def test_value_dump_load():
conf = configs.SparseServiceConfig()
assert conf.downsampling_factor == 1
conf.downsampling_factor = 2
dump = conf._dumps()
conf = configs.SparseServiceConfig()
assert conf.downsampling_factor == 1
conf._loads(dump)
assert conf.downsampling_factor == 2
conf = configs.EnvelopeServiceConfig()
with pytest.raises(AssertionError):
conf._loads(dump)
def get_sensor_config():
config = configs.SparseServiceConfig()
config.profile = configs.SparseServiceConfig.Profile.PROFILE_3
config.range_interval = [0.48, 0.72]
config.sweeps_per_frame = 16
config.hw_accelerated_average_samples = 60
config.update_rate = 60
return config
def get_sensor_config():
sensor_config = configs.SparseServiceConfig()
sensor_config.range_interval = [0.24, 1.20]
sensor_config.update_rate = 60
sensor_config.sampling_mode = configs.SparseServiceConfig.SamplingMode.A
sensor_config.profile = configs.SparseServiceConfig.Profile.PROFILE_3
sensor_config.hw_accelerated_average_samples = 60
return sensor_config
def test_value_get_set():
conf = configs.SparseServiceConfig()
assert conf.downsampling_factor == 1
conf.downsampling_factor = 2
assert conf.downsampling_factor == 2
del conf.downsampling_factor
assert conf.downsampling_factor == 1
def get_sensor_config():
config = configs.SparseServiceConfig()
config.profile = configs.SparseServiceConfig.Profile.PROFILE_3
config.sampling_mode = configs.SparseServiceConfig.SamplingMode.B
config.range_interval = [0.3, 1.3]
config.update_rate = 80
config.sweeps_per_frame = 32
config.hw_accelerated_average_samples = 60
return config
def get_sensor_config():
config = configs.SparseServiceConfig()
config.profile = configs.SparseServiceConfig.Profile.PROFILE_3
config.sampling_mode = configs.SparseServiceConfig.SamplingMode.A
config.range_interval = [0.24, 0.48]
config.sweeps_per_frame = 64
config.sweep_rate = 3e3
config.hw_accelerated_average_samples = 60
return config
def get_sensor_config():
config = configs.SparseServiceConfig()
config.profile = configs.SparseServiceConfig.Profile.PROFILE_4
config.sampling_mode = configs.SparseServiceConfig.SamplingMode.A
config.range_interval = [0.36, 0.54]
config.downsampling_factor = 3
config.sweeps_per_frame = 512
config.hw_accelerated_average_samples = 60
return config
def main():
args = utils.ExampleArgumentParser().parse_args()
utils.config_logging(args)
if args.socket_addr:
client = clients.SocketClient(args.socket_addr)
elif args.spi:
client = clients.SPIClient()
else:
port = args.serial_port or utils.autodetect_serial_port()
client = clients.UARTClient(port)
client.squeeze = False
sensor_config = configs.SparseServiceConfig()
sensor_config.sensor = args.sensors
sensor_config.range_interval = [0.24, 1.20]
sensor_config.sweeps_per_frame = 16
sensor_config.hw_accelerated_average_samples = 60
sensor_config.sampling_mode = sensor_config.SamplingMode.A
sensor_config.profile = sensor_config.Profile.PROFILE_3
sensor_config.gain = 0.6
session_info = client.setup_session(sensor_config)
pg_updater = PGUpdater(sensor_config, None, session_info)
pg_process = PGProcess(pg_updater)
pg_process.start()
client.start_session()
interrupt_handler = utils.ExampleInterruptHandler()
print("Press Ctrl-C to end session")
while not interrupt_handler.got_signal:
data_info, data = client.get_next()
try:
pg_process.put_data(data)
except PGProccessDiedException:
break
print("Disconnecting...")
pg_process.close()
client.disconnect()
def __init__(self, demo_ctrl, params):
super().__init__(demo_ctrl, self.detector_name)
self.config = configs.SparseServiceConfig()
self.config.sampling_mode = configs.SparseServiceConfig.SamplingMode.A
self.update_config(params)
def main():
args = utils.ExampleArgumentParser().parse_args()
utils.config_logging(args)
if args.socket_addr:
client = clients.SocketClient(args.socket_addr)
elif args.spi:
client = clients.SPIClient()
else:
port = args.serial_port or utils.autodetect_serial_port()
client = clients.UARTClient(port)
client.squeeze = False
range_start = 0.18
range_end = 0.60
num = int((range_end*100-range_start*100)/6)+1
sensor_config = configs.SparseServiceConfig()
sensor_config.sensor = args.sensors
sensor_config.range_interval = [range_start, range_end]
sensor_config.sweeps_per_frame = 16
sensor_config.hw_accelerated_average_samples = 60
sensor_config.sampling_mode = sensor_config.SamplingMode.A
sensor_config.profile = sensor_config.Profile.PROFILE_2
sensor_config.gain = 0.6
session_info = client.setup_session(sensor_config)
# pg_updater = PGUpdater(sensor_config, None, session_info)
# pg_process = PGProcess(pg_updater)
# pg_process.start()
client.start_session()
storage = []
counter = 0
sample = 300
interrupt_handler = utils.ExampleInterruptHandler()
print("Press Ctrl-C to end session")
temp = np.zeros(num)
#端末からデータを受け取り、フレームに入っているsweepの平均を取得し、tempに追加
while not interrupt_handler.got_signal:
data_info, data = client.get_next()
counter += 1
for sweep in data[0]:
temp = temp + sweep
temp = temp/len(data[0])
storage.append(temp)
temp = np.zeros(int(num))
if(counter >= sample):
break
#300個のフレームから距離ごとに平均を取得
# result = np.zeros(len(storage[0]))
# for data in storage:
# result = result + data
# result = result/len(storage)
# clr=plt.rcParams['axes.prop_cycle'].by_key()['color']
roop_count = 0
prev_getData = np.loadtxt('sparse.csv')
#生データ300個のフレームの平均を表示
show_raw_data = np.zeros(int(num))
for frame in storage:
show_raw_data += frame
show_curr_RawData = show_raw_data/sample
##保存していたデータの生データ300個のフレームの平均
show_prev_RawData = []
for data in prev_getData:
show_prev_RawData.append(np.sum(data)/sample)
x = np.arange(range_start*100,range_end*100+1,6)
plt.subplot(1,3,1)
plt.plot(x,show_curr_RawData,color='r')
plt.title("Raw Data(Current)")
plt.xlabel("Distance(cm)")
plt.ylabel("Amptitude")
plt.subplot(1,3,2)
plt.plot(x,show_prev_RawData,color='b')
plt.title("Raw Data(Previous)")
plt.xlabel("Distance(cm)")
plt.ylabel("Amptitude")
plt.subplot(1,3,3)
plt.title("Raw Data(Compare)")
plt.plot(x,show_curr_RawData,color='r')
plt.plot(x,show_prev_RawData,color='b')
plt.xlabel("Distance")
plt.ylabel("Amptitude")
# plt.tight_layout()
# plt.show()
# exit(1)
#データを別々に表示
for i in range(num):
show_data = []
plt.subplot(math.ceil(num/2),4,roop_count*2+1)
plt.title(str(range_start*100+i*6)+"cm(Current)")
plt.xlabel("Quantity")
plt.ylabel("Amptitude")
for j in range(len(storage)):
show_data.append(storage[j][i])
plt.plot(range(1,sample+1),show_data,color='r')
plt.subplot(math.ceil(num/2),4,roop_count*2+2)
plt.title(str(range_start*100+i*6)+"cm(Previous)")
plt.plot(range(1,sample+1),prev_getData[i],color='b')
plt.xlabel("Quantity")
plt.ylabel("Amptitude")
roop_count += 1
# plt.tight_layout()
# plt.show()
#データをあわせて表示
for i in range(num):
show_data = []
plt.subplot(math.ceil(num/2),2,i+1)
for j in range(len(storage)):
show_data.append(storage[j][i])
plt.plot(range(1,sample+1),show_data,color='r',label="Current Data")
plt.title(str(range_start*100+i*6)+"cm(Combination)")
plt.plot(range(1,sample+1),prev_getData[i],color='b',label="Previous Data")
plt.xlabel("Quantity")
plt.ylabel("Amptitude")
plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left', borderaxespad=0,fontsize='small')
# plt.tight_layout()
# plt.show()
#ある距離の時系列データの遷移とそのデータのヒストグラムの表示
now_getData = []
for i in range(num):
show_data = []
# plt.subplot(math.ceil(num/2),2,i+1)
plt.subplot(math.ceil(num/2),4,i*2+1)
plt.title(str(range_start*100+i*6)+"cm")
plt.xlabel("Distance")
plt.ylabel("Amptitude")
for j in range(len(storage)):
show_data.append(storage[j][i])
now_getData.append(show_data)
# ヒストグラムの作成と表示
# hist, bins = np.histogram(show_data,density=True)
plt.plot(range(1,sample+1),show_data,color = "tomato")
plt.subplot(math.ceil(num/2),4,i*2+2)
plt.hist(show_data,bins=10,density=True,color = "aqua")
plt.title(str(range_start*100+i*6)+"cm")
plt.xlabel("Class")
plt.ylabel("Frequency")
# print("ヒストグラムの度数"+str(hist[0]))
# print("階級を区切る値"+str(hist[1]))
# plt.tight_layout()
# plt.show()
#ndarray型に変換し、保存
# now_getData = np.array(now_getData)
# np.savetxt('sparse.csv',now_getData)
# exit(1)
roop_count = 0
prev_getData = np.loadtxt('sparse.csv')
KL_storage = []
JS_storage = []
#ヒストグラムの度数からKLを算出
for i in zip(prev_getData,now_getData):
plt.subplot(math.ceil(num/2),4,roop_count*2+1)
plt.title(str(range_start*100+roop_count*6)+"cm(Previous)")
plt.xlabel("Distance")
plt.ylabel("Frequency")
now_hist = plt.hist(i[0],bins=10,color = "lime")
now_hist = normalization(np.array(now_hist[0]))
# now_hist_normalization = normalization(np.array(now_hist[0]))
# now_hist[0] = now_hist_normalization
# print(now_hist)
plt.subplot(math.ceil(num/2),4,roop_count*2+2)
plt.title(str(range_start*100+roop_count*6)+"cm(Now)")
plt.xlabel("Distance")
plt.ylabel("Frequency")
prev_hist = plt.hist(i[1],bins=10,color = "deepskyblue")
prev_hist = normalization(np.array(prev_hist[0]))
# print("now_histの要素の数: "+str(len(now_hist[0])))
# print("prev_histの要素の数: "+str(len(now_hist[0])))
KL_value = KLdivergence(now_hist,prev_hist)
print(str(range_start*100+roop_count*6)+"cm時のKL_Divergence: "+str(KL_value))
JS_value = JSdivergence(now_hist,prev_hist)
print(str(range_start*100+roop_count*6)+"cm時のJS_Divergence: "+str(JS_value)+"\n")
KL_storage.append(KL_value)
JS_storage.append(JS_value)
roop_count += 1
# plt.tight_layout()
# plt.show()
temp_KLlist = []
temp_store_KLlist = []
temp_JSlist = []
temp_store_JSlist = []
temp_KLlist.append(np.array(KL_storage))
temp_JSlist.append(np.array(JS_storage))
#ファイルからデータを読みこむ
store_KLarr = np.loadtxt("KLDivergence_Sparse.csv",delimiter = ",")
store_JSarr = np.loadtxt("JSDivergence_Sparse.csv",delimiter = ",")
temp_store_KLlist.append(store_KLarr)
temp_store_JSlist.append(store_JSarr)
#データを追加
store_KLarr = np.empty((0,num),int)
store_JSarr = np.empty((0,num),int)
for data in temp_store_KLlist:
print(data)
store_KLarr = np.append(store_KLarr,np.array(data),axis=0)
# store_KLarr = np.append(store_KLarr,np.array(temp_store_KLlist),axis=0)
for data in temp_store_KLlist:
store_JSarr = np.append(store_JSarr,np.array(data),axis=0)
# store_JSarr = np.append(store_JSarr,np.array(temp_store_JSlist),axis=0)
store_KL = np.append(store_KLarr,np.array(temp_KLlist),axis=0)
store_JS = np.append(store_JSarr,np.array(temp_JSlist),axis=0)
#ファイルを保存
store_KL = np.savetxt("KLDivergence_Sparse.csv",store_KL,delimiter=",")
store_JS = np.savetxt("JSDivergence_Sparse.csv",store_JS,delimiter=",")
#KL,JSの平均値
# print("KL_Divergenceの平均値: "+str(np.sum(np.array(KL_storage))/num)+"\n")
# print("JS_Divergenceの平均値: "+str(np.sum(np.array(JS_storage))/num)+"\n")
#KLとJSを可視化
X = list(range(int(range_start*100),int(range_end*100+1),6))
plt.subplot(1,3,1)
# plt.plot(range(len(KL_storage)),KL_storage,color='r',marker="o")
plt.plot(X,KL_storage,color='r',marker="o")
plt.title("KL Divergence")
plt.xlabel("Distance")
plt.ylabel("Amptitude")
plt.subplot(1,3,2)
plt.plot(X,JS_storage,color='b',marker="o")
plt.title("JS Divergence")
plt.xlabel("Distance")
plt.ylabel("Amptitude")
plt.subplot(1,3,3)
plt.plot(X,KL_storage,color='r',marker="o")
plt.plot(X,JS_storage,color='b',marker="o")
# plt.plot(X,KL_strage,color='r',marker="o",linewidth=0)
# plt.plot(X,JS_strage,color='b',marker="o",linewidth=0)
plt.title("Compare")
plt.xlabel("Distance")
plt.ylabel("Amptitude")
# plt.tight_layout()
# plt.show()
print("Disconnecting...")
client.disconnect()
def test_repetition_mode(setup):
client, sensor = setup
def measure(config):
client.start_session(config, check_config=False)
client.get_next()
t0 = time.time()
missed = False
n = 5
for _ in range(n):
info, data = client.get_next()
if info["missed_data"]:
missed = True
t1 = time.time()
client.stop_session()
dt = (t1 - t0) / n
return (dt, missed)
config = configs.SparseServiceConfig()
config.sensor = sensor
config.range_interval = [0.3, 0.36]
config.sweeps_per_frame = 50
config.sweep_rate = 1e3
nominal_f = config.sweep_rate / config.sweeps_per_frame
nominal_dt = 1.0 / nominal_f
# on demand / host driven
config.repetition_mode = configs.SparseServiceConfig.RepetitionMode.HOST_DRIVEN
# no rate limit
config.update_rate = None
dt, missed = measure(config)
assert not missed
assert dt == pytest.approx(nominal_dt, rel=0.15)
# ok rate limit
config.update_rate = 0.5 / nominal_dt
dt, missed = measure(config)
assert not missed
assert dt == pytest.approx(nominal_dt * 2.0, rel=0.15)
# too high rate limit
config.update_rate = 2.0 / nominal_dt
dt, missed = measure(config)
if isinstance(client, clients.SocketClient): # TODO
assert missed
assert dt == pytest.approx(nominal_dt, rel=0.15)
# streaming / sensor driven
config.repetition_mode = configs.SparseServiceConfig.RepetitionMode.SENSOR_DRIVEN
# ok rate
config.update_rate = 0.5 / nominal_dt
dt, missed = measure(config)
assert not missed
assert dt == pytest.approx(nominal_dt * 2.0, rel=0.15)
# too high rate
config.update_rate = 2.0 / nominal_dt
dt, missed = measure(config)
assert missed
def test_mp():
conf = configs.SparseServiceConfig()
conf.downsampling_factor = 2
[dump] = mp.Pool(1).map(dump_fun, [conf])
assert dump == conf._dumps()