or use the default which is 100ms. The idea here is to let time for the power amplifier to ramp up before transmitting. Padding the packet also helps. You do not need to identify yourself, since your packet already does identify you!
p = pyaudio.PyAudio()
Qout = Queue.Queue()
ctrlQ = Queue.Queue()
Qout.put("KEYON")
Qout.put(msg*0.2) # pick the gain that you calibrated in lab 5 part I
Qout.put("KEYOFF")
Qout.put("EOT")
play_audio( Qout ,ctrlQ ,p, 48000 , dusb_out, s, keydelay=0.3)
time.sleep(1)
p.terminate()
Now that we know how to create AX.25 and APRS packets, know how to AFSK1200 modulate them, know how to demodulate AFSK1200 as well, we can move forward to receiving and decoding packets. By the end of that we will have a fully functioning communication system.
Download the file ISSpkt.wav
. It contains an APRS packet I recorded on one of the ISS flybyes. Load it to your workspace using the function wavread
, which we imported from scipy.io
.
(ISSpkt.wav)
testfile = urllib.URLopener()
testfile.context = ssl._create_unverified_context()
testfile.retrieve("https://inst.eecs.berkeley.edu/~ee123/sp16/lab/lab5/ISSpkt.wav", 'ISSpkt.wav')
fs,sig = wavread("ISSpkt.wav")
We will now automate the packet decoding by writing some functions that implement portions of the process.
nc_afsk1200Demod
on the ISS packet to get the demodulated "analog" NRZI# your code
Once we demodulated the signal, the next task is to sample and look for packets in it! There are many ways to do so. We chose one similar to the DireWolfe implementation:
Below are several functions we wrote for you:
findPackets(bits)
- Searches a bitstream for possible valid packetsgenfcs(bits)
- generates a checksum for validating packetsdecodeAX25(bits)
- Parses a packet bitstream into fields and decodes themdef findPackets(bits):
# function take a bitarray and looks for AX.25 packets in it.
# It implements a 2-state machine of searching for flag or collecting packets
flg = bitarray.bitarray([0,1,1,1,1,1,1,0])
packets = []
n = 0
pktcounter = 0
packet = []
state = 'search'
# Loop over bits
while (n < len(bits)-7) :
# default state is searching for packets
if state is 'search':
# look for 1111110, because can't be sure if the first zero is decoded
# well if the packet is not padded.
if bits[n:n+7] == flg[1:]:
# flag detected, so switch state to collecting bits in a packet
# start by copying the flag to the packet
# start counter to count the number of bits in the packet
state = 'pkt'
packet=flg.copy()
pktcounter = 8
# Advance to the end of the flag
n = n + 7
else:
# flag was not found, advance by 1
n = n + 1
# state is to collect packet data.
elif state is 'pkt':
# Check if we reached a flag by comparing with 0111111
# 6 times ones is not allowed in a packet, hence it must be a flag (if there's no error)
if bits[n:n+7] == flg[:7]:
# Flag detected, check if packet is longer than some minimum
if pktcounter > 200:
# End of packet reached! append packet to list and switch to searching state
# We don't advance pointer since this our packet might have been
# flase detection and this flag could be the beginning of a real packet
state = 'search'
packet.extend(flg)
packets.append(packet.copy())
else:
# packet is too short! false alarm. Keep searching
# We don't advance pointer since this this flag could be the beginning of a real packet
state = 'search'
# No flag, so collect the bit and add to the packet
else:
# check if packet is too long... if so, must be false alarm
if pktcounter < 2680:
# Not a false alarm, collect the bit and advance pointer
packet.append(bits[n])
pktcounter = pktcounter + 1
n = n + 1
else:
#runaway packet, switch state to searching, and advance pointer
state = 'search'
n = n + 1
return packets
# function to generate a checksum for validating packets
def genfcs(bits):
# Generates a checksum from packet bits
fcs = ax25.FCS()
for bit in bits:
fcs.update_bit(bit)
digest = bitarray.bitarray(endian="little")
digest.frombytes(fcs.digest())
return digest
# function to parse packet bits to information
def decodeAX25(bits):
ax = ax25.AX25()
ax.info = "bad packet"
bitsu = ax25.bit_unstuff(bits[8:-8])
if (genfcs(bitsu[:-16]).tobytes() == bitsu[-16:].tobytes()) == False:
#print("failed fcs")
return ax
bytes = bitsu.tobytes()
ax.destination = ax.callsign_decode(bitsu[:56])
source = ax.callsign_decode(bitsu[56:112])
if source[-1].isdigit() and source[-1]!="0":
ax.source = b"".join((source[:-1],'-',source[-1]))
else:
ax.source = source[:-1]
digilen=0
if bytes[14]=='\x03' and bytes[15]=='\xf0':
digilen = 0
else:
for n in range(14,len(bytes)-1):
if ord(bytes[n]) & 1:
digilen = (n-14)+1
break
# if digilen > 56:
# return ax
ax.digipeaters = ax.callsign_decode(bitsu[112:112+digilen*8])
ax.info = bitsu[112+digilen*8+16:-16].tobytes()
return ax
findPackets
to find packet in the bitstreamYou should get this when decoding properly:
1) |DEST:CQ |SRC:RS0ISS |DIGI: |>ARISS - International Space Station|
# your code here:
# code to display the packets
# Iterate over all packets, print the valid ones
npack = 0
for pkt in packets:
if len(pkt) > 200:
ax = decodeAX25(pkt)
if ax.info != 'bad packet':
npack = npack+1
print(str(npack)+") |DEST:"+ax.destination[:-1]+" |SRC:"+ax.source + " |DIGI:"+ax.digipeaters+" |",ax.info,"|")
Now, lets try our decoder on a noisy recording with many more packets.
ISS.wav
. It's a 3min recording. It should decode in less than about a minute. testfile = urllib.URLopener()
testfile.context = ssl._create_unverified_context()
testfile.retrieve("https://inst.eecs.berkeley.edu/~ee123/sp16/lab/lab5/ISS.wav", 'ISS.wav')
fs, sig = wavread("ISS.wav")
# your code here:
npack = 0
for pkt in packets:
if len(pkt) > 200:
ax = decodeAX25(pkt)
if ax.info != 'bad packet':
npack = npack+1
print(str(npack)+") |DEST:"+ax.destination[:-1]+" |SRC:"+ax.source + " |DIGI:"+ax.digipeaters+" |",ax.info,"|")
The next step would be to test a loop-back in which we send audio through the USB audio device when its output and input are connected through a cable. To do so,
This should look like this pictude below:
Tips:
npack = 0
for pkt in packets:
if len(pkt) > 200:
ax = decodeAX25(pkt)
if ax.info != 'bad packet':
npack = npack+1
print(str(npack)+") |DEST:"+ax.destination[:-1]+" |SRC:"+ax.source + " |DIGI:"+ax.digipeaters+" |",ax.info,"|")
Congratulations, you created a modem!
If you would like to attempt to get the GUI application working (just for fun), which uses your code to decode and encode APRS messages, with an easy to use interface, contact me (jamescarlson@berkeley.edu) and I will point you in the right direction.