Measure types

carp_movisens_package has six measure types in the dk.cachet.carp.movisens namespace, which CARP Mobile Sensing (CAMS) collects from a Movisens sensor. Add each one to a protocol as Measure(type: MovisensSamplingPackage.<constant>), in a task that runs on a MovisensDevice:

All six are event-based; see Event-based and one-time measures. They stream while the sensor is connected, so start them with an ImmediateTrigger. The measures have no sampling configuration.

For the measurement envelope around the data, see Measurements and data.

Measures and data types

A measure type says what to collect, and the data type of each measurement says what it holds. For four measures, the two are the same. ACTIVITY and HR produce several data types:

To select measurements, use the data type, not the measure type, for example in measurementsByType() or in a SamplingEventTrigger. MovisensData has constants for these data types, such as MovisensData.STEPS:

client
    .getStudyController(study)
    ?.measurementsByType(MovisensData.STEPS)
    .listen((measurement) => print(jsonEncode(measurement)));

Delays and timestamps

The sensor computes most values over one minute and sends each one 0–84 seconds after measuring it, depending on the sensor and the value. Tap markers arrive without delay. Movisens lists the delay of each signal in Available signals per sensor.

The package does not correct for the delay: timestamp and sensorStartTime both give the phone's time when the value arrived. timestamp is in the phone's local time, without a time zone offset, and sensorStartTime is in microseconds since the epoch.

The package reads only live values. Values that the sensor measures while it is disconnected are not transferred later.

Common fields

Every Movisens data class extends MovisensData, which has these fields:

Physical activity

ACTIVITY collects the values of the sensor's physical activity service, about once per minute each. Body position, the metabolic equivalent of task (MET) and MET levels depend on the wearer's data in the MovisensDevice. See Add the device to the protocol.

The six data classes add these fields to the common fields:

{
  "__type": "dk.cachet.carp.movisens.activity.steps",
  "timestamp": "2026-10-26T10:23:20.123456",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "steps",
  "steps": 56
}

{
  "__type": "dk.cachet.carp.movisens.activity.body_position",
  "timestamp": "2026-10-26T10:23:20.894110",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "bodyPosition",
  "bodyPosition": "standing"
}

{
  "__type": "dk.cachet.carp.movisens.activity.met_level",
  "timestamp": "2026-10-26T10:23:21.012345",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "metLevel",
  "sedentary": 12,
  "light": 38,
  "moderate": 10,
  "vigorous": 0
}

Heart rate and HRV

HR collects the values of the heart rate variability (HRV) service of an EcgMove4, about once per minute each. The package does not collect the beat-by-beat heart rate that the sensor also sends.

The three data classes add these fields to the common fields:

All three classes use the data type dk.cachet.carp.movisens.hr.hr_mean. The constants MovisensData.HRV and MovisensData.IS_HRV_VALID exist, but no data uses them. Tell the three apart by the type field, or by class:

client.measurements.listen((measurement) {
  final data = measurement.data;
  if (data is MovisensHR) print('Mean heart rate: ${data.hr} BPM');
  if (data is MovisensHRV) print('RMSSD: ${data.hrv} ms');
  if (data is MovisensIsHrvValid) print('HRV valid: ${data.isHrvValid}');
});

{
  "__type": "dk.cachet.carp.movisens.hr.hr_mean",
  "timestamp": "2026-10-26T10:24:30.218930",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "hrMean",
  "hr": 72
}

{
  "__type": "dk.cachet.carp.movisens.hr.hr_mean",
  "timestamp": "2026-10-26T10:24:44.731002",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "rmssd",
  "hrv": 42
}

Electrodermal activity

EDA collects the mean skin conductance level (SCL) of an EdaMove4, about once per minute.

MovisensEDA adds this field to the common fields:

{
  "__type": "dk.cachet.carp.movisens.eda",
  "timestamp": "2026-10-26T10:25:44.731002",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "edaSclMean",
  "edaSclMean": 2.447584459974975
}

Skin temperature

SKIN_TEMPERATURE collects the mean skin temperature of the last minute, about once per minute.

MovisensSkinTemperature adds this field to the common fields:

{
  "__type": "dk.cachet.carp.movisens.skin_temperature",
  "timestamp": "2026-10-26T10:26:12.004518",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "skinTemperature",
  "skinTemperature": 33.12
}

Respiration

RESPIRATION reads the respiratory movement value of the sensor's respiration service.

MovisensRespiration adds this field to the common fields:

{
  "__type": "dk.cachet.carp.movisens.respiration",
  "timestamp": "2026-10-26T10:26:12.004518",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "respiratoryMovement",
  "value": 1342
}

Tap markers

TAP_MARKER reports each tap on the sensor without delay, for example when the participant marks a moment in the day.

MovisensTapMarker adds this field to the common fields:

A SamplingEventTrigger with measureType: MovisensSamplingPackage.TAP_MARKER starts a task on every tap. Leave out triggerCondition: Movisens data classes do not support conditions, so a trigger with one never fires.

{
  "__type": "dk.cachet.carp.movisens.tap_marker",
  "timestamp": "2026-10-26T10:23:20.123456",
  "deviceId": "88:6B:0F:3A:52:C1",
  "type": "tapMarker",
  "tapMarker": 1203
}

Data transformers

Registering the package adds transformers to the Open mHealth (OMH) and Fast Healthcare Interoperability Resources (FHIR) formats of CAMS:

To store data in one of these formats, set dataFormat on the data endpoint:

final protocol = SmartphoneStudyProtocol(
  ownerId: 'owner@example.com',
  name: 'Heart rate and steps in Open mHealth format',
  dataEndPoint: SQLiteDataEndPoint(dataFormat: NameSpace.OMH),
);

Data of the other data types keeps the CARP format. See Data transformation.