UWHandles is a dataset for 6D object pose estimation in underwater fisheye images. It provides 6D pose and 2D bounding box annotations for 3 different graspable handle objects used for ROV manipulation. The dataset consists of 28 image sequences collected in natural seafloor environments with a total of 20,427 annotated frames. and Meta repository for the dataset
https://github.com/gidobot/UWHandles
Billings, G., & Johnson-Roberson, M. (2020). SilhoNet-fisheye: Adaptation of a ROI based object pose estimation network to monocular fisheye images. IEEE Robotics and Automation Letters, 5(3), 4241-4248.
UWslam is a dataset for underwater stereo and hybrid monocular fisheye + stereo SLAM in natural seafloor environments. The dataset includes a spiral survey of a shallow reef captured with a diver operated stereo rig and 4 hybrid image sequences captured with a deep ocean ROV in different deep ocean environments. Ground truth pose estimates for the spiral stereo trajectory were obtained by processing the image sequence through COLMAP. Ground truth pose estimates for the hybrid sequences were obtained by distributing fiducials on the seafloor before capturing an image sequence and processing the image sequences with the ROS based TagSLAM package.
G. Billings, R. Camilli and M. Johnson-Roberson, "Hybrid Visual SLAM for Underwater Vehicle Manipulator Systems," in IEEE Robotics and Automation Letters, vol. 7, no. 3, pp. 6798-6805, July 2022, doi: 10.1109/LRA.2022.3176448.
The data and scripts are meant to show how burster dynamics determine response to a single biphasic stimulus. The files include data which show trends in the propensity of termination for different burster types and the MATLAB scripts used to generate this data. The MATLAB scripts also allow the user to generate their own data sets for alternative bursting paths and stimulus parameter combinations. Furthermore, they allow the user to visually examine the effects of single stimuli in the voltage timeseries and in state space. How the user can access these features of the script is described in the file "ReadMe.pdf."
In this work, we perform Global Sensitivity Analysis (GSA) for the background solar wind in order to quantify contributions from uncertainty of different model parameters to the variability of in-situ solar wind speed and density at 1au, both of which have a major impact on CME propagation and strength. Scripts written in the Julia language are used to build the PCE and calculate the sensitivity results. Data is available in csv, NetCDF and JLD files. A `Project.toml` file is included to activate and install all required dependencies (See README for details).
The goal of the research was to train a surrogate model for the prediction of electric field distribution for a given electrospray emitter geometry design. The surrogate is to be used in reduced-fidelity modeling of electrospray thruster arrays. The code repository is included in the README.txt file.
J.D. Eckels, C.B. Whittaker, B.A. Jorns, A.A. Gorodetsky, B. St. Peter, R.A. Dressler, “Simulation-based surrogate methodology of electric field for electrospray emitter geometry design and uncertainty quantification”, presented at the 37th International Electric Propulsion Conference, Boston, MA USA, June19-23, 2022 Available: https://www.electricrocket.org/IEPC_2022_Papers.html
To produce this dataset, three modes of the flywheel were tested. The first was with the flywheel off, which produced a baseline for roll without stabilization. The second mode was active stabilization with the flywheel spinning. An IMU on board took in roll in degrees. An Arduino uno used the roll angle to precess the flywheel to a degree that countered the roll. The last mode was passive stabilization with the flywheel on. Here, the precession belt was removed which allowed the flywheel to freely precess and counter the moment generated by the roll.
This data contains 3 magnetometer signals of 4 noise sources. It was created to test a Underdetermined Blind Source Separation algorithm for magnetic signals.
The data was used to calibrate and simulate pattern recognition algorithms for the following publication: Surgically Implanted Electrodes Enable Real-Time Finger and Grasp Pattern Recognition for Prosthetic Hands (medRxiv 2020, IEEE TRO in review). Each data file is named as follows Px_PostureSet.csv. Where Px is the patient number. The 1 of 10 posture set contains individual finger and intrinsic thumb movements, the grasps posture set contains a fewer number of combined finger movements. P1’s calibration data for individual fingers is labelled 1 of 12 because it also includes two grasps, which were removed for analysis in the publication. The first column of each .csv file is the experiment time in seconds. The second column is the posture of the cue hand at that timestamp. The rest of the columns are the raw EMG data in microvolts sampled at 30KSps. A legend of the movement postures, each patients EMG channels, and suggested signal processing and filtering is included in DataLabellingAndProcessing.pdf
Surgically Implanted Electrodes Enable Real-Time Finger and Grasp Pattern Recognition for Prosthetic Hands A. K. Vaskov, P. P. Vu, N. North, A. J. Davis, T. A. Kung, D. H. Gates, P. S. Cederna, C. A. Chestek medRxiv 2020.10.28.20217273; doi: https://doi.org/10.1101/2020.10.28.20217273
Statistical study of Swarm observations and two Earth magnetic field models: IGRF-12 and CHAOS-6 categorized by Kp*10 index. Data analysis done on https://viresclient.readthedocs.io/en/latest/ JupyterLab.
These data were produced in an attempt to characterize the turning and steering behaviors of 1-DoF multi-legged (hexpedal in this case) robots. Such turning behaviors require sliding contact points.
The .tar file contains multiple trials in .csv.gz format, with names following an informative naming convention documented in the README.
Additional metadata for the trials is given in the metadata.py file in both machine and human readable form.