This is a subset of the SHIFDR dataset collection containing data from 14 buildings in Southeast Michigan. The full dataset collection can be found at https://deepblue.lib.umich.edu/data/collections/vh53ww273?locale=en and Organization: We include a subfolder for each building, identified by name. All buildings have been renamed after lakes to protect the identity of the building. Within each building subfolder, there is fan power (i.e. current measurements from which fan power can be computed), building automation system (BAS), whole building electrical load (WBEL), and voltage data collected over the course of our experimentation from 2017 to 2021. All experiments were conducted in the summer months and a full schedule of Demand Response (DR) events is included along with each building in the ‘Event_Schedule.csv’ file. The building information file contains general information about the buildings, pertinent to the experiments we conducted.
There is also a folder labeled ‘2021 Preprocessed data’ which contains combined BAS and fan power data from the summer of 2021. This data has been lightly processed to calculate fan power from current measurements and interpolate BAS data to 1 minute intervals. These act as an easy-to-use starting point for data analysis.
Citation to related publication:
A.J. Lin, S. Lei, A. Keskar, I.A. Hiskens, J.X. Johnson, and J.L. Mathieu. “The Sub-metered HVAC Implemented For Demand Response (SHIFDR) Dataset,” Submitted, 2023.
Data included in raw format in addition to the MATLAB scripts used for processing into final results. If there are issues or confusion regarding this data or the codes, feel free to contact me at tategill@umich.edu.
To enhance environmental turbulence measurements, we have designed and constructed a novel Particle Image Velocimetry (PIV) instrument intended for field use. The data contained here was used for either validation of the instrument, or was produced by the instrument in proof-of-concept field testing.
PedX is a large-scale multi-modal collection of pedestrians at complex urban intersections. The dataset provides high-resolution stereo images and LiDAR data with manual 2D and automatic 3D annotations. The data was captured using two pairs of stereo cameras and four Velodyne LiDAR sensors.
The work guides the processing of CAM6 data for use in machine learning applications. We also provide workflow scripts for training both random forests and neural networks to emulate physic s schemes from the data, as well as analysis scripts written in both Python and NCL in order to process our results.
Limon, G. C., Jablonowski, C. (2022) Probing the Skill of Random Forest Emulators for Physical Parameterizations via a Hierarchy of Simple CAM6 Configurations [Pre Print]. ESSOAr. https://10.1002/essoar.10512353.1
The collection contains the code and the data used to train machine learning algorithms to emulate simplified physical parameterizations within the Community Atmosphere Model (CAM6). CAM6 is the atmospheric general circulation model (GCM) within the Community Earth System Model (CESM) framework, developed by the National Center for Atmospheric Research (NCAR). GCMs are made up of a dynamical core, responsible for the geophysical fluid flow calculations, and physical parameterization schemes, which estimate various unresolved processes. Simple physics schemes were used to train both random forests and neural networks in the interest of exploring the feasibility of machine learning techniques being used in conjunction with the dynamical core for improved efficiency of future climate and weather models. The results of the research show that various physical forcing tendencies and precipitation rates can be effectively emulated by the machine learning models.
The data represents weekly output from three 60-year CAM6 model runs. The output includes state (.h0. files) and tendency (.h1. files) fields for three difference model configurations of increasing complexity. State fields include temperature, surface pressure, specific humidity, among others; while tendencies include temperature tendencies, specific humidity tendencies, as well as precipitation rates. Using the state variables at a given time step, machine learning techniques can be trained to predict the following tendency field, which can then be applied to the state variables to provide the state at the next physics time step of the model.
Limon, G. C., Jablonowski, C. (2022) Probing the Skill of Random Forest Emulators for Physical Parameterizations via a Hierarchy of Simple CAM6 Configurations [Preprint]. ESSOAr. https://10.1002/essoar.10512353.1
The Sub-metered HVAC Implemented for Demand Response (SHIFDR) dataset is a massive dataset that captures the response of individual commercial building HVAC system components to demand response events. The dataset includes device-level power consumption during demand response events as well as during normal operation. We have organized the data into subsets, with each subset containing data from buildings in different parts of the world. Kindly refer to the README file within each subsection for specific information about how data is organized. Please reach out if you have data that you would like to share, find any mistakes in the data, or have any questions. We are always trying to improve SHIFDR.
This research introduces a novel method to produce biomimetic shapes using low cost soluble 3D printed molds. Mesenchymal stem cells in alginate matrix cell viability was studied. The alginate stem cell structure is made in a construct that is 21 mm wide, 6 mm high, with an arbor diameter of 1 mm (see Combined_Test_Channels.stl). The cells showed 64% survivability at 7 days in the 3D constructs.
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.
2Works