https://doi.org/10.5061/dryad.15dv41p49

The dataset consists of acceleration and position measurements taken using smartphones as vehicles drove across various bridges. Current projects using this data revolve around mobile sensing and system identification. The accelerations were measured using the smartphone’s onboard triaxial accelerometer. However, the position measurements are a process version of the GPS measurements. The position measurements are 1D distance from a bridge pier, measured in meters, rather than the raw 2D GPS measurements from the phone. Within this repository, there are multiple case studies, and between cases, the method of data collection varied. The methods for data collection can be found in the related work.  A Jupyter Notebook accompanies the datasets to reproduce the Golden Gate Bridge absolute mode shape results in the related work.

Description of the data and file structure

The data is saved in MAT file format and can be accessed using MATLAB or various Python libraries. The file names are labeled with the bridge identifier in the first three digits, and trip index in the second three digits. In each file, the acceleration measurements are saved in a variable called “aOutput,” a matrix 3xN where three rows represent the three axes of the accelerometer. Next, the position measurements are saved in a variable named “S_x,” a 1xN matrix. Last, there is a variable named “dt” for the timestep of the measurements. These are the necessary variables to replicate the results in the associated works. The bridge ID and variables in the respective mat files are listed below:

ggb_params.npy

• Parameters saved in a python dictionary to recreate controlled GGB results in JupyterNotebook supplied in repository 

Cadore (Bridge ID: 332)

• aOutput
• S_x
• dt

Ciampino (Bridge ID: 346)

• aOutput
• S_x
• dt

GGB Controlled (Bridge ID: 198)

• aOutput
• S_x
• dt
• uuid, seq, coverage, bridgeId (misc. indexes and variables used by research team, type sting or float)

GHMB (Bridge ID: 334)

• aOutput
• S_x
• dt
• uuid, seq, coverage, bridgeId, date, phoneID, start_time (misc. indexes and variables used by research team, type sting or float)

UBER (Bridge ID: 298)

• aOutput
• S_x
• dt

Code

The Jupyter Notebook supplied with the code is self-contained and can be run using Google Colab. To run the code, simply upload the code and data into google drive. Open the notebook using the Google application Google Colab and follow the instructions within the notebook. The “ggb_params.npy” is included in the repository to replicate the controlled GGB absolute mode shape results.