Data Sheet

October 2017 BNO080 Datasheet 1000-3927
www.hillcrestlabs.com © 2017 Hillcrest Laboratories, Inc. All rights reserved. 35 / 57
2.4.4 Step Counter
The step counter uses the step detector to detect and count steps. It provides a more accurate indication of steps
taken than the step detector. It provides increased accuracy by evaluating the data around each step event,
possibly reclassifying previous samples as either steps or non-steps, depending upon the patterns perceived. It
outputs a 16-bit step count. The application processor maintains a 64-bit count of the total number of steps taken
and must therefore manage any roll-overs in the step counter output. The step counter will assert the host
interrupt while the host processor is asleep if the 16-bit step counter is close to overflow to ensure steps are not
lost.
The step counter has two optimizations, one for wrist worn devices and the second for all other use cases.
2.4.5 Activity Classification
The activity classifier uses the accelerometer to determine if the user is walking, running or climbing stairs. The
climbing stairs feature requires the use of an external pressure sensor. The sensor is configurable to allow
customization for a particular device and/or a particular individual. The configurable options relate to the step rate
for running versus walking and to the variation of the signal strength of the accelerometer.
The configurable options are:
Thresholds related to Step Rates
• runLowerThldStep: The lowest step rate for running. The default is 1.25 steps/s
• walkUpperThldStep: The highest step rate for walking. The default is 2.25 steps/s
• walkLowerThldStep: The lowest step rate for walking. The default is 0.75 step/s
Thresholds related to Step Signal Variation Strength:
• runUpperThldStd: The highest signal variation strength for running. The default is 23.50m/s
2
• runLowerThldStd: The lowest signal variation strength for running. The default is 3.20m2/s
2
• walkUpperThldStd: The highest signal variation strength for walking. The default is 4.50m2/s
2
• walkLowerThldStd: The lowest signal variation strength for running. The default is 0.75m2/s
2
The classification considers steps within a 4s window and the variation in signal strength is determined by taking
the standard deviation of the accelerometer normal over the 4s window.
With the default settings a classification matrix as depicted in Figure 2-4 can be created.
Figure 2-4: Activity classification matrix