This library will estimate four gait parameters when given a dataset of vertical acceleration values (m/s^2).

The four parameters estimated are:

• Step Length (m)
• Cadence (step/min)
• Total time (s)
• Gait speed (m/s)

The collected data must be from an accelerometer mounted close to the patient's L5 vertebrae. The sensor should be pre calibrated against a 90 degree block. Only data from the vertical axis is required.

Simply copy both the gait_func.c and gait_func.h files to the same directory as the Arduino script.

Include the library header near the top of the Arduino script:

 #include "gait_func.h"

Session calibration is performed everytime time device is started. It compensates for the mounting angle of the sensor once it is on the person's back. To perform the calibration:-

1. Take X and Y axis measurements. Either single readings or an average. These values must be stored for use in step 4.
2. Pass these readings to the get_calibration_angle() function. Save this angle value for use in step 4.
3. Collect your data set.
4. Each X or Y axis value needs to be passed into either get_corrected_x() or get_corrected_y(). These return the corrected values using the angle measured in step 2.

The following pseudocode demonstrates how to use the library functions. The example data is collected at 100Hz for 5 seconds. Therefore n = 500 and ts (sample time) = 0.01. The value of l is the height of the sensor from the ground and is set as 1.1m in this example.

 float y-data[500] = vertical_acceleration_data();
 step_t my_step = calculate_step(500, 0.01, 1.1, y-data);

 Serial.println(my_step.length);
 Serial.println(my_step.cadence);
 Serial.println(my_step.time);
 Serial.println(my_step.gait_time);

If excessive integration error is present it might be better to use linear regression to improve the data. The method to use linear regression is the same execpt:

 step_t my_step = calculate_step_linreg(500, 0.01, 1.1, y-data);