def __init__(self,**kwargs):
""" Constructor
Parameters
----------
fs : sampling frequency
fc : center frequency
beta : scale factor :math:`$\beta$`
Tns : time integration in nanoseconds
BGHz : Bandwidth in GHz
pfa : false alarm probability
wt : filter B ripple
R : Resistance for Tension/Power conversion
NF : Noise factor in dB (default 0dB)
gpass: gain in the pass band
gstop : gain in the stop band
Notes
-----
An energy detector is determined by 2 filter a bandpass elliptic filter
:math:`H_B(f)` and a lowpass filter which corresponds to the time
integration over duration :math:`T` :math:`H_T(f)`
"""
defaults = {'fsGHz':50,
'fcGHz':4,
'BGHz':3,
'Tns':10,
'pfa':0.01,
'beta':1,
'wt':0.01,
'gpass':0.5,
'gstop':30,
'NF':0,
'R':50}
for k in defaults:
if k not in kwargs:
kwargs[k]=defaults[k]
self.fsGHz = kwargs['fsGHz']
self.fcGHz = kwargs['fcGHz']
self.R = kwargs['R']
BGHz = kwargs['BGHz']
Tns = kwargs['Tns']
wt = kwargs['wt']
#
# Filter B : Input Band Pass Filter
#
ts = 1./self.fsGHz
fN = self.fsGHz/2.
w1 = (self.fcGHz-BGHz/2.)/fN
w2 = (self.fcGHz+BGHz/2.)/fN
wp = [w1,w2]
ws = [w1-wt,w2+wt]
self.filterB = DF()
self.filterB.ellip_bp(wp,ws,kwargs['gpass'],kwargs['gstop'])
#self.filterB.butter(order=6,w=wp,typ='bandpass')
#
# Filter T : Time Averaging Filter (Integrator) (Tns => N samples)
#
N = np.ceil(Tns/ts)
b = (Tns*1e-9/N)*ones(N)
self.NfilterT = len(b)
a = array([1])
self.filterT = DF(b,a)
#
#
#
self.BGHz = BGHz
self.Tns = Tns
self.beta = kwargs['beta']
self.pfa = kwargs['pfa']
self.NF = kwargs['NF']
#
# calculates ED moments, order and scale
#
# self.moment(typ='struc')
# kB = 1,3806488 × 10-23
kB = 1.3806488e-23
# TB : Noise temperature
TB = 290*10**(self.NF/10.)
N0 = kB*TB
self.order = 2*self.BGHz*self.Tns
self.muy = self.order*N0
self.vary = 4*self.BGHz*self.Tns*N0**2
self.scale = np.sqrt(self.vary/(2*self.order))
IP = ip.psd(Tpns=500) IP.plotdB(mask=True) IP = ip.psd flt=DF() fc = 4 fs = fsGHz fN = 50 B = 0.5 w1 = (fc-B/2)/fN w2 = (fc+B/2)/fN flt.butter(w=[w1,w2],typ='bandpass') flt.freqz() ### Creating a simple channel #fig = plt.figure(figsize=(10,3))
class ED(PyLayers):
"""
Energy Detector Class
This class implements an energy detector receiver
"""
def __init__(self,**kwargs):
""" Constructor
Parameters
----------
fs : sampling frequency
fc : center frequency
beta : scale factor :math:`$\beta$`
Tns : time integration in nanoseconds
BGHz : Bandwidth in GHz
pfa : false alarm probability
wt : filter B ripple
R : Resistance for Tension/Power conversion
NF : Noise factor in dB (default 0dB)
gpass: gain in the pass band
gstop : gain in the stop band
Notes
-----
An energy detector is determined by 2 filter a bandpass elliptic filter
:math:`H_B(f)` and a lowpass filter which corresponds to the time
integration over duration :math:`T` :math:`H_T(f)`
"""
defaults = {'fsGHz':50,
'fcGHz':4,
'BGHz':3,
'Tns':10,
'pfa':0.01,
'beta':1,
'wt':0.01,
'gpass':0.5,
'gstop':30,
'NF':0,
'R':50}
for k in defaults:
if k not in kwargs:
kwargs[k]=defaults[k]
self.fsGHz = kwargs['fsGHz']
self.fcGHz = kwargs['fcGHz']
self.R = kwargs['R']
BGHz = kwargs['BGHz']
Tns = kwargs['Tns']
wt = kwargs['wt']
#
# Filter B : Input Band Pass Filter
#
ts = 1./self.fsGHz
fN = self.fsGHz/2.
w1 = (self.fcGHz-BGHz/2.)/fN
w2 = (self.fcGHz+BGHz/2.)/fN
wp = [w1,w2]
ws = [w1-wt,w2+wt]
self.filterB = DF()
self.filterB.ellip_bp(wp,ws,kwargs['gpass'],kwargs['gstop'])
#self.filterB.butter(order=6,w=wp,typ='bandpass')
#
# Filter T : Time Averaging Filter (Integrator) (Tns => N samples)
#
N = np.ceil(Tns/ts)
b = (Tns*1e-9/N)*ones(N)
self.NfilterT = len(b)
a = array([1])
self.filterT = DF(b,a)
#
#
#
self.BGHz = BGHz
self.Tns = Tns
self.beta = kwargs['beta']
self.pfa = kwargs['pfa']
self.NF = kwargs['NF']
#
# calculates ED moments, order and scale
#
# self.moment(typ='struc')
# kB = 1,3806488 × 10-23
kB = 1.3806488e-23
# TB : Noise temperature
TB = 290*10**(self.NF/10.)
N0 = kB*TB
self.order = 2*self.BGHz*self.Tns
self.muy = self.order*N0
self.vary = 4*self.BGHz*self.Tns*N0**2
self.scale = np.sqrt(self.vary/(2*self.order))
def __repr__(self):
st = 'Energy Detector'+'\n'
st = st+'---------------'+'\n\n'
st = st+'Center frequency : '+str(self.fcGHz)+' GHz\n'
st = st+'Bandwidth : '+str(self.BGHz)+' GHz \n'
st = st+'Integration Time : '+str(self.Tns)+' ns \n'
st = st+'Resistance : '+str(self.R)+' Ohms\n'
st = st+'Mean (y) : '+str(self.muy/1e-9)+' nJ \n'
st = st+'Std(y) : '+str(np.sqrt(self.vary)/1e-9)+' nJ \n'
st = st+'Order (df) : '+str(self.order)+' \n'
st = st+'Scale : '+str(self.scale)+' \n'
st = st+'2BT : '+str(2*self.BGHz*self.Tns)+' \n'
return(st)
def apply(self,x,typ='ideal'):
r"""
Parameters
----------
x : input signal
Returns
-------
y : :math:`F_2\{\beta^2 F_1\{x\}^2\}`
:math:`F{x}` means a filtering of x with filter :math:`F`
"""
if typ=='real':
# B filtering
self.xf1 = self.filterB.filter(x.y)
# quadrator and scaling
self.xq = ((self.beta)**2)*(self.xf1*self.xf1)/self.R
# T filtering
self.y = bs.TUsignal(x.x,self.filterT.filter(self.xq))
if typ=='ideal':
# B filtering --> FH signal
X = x.fft()
self.gate = np.zeros(len(x.x))
u1 = np.where((x.x>(self.fcGHz-self.BGHz/2.)) & (x.x<(self.fcGHz+self.BGHz/2)))[0]
u2 = np.where((x.x>(self.fsGHz-self.fcGHz-self.BGHz/2.)) & (x.x<(self.fsGHz-self.fcGHz+self.BGHz/2)))[0]
self.gate[u1] = 1
self.gate[u2] = 1
X.y = X.y*self.gate
self.y = X.ifft()
#self.xq = ((self.beta)**2)*(self.xf1*self.xf1)/self.R
#self.y = bs.TUsignal(x.x,self.filterT.filter(self.xq))
return(self.y)
def moment(self,typ='struc'):
r""" evaluate mean and variance of output signal y
Parameters
----------
typ : string
'struc' structural evaluation of moments
'emp' empirical evaluation of moments
Notes
-----
:math:`R_y(\tau)=2R_x^2(\tau)+R_x^2(0)`
This function updates the order and scale or the Energy Detector
"""
# filterB bandpass filter
self.filterB.freqz(display=False)
# filterT time integration (=low pass filter)
self.filterT.freqz(display=False)
w = bs.Noise(fsGHz=self.fsGHz)
if typ=='struc':
# ACF of filter B
self.RB = self.filterB.wnxcorr(var=w.var)
# F.de Coulom T&T signal eq 8.148
# ACF of xq
self.Rx = self.RB*self.RB*2
# Power of xq
RB0= self.RB.y[0]
#
self.Rx.y = self.Rx.y+RB0**2
Phix = np.real(fft.fft(self.Rx.y))
HH = self.filterT.H.symH(0)
HH.y = HH.y/(self.Tns*1e-9)
Phiy = Phix * np.real(HH.y * np.conj(HH.y))
self.muy = self.RB.max()[0]*self.beta*self.beta
self.vary = mean(Phiy)
if typ=='emp':
self.muy = self.y.y[self.NfilterT:].mean()
self.vary = self.y.y[self.NfilterT:].var()
self.order = (2*(self.muy)**2)/self.vary
self.scale = np.sqrt(self.vary/(2*self.order))
def pdf(self,E,Ei=0):
""" calculates decision problem densities under H0 and H1
Parameters
----------
E : float
Legitimate Signal Energy
Ei : float
Interferer Signal Energy
See Also
--------
scipy.stats.chi2
scipy.stats.ncx2
"""
self.E = E
self.Ei = Ei
# H0 : Noise only (no interferer)
# central chi square
if Ei==0:
self.pH0 = st.chi2(self.order,
scale=self.scale)
else:
# H0 : Noise + interference
# non central chi square
self.pH0 = st.ncx2(self.order,
scale=self.scale,
nc=Ei/(self.scale))
# H1 : Signal + interference + Noise
# non central chi square
self.pH1 = st.ncx2(self.order,
scale=self.scale,
nc=(E+Ei)/(self.scale))
def errprob(self,E,Ei=0,thresh=0,typ='thresh'):
""" calculates error probability
Parameters
----------
E : float
Legitimate Signal Energy
Ei : float
Interferer Signal Energy
thresh : float
typ : string
'thresh'
'scale': for IEEE.802.11.6 standard
'both' : both kind of decision
Returns
-------
errp : error probability
Notes
-----
See Also
--------
scipy.stats.ncf
ED.pdf
"""
self.pdf(E,Ei=Ei)
if typ=='thresh':
self.p10 = self.pH0.sf(thresh)
self.p01 = 1-self.pH1.sf(thresh)
errp = 0.5*(self.p10+self.p01)
if typ=='scale':
ncf = st.ncf(dfn = self.order,
dfd = self.order,
scale = 1,
nc = self.E/self.scale)
errp = 1 - ncf.sf(1)
if typ=='both':
self.p10 = self.pH0.sf(thresh)
self.p01 = 1-self.pH1.sf(thresh)
pe1 = 0.5*(self.p10+self.p01)
ncf = st.ncf(dfn=self.order,
dfd=self.order,
scale=1,
nc=self.E/self.scale)
pe2= 1 - ncf.sf(1)
errp =(pe1,pe2)
return(errp)
def plot(self,E):
""" plot decision problem densities
Parameters
----------
E : float
value of energy
"""
self.pdf(E)
e = np.linspace(0,max(self.muy*5,2*self.E),200)
p0 = self.pH0.pdf(e)
p1 = self.pH1.pdf(e)
#plt.plot(10*np.log10(E),p0,'b',linewidth=4)
#plt.plot(10*np.log10(E),p1,'r',linewidth=4)
plt.plot(e,p0,'b',linewidth=4)
plt.plot(e,p1,'r',linewidth=4)
def roc(self,E,typ='thresh'):
""" determine receiver operating curve
Parameters
----------
E : float
Energy
typ : string
'thresh' | 'both' | 'scale'
Returns
-------
terr : error probability
tpfa : false alarm probability
tpd : detection probability
Notes
-----
Iterate on energy from 0 to 10*E
"""
thresh = np.linspace(10*E,0,300)
terr = []
tpfa = []
tpd =[]
for t in thresh:
perr = self.errprob(E,thresh=t,typ=typ)
terr.append(perr)
tpfa.append(self.p10)
tpd.append(1-self.p01)
terr = np.array(terr)
tpfa = np.array(tpfa)
tpd = np.array(tpd)
return(terr,tpfa,tpd)
t = linspace(0,1,Npoints)
wn = sqrt(Var_w)*randn(Npoints)
#
# Filter 1 : Input Band Pass Filter
#
w1 = 0.05
w2 = 0.1
wt = 0.01
wp = [w1,w2]
ws = [w1-wt,w2+wt]
gpass = 0.5
gstop = 30
type = 'ellip'
filter1 = DF()
filter1.ellip_bp(wp,ws,gpass,gstop)
#
# Filter 2 : averaging filter
#
N = 40
b = (1./N)*ones(N)
a = array([1])
filter2 = DF(b,a)
#
#
#
beta = 1
pfa = 0.001