def calc_estimator_status_metrics(self) -> Dict[str, list]:
"""
calculates the estimator status metrics
:return:
"""
estimator_status_metrics = dict()
estimator_status_data = self.ulog.get_dataset('estimator_status').data
# add windowed metrics
estimator_status_metrics['{:s}_percentage_red_windowed'.format(self._check_id)] = \
calculate_windowed_mean_per_airphase(
estimator_status_data, 'estimator_status', self._test_ratio_name,
self._in_air_detector, threshold=params.ecl_red_thresh(),
window_len_s=params.ecl_window_len_s())
estimator_status_metrics['{:s}_percentage_amber_windowed'.format(self._check_id)] = \
calculate_windowed_mean_per_airphase(
estimator_status_data, 'estimator_status', self._test_ratio_name,
self._in_air_detector, threshold=params.ecl_amb_thresh(),
window_len_s=params.ecl_window_len_s())
estimator_status_metrics['{:s}_test_windowed_mean'.format(self._check_id)] = \
calculate_windowed_mean_per_airphase(
estimator_status_data, 'estimator_status', self._test_ratio_name,
self._in_air_detector, window_len_s=params.ecl_window_len_s())
return estimator_status_metrics
def calc_estimator_statistics(self) -> None:
"""
:return:
"""
if self._test_ratio_name is None:
return
estimator_status_data = self.ulog.get_dataset('estimator_status').data
estimator_status_metrics = self.calc_estimator_status_metrics()
innov_red_pct = self.add_statistic(
CheckStatisticType.INNOVATION_RED_PCT, statistic_instance=0)
innov_red_pct.value = float(
calculate_stat_from_signal(
estimator_status_data, 'estimator_status',
self._test_ratio_name, self._in_air_detector,
lambda x: 100.0 * np.mean(x > params.ecl_red_thresh())))
#TODO: remove subtraction of innov_red_pct and tune parameters
innov_amber_pct = self.add_statistic(
CheckStatisticType.INNOVATION_AMBER_PCT, statistic_instance=0)
innov_amber_pct.value = float(
calculate_stat_from_signal(
estimator_status_data, 'estimator_status',
self._test_ratio_name, self._in_air_detector, lambda x: 100.0 *
np.mean(x > params.ecl_amb_thresh()))) - innov_red_pct.value
innov_amber_pct.thresholds.warning = thresholds.ecl_amber_warning_pct(
self._check_id)
innov_amber_pct.thresholds.failure = thresholds.ecl_amber_failure_pct(
self._check_id)
innov_red_windowed_pct = self.add_statistic(
CheckStatisticType.INNOVATION_RED_WINDOWED_PCT,
statistic_instance=0)
innov_red_windowed_pct.value = float(
max([
np.max(metric) for _, metric in estimator_status_metrics[
'{:s}_percentage_red_windowed'.format(self._check_id)]
]))
innov_amber_windowed_pct = self.add_statistic(
CheckStatisticType.INNOVATION_AMBER_WINDOWED_PCT,
statistic_instance=0)
innov_amber_windowed_pct.value = float(
max([
np.max(metric) for _, metric in estimator_status_metrics[
'{:s}_percentage_amber_windowed'.format(self._check_id)]
]))
# the max and mean ratio of samples above / below std dev
test_ratio_max = self.add_statistic(
CheckStatisticType.ESTIMATOR_FAILURE_MAX, statistic_instance=0)
test_ratio_max.value = float(
calculate_stat_from_signal(estimator_status_data,
'estimator_status',
self._test_ratio_name,
self._in_air_detector, np.amax))
test_ratio_avg = self.add_statistic(
CheckStatisticType.ESTIMATOR_FAILURE_AVG, statistic_instance=0)
test_ratio_avg.value = float(0.0)
if test_ratio_max.value > 0.0:
test_ratio_avg.value = float(
calculate_stat_from_signal(estimator_status_data,
'estimator_status',
self._test_ratio_name,
self._in_air_detector, np.mean))
test_ratio_windowed_avg = self.add_statistic(
CheckStatisticType.ESTIMATOR_FAILURE_WINDOWED_AVG,
statistic_instance=0)
test_ratio_windowed_avg.value = float(
max([
float(np.mean(metric))
for _, metric in estimator_status_metrics[
'{:s}_test_windowed_mean'.format(self._check_id)]
]))