READ ME File For
'Delphin2 Testwood Experimental Data Altitude Tracking November 2017'

Licence: CC-By
Dataset DOI: https://doi.org/10.5258/SOTON/D1157

Provided by: Sophia Schillai, University of Southampton, sms4g13@soton.ac.uk, ORCID ID 0000-0003-3349-4169
Date of data collection: 2017-11-01, files are dated individually
Location: Testwood Lake, Hampshire, UK, origin of the local coordinate frame is: (lat: 50.93664083, lon: -1.50766432)
Language: CSV, plaintext, rosbag data format
Funding: National Oceanography Centre, University of Southampton

This dataset supports the thesis entitled "Auonomous Underwater Vehicle Photographic Surveys in Complex Terrain"
Awarded by: University of Southampton
Award year: 2019
Publications:
- Experimental analysis of low altitude terrain following for hover capable flight style autonomous underwater vehicles. Schillai, Sophia, Maria, Turnock, Stephen, Rogers, Eric and Phillips, Alexander (2019) Experimental analysis of low altitude terrain following for hover capable flight style autonomous underwater vehicles. Journal of Field Robotics. (doi:10.1002/rob.21910).
- PhD thesis "Autonomous Underwater Vehilce Photographic Surveys in Complex Terrain". Sophia Schillai, University of Southampton, October 2018.

Summary: This dataset contains part 2 (of 3) of experimental data collected with the Delphin2 AUV at Testwood Lake in 2017 and 2018. The vehicle repeatedly
crosses a step in the lakebed, using different methods for terrain detection and altitude tracking.
The data_overview.csv file provides a machine readable overview of the data.
For each experiment in the table, a rosbag and a parameter file are included; the filename for each file includes a date- and timestamp.
The rosbag uses the ROS standard compatible with at least ROS indigo, ROS melodic.
The parameter file includes all parameters that were configured outside of the code.
The code running on Delphin2 is available on github: https://github.com/Southampton-Maritime-Robotics/DelphinROSv3

Data collection method: Data collected by the Autonomous Underwater Vehicle Delphin2, collection is described in detail in the Methods section of the PhD thesis.

data_overview.csv provides a machine readable overview of the data.
The columns are:
- index: the index of the experiment, a continuous, unique count, used to easier reference each of the files
- month, day, hour, minute: the month, day, hour, and minute of the experiment start (the experiments can be referenced to Appendix A with this information)
- case: the experiment case, as described in the experiment matrix in Table 4.4
- t_up: manually read timestamp when the upwards step was crossed (C or D, depending on the direction), from the sum of the altitude and the depth; in seconds
- t_down: manually read timestamp when the downwards step was crossed (D or C, depending on the direction), from the sum of the altitude and the depth; in seconds
- detect_range: estimate of the detection range of the sonar, manually read from a plot of all sonar scans with detections; in metres
- error_rate: number of altimeter measurements that was removed due to the altimeter jumping by an unreasonable amount relative to the previous measurement
- prop: propeller setpoint in RPM
- direction: positive when traveling from A to B, negative in the reverse direction
- goal_alt: goal altitude in metres
- pseudo_alt: -1 if only the altimeter value and not the pseudo altitude was used, 1 if the forwards detectionw was used for obstacle avoidance
- sonar_threshold: detection threshold S_th of the sonar detection, 0 if forwards detection was not used
- sonar_slide: sliding threshold S_r, threshold reduction used for cases 11 and 16; 0 else
- note: manual note added about the experiment
- filename: name of the rosbag for the experiment
The rest of the data was obtained through automated processing of the previous columns:
- step_time: t_down - t_up
- scale_time_to_x: step_time/length of the step, between C and D; used to correct for the vehicle speed
- depth_offset: comparing the sum of the altitude and depth sensor, the error in water level was corrected with this value
- energy/power levels: calculated through integration over the sensor values between the start and end time of the experiment