class Kalman_SNN:
def __init__(self):
self.A_k = None
self.B_k = None
self.dt = 0.02 # simulation time step
self.t_rc = 0.005 # membrane RC time constant
self.t_ref = 0.001 # refractory period
self.tau = 0.014 # synapse time constant for standard first-order lowpass filter synapse
self.N_A = 1000 # number of neurons in first population
self.rate_A = 200, 400 # range of maximum firing rates for population A
self.model = nengo.Network(label="Kalman SNN")
self.sim = None
self.kalman = Kalman()
self.output = None
self.testX = None
self.count = 0
self.chN = None
def train(self, NeuralData, NeuralDim, KinData, KinDim):
NeuralData = np.reshape(np.asarray(NeuralData), (NeuralDim, -1))
KinData = np.reshape(np.asarray(KinData), (KinDim, -1))
self.chN = np.where(np.sum(NeuralData, axis=1) != 0)
NeuralData = np.squeeze(NeuralData[self.chN, :])
self.kalman.calculate(NeuralData,
KinData,
pool=0,
dt=self.dt,
tau=self.tau)
def build(self, testY):
self.count = 0
def update(x):
"""
Kalman Filter: X_k = A * X_k_1 + B * Y_k
"""
Externalmat = np.mat(x[2:4]).T
Inputmat = np.mat(x[0:2]).T
Controlmat = np.matrix([[x[4], x[5]], [x[6], x[7]]])
next_state = np.squeeze(
np.asarray(Controlmat * Inputmat + Externalmat))
return next_state
with self.model:
Dir_Nurons = nengo.Ensemble(1,
dimensions=2 + 2 + 4,
neuron_type=nengo.Direct())
LIF_Neurons = nengo.Ensemble(
self.N_A,
dimensions=2,
intercepts=Uniform(-1, 1),
max_rates=Uniform(self.rate_A[0], self.rate_A[1]),
neuron_type=nengo.LIFRate(tau_rc=self.t_rc,
tau_ref=self.t_ref))
state_func = Piecewise({
0.0: [0.0, 0.0],
self.dt:
np.squeeze(np.asarray(np.mat([testY[0], testY[1]]).T)),
2 * self.dt: [0.0, 0.0]
})
state = nengo.Node(output=state_func)
# state_probe = nengo.Probe(state)
external_input = nengo.Node(output=lambda t: self.data(t))
# external_input_probe = nengo.Probe(external_input)
control_signal = nengo.Node(output=lambda t: self.control(t))
conn0 = nengo.Connection(state, Dir_Nurons[0:2])
#
conn1 = nengo.Connection(external_input, Dir_Nurons[2:4])
conn2 = nengo.Connection(control_signal, Dir_Nurons[4:8])
conn3 = nengo.Connection(Dir_Nurons,
LIF_Neurons[0:2],
function=update,
synapse=self.tau)
conn4 = nengo.Connection(LIF_Neurons[0:2], Dir_Nurons[0:2])
self.output = nengo.Probe(LIF_Neurons[0:2])
self.sim = nengo.Simulator(self.model, dt=self.dt)
def data(self, t):
if t == 0.0:
return [0.0, 0.0]
yt = np.mat(self.testX)
out = np.transpose(self.B_k * yt.T)
return np.squeeze(np.asarray(out))
def control(self, t):
"""
Matrix A_0 = (I-KC)A, K: Kalman Gain, A,C: State and observation matrix
"""
if t == 0.0:
return [0.0, 0.0, 0.0, 0.0]
return np.squeeze(np.asarray(self.A_k.ravel()))
def test(self, testX):
self.A_k, self.B_k = self.kalman.K_update(dt=self.dt, tau=self.tau)
testX = np.asarray(testX)
self.testX = testX[self.chN]
self.sim.step()
res = self.sim.data[self.output][self.count]
self.count = self.count + 1
res = array.array('d', res)
return res
def save(self, name='data.npy'):
path = './' + name
self.kalman.save(path)
def load(self, name='data.npy'):
path = './' + name
self.kalman.load(path)
def getParam(self):
return self.kalman.getParam()
def standard_kalman(self, testX, testY, length=None):
self.kalman.standard_Kalman_Filter(testX, testY, length)
dataloader = Dataloader()
kalman = Kalman()
trainX, trainY, testX, testY = dataloader.getData()
ChN = np.where(np.sum(trainX, axis=1) != 0)
trainX = np.squeeze(trainX[ChN, :])
testX = np.squeeze(testX[ChN, :])
kalman.calculate(trainX, trainY, pool=pool, dt=dt, tau=tau)
# kalman.Kalman_Filter(testX, testY)
kalman.standard_Kalman_Filter(testX, testY)
global A_0, B_0
A_0, B_0 = kalman.K_update(dt, tau)
def data(t):
"""
Neuron records, Y_k, and calculate B * Y_k
"""
if t == 0.0:
return [0.0, 0.0]
yt = np.mat(testX[:, int(50 * t)])
out = np.transpose(B_0 * yt.T)
return np.squeeze(np.asarray(out))
def control(t):
