/
controllerSRV.py
258 lines (211 loc) · 11.8 KB
/
controllerSRV.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
# =================================================================================================================================================
# Import modules
import sys
sys.path.append('/home/ercelik/opt1/nest/lib/python3.4/site-packages/')
import nest
import pickle
import random
import numpy as np
def PickleIt(data,fileName):
with open(fileName, 'wb') as f:
pickle.dump(data, f, pickle.HIGHEST_PROTOCOL)
def GetPickle(fileName):
with open(fileName, 'rb') as f:
data = pickle.load(f)
return data
def RandomConnect(Source,Dest,numCon,Weight,Delay,Model):
#SourceNeurons=random.sample(Source,numCon)
for i in range(len(Source)):
DestNeurons=random.sample(Dest,numCon)
nest.Connect(Source[i:i+1],DestNeurons,{'rule':'all_to_all'},{'weight':Weight,'model':Model,'delay':Delay})
def PoissonRate(poi,rate):
nest.SetStatus(poi,"rate",rate+0.0)
def SpikeNum(SpikeGen,Stime,Ftime):
timeArray=nest.GetStatus(SpikeGen,"events")[0]['times']
betweenTimeInt=[j for j in range(len(timeArray)) \
if (timeArray[j]>=Stime and timeArray[j]<=Ftime)]
return len(betweenTimeInt)+.0
def SetNeuronInput(Neuron,Sti):
nest.SetStatus(Neuron,"I_e",Sti)
def CalcPoiRate(coeff,offset,maxRate,minRate,value):
return np.exp(-coeff*(value+offset)**2)*(maxRate-minRate)+minRate
def ActFunc(x):
return -0.135*x*np.exp(0.05*x)#-0.675*x*np.exp(0.25*x)#
def InputFunc(n,maxim,minim,value,off):
return np.exp((-n/(3*(maxim-minim)))*(value-off)**2)
def InputAcc(value):
global nPoi,minPoiRate,maxPoiRate,maxAcc,minAcc
n=nPoi
step=(maxAcc-minAcc)/n
coeff=1/(step*4.)
offset=np.arange(minAcc,maxAcc,step)+step/2.
Poi=[CalcPoiRate(coeff,offset[i],maxPoiRate,minPoiRate,value) for i in range(n)]
return Poi
bpy.context.scene.game_settings.fps=50.
dt=1000./bpy.context.scene.game_settings.fps
#nest.sli_func('synapsedict info')
# =================================================================================================================================================
# Creating muscles
#~ muscle_ids = {}
#~ [ muscle_ids["forearm.L_FLEX"], muscle_ids["forearm.L_EXT"] ] = setTorqueMusclePair(reference_object_name = "obj_forearm.L", attached_object_name = "obj_upper_arm.L", maxF = 5.0)
#~ servo_ids = {}
#~ servo_ids["forearm.L"] = setVelocityServo(reference_object_name = "obj_forearm.L", attached_object_name = "obj_upper_arm.L", maxV = 10.0)
PP=60.
servo_ids = {}
servo_ids["wrist.L"] = setPositionServo(reference_object_name = "obj_wrist.L", attached_object_name = "obj_forearm.L", P = PP)
servo_ids["wrist.R"] = setPositionServo(reference_object_name = "obj_wrist.R", attached_object_name = "obj_forearm.R", P = PP)
servo_ids["forearm.L"] = setPositionServo(reference_object_name = "obj_forearm.L", attached_object_name = "obj_upper_arm.L", P = PP)
servo_ids["forearm.R"] = setPositionServo(reference_object_name = "obj_forearm.R", attached_object_name = "obj_upper_arm.R", P = PP)
servo_ids["upper_arm.L"] = setPositionServo(reference_object_name = "obj_upper_arm.L", attached_object_name = "obj_shoulder.L", P = PP)
servo_ids["upper_arm.R"] = setPositionServo(reference_object_name = "obj_upper_arm.R", attached_object_name = "obj_shoulder.R", P = PP)
servo_ids["shin_lower.L"] = setPositionServo(reference_object_name = "obj_shin_lower.L", attached_object_name = "obj_shin.L", P = PP)
servo_ids["shin_lower.R"] = setPositionServo(reference_object_name = "obj_shin_lower.R", attached_object_name = "obj_shin.R", P = PP)
servo_ids["shin.L"] = setPositionServo(reference_object_name = "obj_shin.L", attached_object_name = "obj_thigh.L", P = PP)
servo_ids["shin.R"] = setPositionServo(reference_object_name = "obj_shin.R", attached_object_name = "obj_thigh.R", P = PP)
servo_ids["thigh.L"] = setPositionServo(reference_object_name = "obj_thigh.L", attached_object_name = "obj_hips", P = PP)
servo_ids["thigh.R"] = setPositionServo(reference_object_name = "obj_thigh.R", attached_object_name = "obj_hips", P = PP)
# =================================================================================================================================================
# Network creation
## Nest Kernel Initialization
T=8
nest.ResetKernel()
nest.SetKernelStatus({"overwrite_files": True, "print_time": True})
nest.SetKernelStatus({"local_num_threads": T})
aa=np.random.randint(1,100)
nest.SetKernelStatus({'rng_seeds' : range(aa,aa+T)})
nest.sr("M_ERROR setverbosity")
## Network Parameters
neuronModel='aeif_cond_exp'
numSC=100
dx=10
dy=5
dz=2
dim=3
inpDim=6
connProb=.1
synType='tsodyks2_synapse'
numInp=(numSC*1.)
numOut=12
Weight=10.
## Input Parameters
maxPoiRate=1000.
minPoiRate=10.
nPoi=20
maxAcc=30.
minAcc=-30.
## Network Definitions and Connections
conn=[(i-1,i+1,i-dx,i+dx,i-dx*dy,i+dx*dy) for i in range(numSC)] # nor recursive conn
SpinCord=nest.Create(neuronModel,numSC)
for i in range(numSC):
for j in range(dim*2):
if np.random.rand()<connProb and conn[i][j]>=0 and conn[i][j]<numSC:
nest.Connect(i+1,conn[i][j]+1,{'weight':Weight,'delay':1.,'model':synType})
## IDs of Input and Output Neurons
InpNeurons=tuple(np.random.randint(np.amin(SpinCord),np.amax(SpinCord)+1,size=(1,numInp))[0])
OutNeurons=tuple(np.random.randint(np.amin(SpinCord),np.amax(SpinCord)+1,size=(1,numOut))[0])
## Recording Devices
outSpikes=nest.Create("spike_detector", 1, {"to_file": False})
nest.ConvergentConnect(OutNeurons, outSpikes, model="static_synapse")
## Input Stimulus
poisson = nest.Create( 'poisson_generator' , nPoi*inpDim , { 'rate' : minPoiRate }) # for Inp1
nest.Connect(poisson,InpNeurons[:],{'rule':'all_to_all'},{'weight':40.,'model':'static_synapse','delay':1.})
## Output Weights
wout=np.random.rand(numOut,numOut)
## Reinforcement Learning Param
wV=np.random.rand(numOut,1)
vOld=.1
count=0
eps=1.
ax=.1
ay=.1
az=.1
ax2=.1
ay2=.1
az2=.1
ax3=.1
ay3=.1
az3=.1
# =================================================================================================================================================
# Evolve function
def evolve():
#~ print("Step:", i_bl, " Time:", t_bl)
# ------------------------------------- Visual ------------------------------------------------------------------------------------------------
#visual_array = getVisual(camera_name = "Meye", max_dimensions = [256,256])
#scipy.misc.imsave("test_"+('%05d' % (i_bl+1))+".png", visual_array)
# ------------------------------------- Olfactory ---------------------------------------------------------------------------------------------
olfactory_array = getOlfactory(olfactory_object_name = "obj_nose", receptor_names = ["smell1", "plastic1"])
# ------------------------------------- Taste -------------------------------------------------------------------------------------------------
taste_array = getTaste( taste_object_name = "obj_mouth", receptor_names = ["smell1", "plastic1"], distance_to_object = 1.0)
# ------------------------------------- Vestibular --------------------------------------------------------------------------------------------
vestibular_array = getVestibular(vestibular_object_name = "obj_head")
#print (vestibular_array)
# ------------------------------------- Sensory -----------------------------------------------------------------------------------------------
# ------------------------------------- Proprioception ----------------------------------------------------------------------------------------
#~ spindle_FLEX = getMuscleSpindle(control_id = muscle_ids["forearm.L_FLEX"])
#~ spindle_EXT = getMuscleSpindle(control_id = muscle_ids["forearm.L_EXT"])
# ------------------------------------- Neural Simulation -------------------------------------------------------------------------------------
# nest.Simulate()
global dt,numOut,wout,wV,vOld,count,eps,ax,ay,az,ax2,ay2,az2,ax3,ay3,az3
#nest.SetStatus(poisson,InputAcc(vestibular_array[3:6]))
rates=np.array(InputAcc(np.hstack((vestibular_array[0:3],[ax,ay,az]))))
rates=list(rates.reshape(1,rates.shape[0]*rates.shape[1])[0])
nest.SetStatus(poisson,'rate',rates)
nest.Simulate(dt)
## Get Events
T_times=nest.GetStatus(outSpikes,'events')[0]['times']
T_senders=nest.GetStatus(outSpikes,'events')[0]['senders']
## Time Shifting and Activity Calc
time=nest.GetKernelStatus()['time']
T_times-=time
T_act=ActFunc(T_times)
## Output to Limbs
outV=[[T_act[j] for j in range(T_senders.size) if OutNeurons[i]==T_senders[j]]for i in range(numOut)]
outV=np.array([sum(outV[i]) for i in range(numOut)])
outV=outV/(numOut+.0)
## tempOutV=outV
## vN=wV.T.dot(outV)/numOut
##
## outV=wout.dot(outV)
outV=np.array([.5*np.random.rand() for i in range(numOut)])
outV=0.5 + 0.5*np.sin(outV*t_bl)
#outV=wout.dot(outV)/numOut
## outV=np.array([np.average(outV[i]) if len(outV[i]) is not 0 else 0. for i in range(numOut)])
#print(outV)
## Delete previous spikes
nest.SetStatus(outSpikes, 'n_events', 0)
# ------------------------------------- Muscle Activation -------------------------------------------------------------------------------------
#outV=outV/max(outV)
#print(outV)
controlActivity(control_id = servo_ids["wrist.L"], control_activity = outV[0])
controlActivity(control_id = servo_ids["wrist.R"], control_activity = outV[1])
controlActivity(control_id = servo_ids["forearm.L"], control_activity = outV[2])
controlActivity(control_id = servo_ids["forearm.R"], control_activity = outV[3])
controlActivity(control_id = servo_ids["upper_arm.L"], control_activity = outV[4])
controlActivity(control_id = servo_ids["upper_arm.R"], control_activity = outV[5])
controlActivity(control_id = servo_ids["shin_lower.L"], control_activity = outV[6])
controlActivity(control_id = servo_ids["shin_lower.R"], control_activity = outV[7])
controlActivity(control_id = servo_ids["shin.L"], control_activity = outV[8])
controlActivity(control_id = servo_ids["shin.R"], control_activity = outV[9])
controlActivity(control_id = servo_ids["thigh.L"], control_activity = outV[10])
controlActivity(control_id = servo_ids["thigh.R"], control_activity = outV[11])
## speed_ = 6.0
## act_tmp = 0.5 + 0.5*np.sin(speed_*t_bl)
## anti_act_tmp = 1.0 - act_tmp
## act_tmp_p1 = 0.5 + 0.5*np.sin(speed_*t_bl - np.pi*0.5)
## anti_act_tmp_p1 = 1.0 - act_tmp_p1
## act_tmp_p2 = 0.5 + 0.5*np.sin(speed_*t_bl + np.pi*0.5)
## anti_act_tmp_p2 = 1.0 - act_tmp_p2
##
##
## controlActivity(control_id = servo_ids["wrist.L"], control_activity = 0.4)
## controlActivity(control_id = servo_ids["wrist.R"], control_activity = 0.4)
## controlActivity(control_id = servo_ids["forearm.L"], control_activity = 0.8*act_tmp)
## controlActivity(control_id = servo_ids["forearm.R"], control_activity = 0.8*anti_act_tmp)
## controlActivity(control_id = servo_ids["upper_arm.L"], control_activity = 1.0*act_tmp_p1)
## controlActivity(control_id = servo_ids["upper_arm.R"], control_activity = 1.0*anti_act_tmp_p1)
## controlActivity(control_id = servo_ids["shin_lower.L"], control_activity = 0.8*anti_act_tmp)
## controlActivity(control_id = servo_ids["shin_lower.R"], control_activity = 0.8*act_tmp)
## controlActivity(control_id = servo_ids["shin.L"], control_activity = 0.5*anti_act_tmp_p1)
## controlActivity(control_id = servo_ids["shin.R"], control_activity = 0.5*act_tmp_p1)
## controlActivity(control_id = servo_ids["thigh.L"], control_activity = 0.5*anti_act_tmp)
## controlActivity(control_id = servo_ids["thigh.R"], control_activity = 0.5*act_tmp)