data and code for associated research manuscript submitted to PLOS ONE for review 01/03/2022
Manuscript Title: A novel smartphone-based activity recognition modeling method for tracked equipment in forest operations

Data use:
License: CC-BY 4.0 https://creativecommons.org/licenses/by/4.0/
Citation for dataset reuse:
Becker, R., & Keefe, R. (2022). Activity Recognition Model: Sensor (AS), GNSS, Time and Motion Study (TM), and Individual Tree Data [Data set]. University of Idaho. https://doi.org/10.7923/1VQA-M812

Data Files:
Becker_Keefe_DOI_2022.txt:
Discovery level metadata utilized to establish DOI for dataset

"AS_TM_GNSS_Combined" folder:
The provided data are the raw sensor data gathered from the Google Pixel smartphones used during the field sampling for the development of the activity recognition models. Additionally, corresponding GNSS locations and observed cycle elements from time and motion study are included and are all matched by time stamps. Data is provided based on the sampling rates used for the study (10, 20, 50Hz) These data are labeled (AS_TM_GNSS_"date collected"_"sampling rate in Hz").
Data Files:
AS_TM_GNSS_8_10_10.csv
AS_TM_GNSS_8_10_20.csv
AS_TM_GNSS_8_10_50.csv
AS_TM_GNSS_8_11_10.csv
AS_TM_GNSS_8_11_20.csv
AS_TM_GNSS_8_11_50.csv
AS_TM_GNSS_9_20_10.csv
AS_TM_GNSS_9_20_20.csv
AS_TM_GNSS_9_20_50.csv
AS_TM_GNSS_9_21_10.csv
AS_TM_GNSS_9_21_20.csv
AS_TM_GNSS_9_21_50.csv
AS_TM_GNSS_9_26_10.csv
AS_TM_GNSS_9_26_20.csv
AS_TM_GNSS_9_26_50.csv
AS_TM_GNSS_9_27_10.csv
AS_TM_GNSS_9_27_20.csv
AS_TM_GNSS_9_27_50.csv
Data Dictionary for headers:
[Col 1 Unlabeled]:
Time:Time: time and motion study timestamp when data point was observed. format-HH:MM:SS using 24hr clock. Timezone-Pacific Daylight Time (PDT) -07:00 from UTC.
Accel.X: Accelerometer measurement in the X-axis - m/sec2
Accel.Y: Accelerometer measurement in the Y-axis - m/sec2
Accel.Z: Accelerometer measurement in the Z-axis - m/sec2
Grav.X: Forces of gravity derived from the accelerometer in the X-axis - m/sec2
Grav.Y: Forces of gravity derived from the accelerometer in the Y-axis - m/sec2
Grav.Z: Forces of gravity derived from the accelerometer in the Z-axis - m/sec2
L.Accel.X: Linear acceleration measurement in the X-axis - m/sec2
L.Accel.Y: Linear acceleration measurement in the Y-axis - m/sec2
L.Accel.Z: Linear acceleration measurement in the Z-axis - m/sec2
Gyro.X: Gyroscope measurement in the X-axis - rad/sec
Gyro.Y: Gyroscope measurement in the Y-axis - rad/sec
Gyro.Z: Gyroscope measurement in the Z-axis - rad/sec
Mag.X: Measurement of the magnetic force in the X-axis - microTelsa (uT)
Mag.Y: Measurement of the magnetic force in the Y-axis - microTelsa (uT)
Mag.Z: Measurement of the magnetic force in the Z-axis - microTelsa (uT)
Orient.Z: Orientation in the Z-axis (roll) - degrees
Orient.X: Orientation in the X-axis (magnetic north) - degrees
Orient.Y: Orientation in the Y-axis (relative to ground) - degrees
Sound: Decibel level (dB) from smartphone based sound pressure meter
Dur.MS: Duration of the cycle element in milliseconds (MS)
AS.Time: Androsensor timestamp when data point was recorded. format-HH:MM:SS.sss using 24hr clock. Timezone-Pacific Daylight Time (PDT) -07:00 from UTC.
X: Sequential cycle number; Null values: NA
Element:Cycle element description observed; Null values: NA
Cam.Lat: Latitude of the dash camera during field sampling; Null values: NA
Cam.Long: Longitude of the dash camera during field sampling; Null values: NA
Cab.Lat: Latitude of the cab-mounted GNSS transponder camera during field sampling; Null values: NA
Cab.Long: Longitude of the cab-mounted GNSS transponder camera during field sampling; Null values: NA
Boom.Lat: Latitude of the boom-mounted GNSS transponder camera during field sampling; Null values: NA
Boom.Long: Longitude of the boom-mounted GNSS transponder camera during field sampling; Null values: NA

"Single_Tree_Data" folder:
Individual tree records derived from the lidar acquisition for the study area. These data are labeled (Unit_#_Trees).
Data Files:
Unit_1_Trees.csv
Unit_2_Trees.csv
Unit_3_Trees.csv
Data Dictionary for headers:
FID: ID used for individual tree
species: Species code for the tree
dbh_in: Diameter at breast height (DBH) in inches
height_ft: Tree height in feet
crown_ht_f: Crown base height in feet
crown_diam: Crown diameter in feet
defect_lik: Defect code
gross_bdft: Gross tree volume in board feet
stand: Management stand where tree is located
tree_long: Longitude of individual tree
tree_lat: Latitude of individual tree


Code:
"Random_forests" folder:
The R code provided gives foundational R-code for random forest model training and final models for all windows (1, 5, 7.5, and 10 seconds) for each of the sampling rates (10, 20, and 50Hz) for natively sampled and downsampled (ds) data.
Code Files:
Random_Forest_10hz_ALL_WINDOWS_ds.R
Random_Forest_10hz_ALL_WINDOWS.R
Random_Forest_20hz_ALL_WINDOWS_ds.R
Random_Forest_20hz_ALL_WINDOWS.R
Random_Forest_50hz_ALL_WINDOWS.R
TRAIN_Random_Forest_10hz_10_ds.R
TRAIN_Random_Forest_10hz_10.R
TRAIN_Random_Forest_20hz_10_ds.R
TRAIN_Random_Forest_20hz_10.R
TRAIN_Random_Forest_50hz_10.R