Introduction

This manual describes the algorithms provided by movisens DataAnalyzer. Some of these algorithms are also implemented directly in the sensor firmware of movisens sensors (generation 3 or higher) and can be calculated live for triggering movisens XS via Sensor Trigger.

In DataAnalyzer you can specify an output interval. The output of most algorithms can be aggregated so that one value is output per output interval. The default is 1min.

Sensor data

Algorithms described here are working on raw sensor data, i.e. data that has not been aggregated in whatever form. This has several advantages:

  • Data is not bound to any proprietary algorithms
  • Data can be reanalyzed if better algorithms are available, without any restriction that could have been applied by any aggregation step
  • The complete series of analysis steps from raw data to final output can be monitored

Data is stored in the free and open unisens data format.

Depending on sensor type the following raw data signals are available from movisens sensors:

Parameter Unisens Name Sample rate [Hz]
Electrocardiogram ecg.bin 1024
Electrodermal Activity eda.bin 32
Acceleration in three dimensions acc.bin 64
Angular Rate in three dimensions angularRate.bin 64
Atmospheric Air Pressure press.bin 8
Ambient Light in 5 channels light.bin 1
Temperature temp.bin 1

Please check the specific sensor manual for details.

Sensor location

Many of the algorithms are dependent on the location where the sensor is worn. The following picture shows the definition of the locations.

Sensor locations

Depending on the sensor location some output parameters can / cannot be calculated (e.g. sitting and standing can only be distinguished when the sensor is worn at the thigh). For many algorithms the sensor location has an effect on quality/accuracy. Please check the descriptions of the output parameters for more information.

Last Updated: 4/10/2019, 4:21:14 PM