def makePrediction(dTemplate, toTest, avgDev):
disregard, testDev = dtw.getDTW(dTemplate, toTest)
avgSlope = linRegress.getSlopeForDifference(
[i for i in range(len(avgDev))], avgDev)
testSlope = linRegress.getSlopeForDifference(
[i for i in range(len(testDev))], testDev)
return 1 if testSlope <= avgSlope * THRESHOLD_PERCENT else 0
def authenticate(deviceID, questionTemplateFile, thresholdPercent): deviceTemplate = getTemplateFromFile(getFilePath(deviceID, DEV_TEMP_FOR_HALF_PATH)) questionTemplate = getTemplateFromFile(questionTemplateFile) avgDevelopmentScore = getTemplateFromFile(getFilePath(deviceID, DEV_DEV_FOR_HALF_PATH)) dtwScore, seqDev = dtw.getDTW(deviceTemplate, questionTemplate) avgDev = getLineForDev(avgDevelopmentScore) seqDev = getLineForDev(seqDev) avgSlope = linRegress.getSlopeForDifference([i for i in range(len(avgDev))], avgDev) seqSlope = linRegress.getSlopeForDifference([i for i in range(len(seqDev))], seqDev) return 1 if seqSlope <= avgSlope*thresholdPercent else 0
def authenticate(sID, dID): deviceTemplate = getTemplate(getFileNamePath(dID, DEV_TEMP_PATH, "d_")) sequenceTemplate = getTemplate(getFileNamePath(sID, SEQ_TEMP_PATH, "s_")) avgDevelopmentScore = getTemplate(getFileNamePath(dID, DEV_DEV_PATH, "d_")) dtwScore, seqDev = dtw.getDTW(deviceTemplate, sequenceTemplate) avgDev = getLineForDev(avgDevelopmentScore) seqDev = getLineForDev(seqDev) # plt.plot([i for i in range(len(deviceTemplate))], deviceTemplate, color='blue') # plt.plot([i for i in range(len(sequenceTemplate))], sequenceTemplate, color='red') # plt.plot([i for i in range(len(avgDev))], avgDev, color='blue') # plt.plot([i for i in range(len(seqDev))], seqDev, color='red') # plt.show() avgSlope = linRegress.getSlopeForDifference([i for i in range(len(avgDev))], avgDev) seqSlope = linRegress.getSlopeForDifference([i for i in range(len(seqDev))], seqDev) return 1 if seqSlope <= avgSlope*THRESHOLD_PERCENT else 0
def authenticate(sID, dID):
deviceTemplate = getTemplate(getFileNamePath(dID, DEV_TEMP_PATH, "d_"))
sequenceTemplate = getTemplate(getFileNamePath(sID, SEQ_TEMP_PATH, "s_"))
avgDevelopmentScore = getTemplate(getFileNamePath(dID, DEV_DEV_PATH, "d_"))
dtwScore, seqDev = dtw.getDTW(deviceTemplate, sequenceTemplate)
avgDev = getLineForDev(avgDevelopmentScore)
seqDev = getLineForDev(seqDev)
# plt.plot([i for i in range(len(deviceTemplate))], deviceTemplate, color='blue')
# plt.plot([i for i in range(len(sequenceTemplate))], sequenceTemplate, color='red')
# plt.plot([i for i in range(len(avgDev))], avgDev, color='blue')
# plt.plot([i for i in range(len(seqDev))], seqDev, color='red')
# plt.show()
avgSlope = linRegress.getSlopeForDifference(
[i for i in range(len(avgDev))], avgDev)
seqSlope = linRegress.getSlopeForDifference(
[i for i in range(len(seqDev))], seqDev)
return 1 if seqSlope <= avgSlope * THRESHOLD_PERCENT else 0
test_data_set_az, log_test_data_set_az[count][0])
test_data_set_gx = remake_test_data_set(
test_data_set_gx, log_test_data_set_gx[count][0])
test_data_set_gy = remake_test_data_set(
test_data_set_gy, log_test_data_set_gy[count][0])
test_data_set_gz = remake_test_data_set(
test_data_set_gz, log_test_data_set_gz[count][0])
if 0.75 < log_test_data_set_az[
count - 50][0] and log_test_data_set_ay[count][0] < -0.05:
print(
"=============================================pick and drop and hand down==============================================="
)
pick_dtw_gz_result = dtw.getDTW(train_data_set_gz,
test_data_set_gz)
drop_dtw_gz_result = dtw.getDTW(drop_train_data_set_gz,
test_data_set_gz)
hand_down_dtw_az_result = dtw.getDTW(hand_down_data_set_az,
test_data_set_az)
if operation_identification(
pick_dtw_gz_result - drop_dtw_gz_result
) == "pick" or operation_identification(
pick_dtw_gz_result - drop_dtw_gz_result) == "drop":
test_data_set_ax = np.zeros_like(test_data_set_ax)
test_data_set_ay = np.zeros_like(test_data_set_ay)
test_data_set_az = np.zeros_like(test_data_set_az)
test_data_set_gx = np.zeros_like(test_data_set_gx)
test_data_set_gy = np.zeros_like(test_data_set_gy)
test_data_set_gz = np.zeros_like(test_data_set_gz)
if operation_identification(pick_dtw_gz_result -
# 加速度を取得
accel_x, accel_y, accel_z = get_accel_data_g()
# 角加速度を取得
gyro_x, gyro_y, gyro_z = get_gyro_data_deg()
# 枠内にデータを作成する
test_data_set_ax = remake_test_data_set(test_data_set_ax, accel_x)
test_data_set_ay = remake_test_data_set(test_data_set_ay, accel_y)
test_data_set_az = remake_test_data_set(test_data_set_az, accel_z)
test_data_set_gx = remake_test_data_set(test_data_set_gx, gyro_x)
test_data_set_gy = remake_test_data_set(test_data_set_gy, gyro_y)
# print (len(test_data_set))
# pickのDTWの値を取得する
pick_dtw_ax_result = dtw.getDTW(train_data_set_ax, test_data_set_ax)
pick_dtw_ay_result = dtw.getDTW(train_data_set_ay, test_data_set_ay)
pick_dtw_az_result = dtw.getDTW(train_data_set_az, test_data_set_az)
pick_dtw_gx_result = dtw.getDTW(train_data_set_gx, test_data_set_gx)
pick_dtw_gy_result = dtw.getDTW(train_data_set_gy, test_data_set_gy)
drop_dtw_ax_result = dtw.getDTW(drop_train_data_set_ax, test_data_set_ax)
drop_dtw_ay_result = dtw.getDTW(drop_train_data_set_ay, test_data_set_ay)
drop_dtw_az_result = dtw.getDTW(drop_train_data_set_az, test_data_set_az)
drop_dtw_gx_result = dtw.getDTW(drop_train_data_set_gx, test_data_set_gx)
drop_dtw_gy_result = dtw.getDTW(drop_train_data_set_gy, test_data_set_gy)
print(pick_dtw_ax_result, pick_dtw_ay_result, pick_dtw_az_result, pick_dtw_gx_result, pick_dtw_gy_result)
print(drop_dtw_ax_result, drop_dtw_ay_result, drop_dtw_az_result, drop_dtw_gx_result, drop_dtw_gy_result)
print(drop_dtw_ax_result - pick_dtw_ax_result, drop_dtw_ay_result - pick_dtw_ay_result, drop_dtw_az_result - pick_dtw_az_result, drop_dtw_gx_result - pick_dtw_gx_result, drop_dtw_gy_result - pick_dtw_gy_result)
if tt > 80 and operation_identification(drop_dtw_ax_result - pick_dtw_ax_result, drop_dtw_gx_result - pick_dtw_gx_result, drop_dtw_gy_result - pick_dtw_gy_result, accel_y, accel_z) == 'pick':
count += 1
# 加速度を取得
accel_x, accel_y, accel_z = get_accel_data_g()
# 角加速度を取得
gyro_x, gyro_y, gyro_z = get_gyro_data_deg()
# 枠内にデータを作成する
test_data_set_ax = remake_test_data_set(test_data_set_ax, accel_x)
test_data_set_ay = remake_test_data_set(test_data_set_ay, accel_y)
test_data_set_az = remake_test_data_set(test_data_set_az, accel_z)
test_data_set_gx = remake_test_data_set(test_data_set_gx, gyro_x)
test_data_set_gy = remake_test_data_set(test_data_set_gy, gyro_y)
# print (len(test_data_set))
# pickのDTWの値を取得する
pick_dtw_ax_result = dtw.getDTW(train_data_set_ax, test_data_set_ax)
pick_dtw_ay_result = dtw.getDTW(train_data_set_ay, test_data_set_ay)
pick_dtw_az_result = dtw.getDTW(train_data_set_az, test_data_set_az)
pick_dtw_gx_result = dtw.getDTW(train_data_set_gx, test_data_set_gx)
pick_dtw_gy_result = dtw.getDTW(train_data_set_gy, test_data_set_gy)
pick_dtw_gx_result_1 = dtw.getDTW(train_data_set_gx_2, test_data_set_gx)
pick_dtw_gy_result_1 = dtw.getDTW(train_data_set_gy_2, test_data_set_gy)
pick_dtw_gx_result_2 = dtw.getDTW(train_data_set_gx_3, test_data_set_gx)
pick_dtw_gy_result_2 = dtw.getDTW(train_data_set_gy_3, test_data_set_gy)
pick_dtw_gx_result_3 = dtw.getDTW(train_data_set_gx_4, test_data_set_gx)
pick_dtw_gy_result_3 = dtw.getDTW(train_data_set_gy_4, test_data_set_gy)
pick_dtw_gx_result_4 = dtw.getDTW(train_data_set_gx_5, test_data_set_gx)
pick_dtw_gy_result_4 = dtw.getDTW(train_data_set_gy_5, test_data_set_gy)
pick_dtw_gx_list.extend([pick_dtw_gx_result, pick_dtw_gx_result_1, pick_dtw_gx_result_2, pick_dtw_gx_result_3, pick_dtw_gx_result_4])
pick_dtw_gy_list.extend([pick_dtw_gy_result, pick_dtw_gy_result_1, pick_dtw_gy_result_2, pick_dtw_gy_result_3, pick_dtw_gy_result_4])
def makePrediction(dTemplate, toTest, avgDev): disregard, testDev = dtw.getDTW(dTemplate, toTest) avgSlope = linRegress.getSlopeForDifference([i for i in range(len(avgDev))], avgDev) testSlope = linRegress.getSlopeForDifference([i for i in range(len(testDev))], testDev) return 1 if testSlope <= avgSlope*THRESHOLD_PERCENT else 0
train_data_set_az = pd.read_csv('pick_train_data_2/pick_accel_z.csv',
usecols=[2]).values.reshape(-1, 1)
train_data_set_gx = pd.read_csv('pick_train_data_2/pick_gyro_x.csv',
usecols=[0]).values.reshape(-1, 1)
train_data_set_gy = pd.read_csv('pick_train_data_2/pick_gyro_y.csv',
usecols=[0]).values.reshape(-1, 1)
drop_train_data_set_ax = pd.read_csv('drop_train_data/drop_accel_x.csv',
usecols=[0]).values.reshape(-1, 1)
drop_train_data_set_ay = pd.read_csv('drop_train_data/drop_accel_y.csv',
usecols=[0]).values.reshape(-1, 1)
drop_train_data_set_az = pd.read_csv('drop_train_data/drop_accel_z.csv',
usecols=[0]).values.reshape(-1, 1)
drop_train_data_set_gx = pd.read_csv('drop_train_data/drop_gyro_x.csv',
usecols=[0]).values.reshape(-1, 1)
drop_train_data_set_gy = pd.read_csv('drop_train_data/drop_gyro_y.csv',
usecols=[0]).values.reshape(-1, 1)
# gx_result = dtw.getDTW(train_data_set_gx, drop_train_data_set_gx)
ax_result = dtw.getDTW(train_data_set_ax, drop_train_data_set_ax)
ay_result = dtw.getDTW(train_data_set_ay, drop_train_data_set_ay)
az_result = dtw.getDTW(train_data_set_az, drop_train_data_set_az)
gx_result = dtw.getDTW(drop_train_data_set_gx, train_data_set_gx)
gy_result = dtw.getDTW(train_data_set_gy, drop_train_data_set_gy)
print(ax_result)
print(ay_result)
print(az_result)
print(gx_result)
print(gy_result)
dtw.getDTWPath(train_data_set_gx, drop_train_data_set_gx)
log_test_data_set_gx = pd.read_csv('sencing_test_result/' + name_list[member_count] + '/log_' + log_num[log_count] + '.csv', usecols=[3]).values.reshape(-1, 1)
log_test_data_set_gy = pd.read_csv('sencing_test_result/' + name_list[member_count] + '/log_' + log_num[log_count] + '.csv', usecols=[4]).values.reshape(-1, 1)
log_test_data_set_gz = pd.read_csv('sencing_test_result/' + name_list[member_count] + '/log_' + log_num[log_count] + '.csv', usecols=[5]).values.reshape(-1, 1)
hash = 0
while len(log_test_data_set_ax) > count:
sec = time.time()
# 枠内にデータを作成する
test_data_set_ax = remake_test_data_set(test_data_set_ax, log_test_data_set_ax[count][0])
test_data_set_ay = remake_test_data_set(test_data_set_ay, log_test_data_set_ay[count][0])
test_data_set_az = remake_test_data_set(test_data_set_az, log_test_data_set_az[count][0])
test_data_set_gx = remake_test_data_set(test_data_set_gx, log_test_data_set_gx[count][0])
test_data_set_gy = remake_test_data_set(test_data_set_gy, log_test_data_set_gy[count][0])
test_data_set_gz = remake_test_data_set(test_data_set_gz, log_test_data_set_gz[count][0])
pick_dtw_gz_result = dtw.getDTW(train_data_set_gz, test_data_set_gz)
drop_dtw_ax_result = dtw.getDTW(drop_train_data_set_ax, test_data_set_ax)
drop_dtw_ay_result = dtw.getDTW(drop_train_data_set_ay, test_data_set_ay)
drop_dtw_gz_result = dtw.getDTW(drop_train_data_set_gz, test_data_set_gz)
flag = "test"
if operation_identification(pick_dtw_gz_result - drop_dtw_gz_result) == "drop" and drop_dtw_ax_result < 14:
flag = "drop"
drop_count += 1
insert_log_data(name_list[member_count], log_num[log_count], log_test_data_set_ax[count][0], log_test_data_set_ay[count][0], log_test_data_set_az[count][0], log_test_data_set_gx[count][0], log_test_data_set_gy[count][0], log_test_data_set_gz[count][0], 0, 0, 0, 0, drop_dtw_ax_result, drop_dtw_gz_result, pick_dtw_gz_result - drop_dtw_gz_result, flag)
test_data_set_ax = np.zeros_like(test_data_set_ax)
test_data_set_ay = np.zeros_like(test_data_set_ay)
test_data_set_az = np.zeros_like(test_data_set_az)
test_data_set_gx = np.zeros_like(test_data_set_gx)
test_data_set_gy = np.zeros_like(test_data_set_gy)
#2秒以上状態を維持する
if elapsed_time > 2:
# 200個のセンシングデータを作成する
while sencing_count <= 200:
#sencing_start_time = time.time()
gyro_x, gyro_y, gyro_z = get_gyro_data_deg()
#print(gyro_x)
# pickのDTWの値を取得する
test_data_set_gx = remake_test_data_set(
test_data_set_gx, gyro_x)
test_data_set_gy = remake_test_data_set(
test_data_set_gy, gyro_y)
pick_dtw_gx_result.append(
dtw.getDTW(train_data_set_gx, test_data_set_gx))
pick_dtw_gy_result.append(
dtw.getDTW(train_data_set_gy, test_data_set_gy))
print(
'=============================================================='
)
print(dtw.getDTW(train_data_set_gx, test_data_set_gx))
#print(gyro_y)
print(dtw.getDTW(drop_train_data_set_gx, test_data_set_gx))
print(
'==============================================================='
)
# DroｐのDTW値を算出
drop_dtw_gx_result.append(
dtw.getDTW(drop_train_data_set_gx, test_data_set_gx))