The Weigh-in-Motion (WIM) Systems can collect data from different kind of devices, like Automatic License Plate Recognition (ALPR), weigh sensors (piezoelectric quartz, piezoelectric ceramic, piezoelectric polymer, etc), temperature sensors, etc. Additionally, the WIM Systems can produce a variety kind of data, like weigh data, classification data, violation information, some analytic reports, etc.
In this paper, it will be treat some data structures and caution about these data, presented on the main WIM literatures.
The goal of this paper is to collect the main data structure and file structure to standardise the OpenWIM input and output data. The OpenWIM is a open science project for weigh-in-motion research. Its goal is to be a standard platform to be used for WIM researchers in their investigation and to be a collaborative repository about this subject.
In the final report of COST 323 (Jacob, OBrien and Jehaes, 2002), in the Section 12 and Appendix IV, some aspects about data storage, processing, transmission and standard format. This information will be the main guideline to this study.
Here will be presentated some recommendations about the data and specifications used to cleaning the data before analysis stage (Enright and OBrien, 2011).
The data that will be treated here is:
The general recommendations is treated in section 1.1, covering the general aspects of data and file formats.
The output data standard layout from WIM systems, tipically has some attributes, treated on the section 1.2.
The section 1.3 has some recommendations to cleaning outpu data.
And, finally, the section 1.4 has the conclusion of this study.
To avoid any confusion while reading the output data files or using these data, it is important to consider some recommendations about the data and files, add some information to improve the completeness of the output data and apply some organization to this output (Jacob, OBrien and Jehaes, 2002).
Recommendations about the data:
The division scales should be, according to the accuracy classes (Jacob, OBrien and Jehaes, 2002):
Recommendations about the file structure:
Some information should be present in the output data file to be more completed and offer some detail to future analysis:
The data package should have some identification information about the WIM site:
To apply some organization to the output data, the first columns should be data common to all vehicle and, the last ones, should be columns when can have variation in the number of axles per vehicle (like axles load, distance between axles, etc).
Basically, the first main input data is the raw data sensors. These data can be acquired using a data acquisition device (DAQ) through the analog channels (weigh sensor, temperature sensors, etc) and the digital channels (i.e., inductive loops).
The three more wideley piezo-eletric weigh sensors used are piezo-ceramic, piezo-polymer and piezo-electric (Jiang, 2009).
It is essencial to store the raw sensor data during the investigation about the input signal and to validate weigh methods. In COST 323 (Jacob, OBrien and Jehaes, 2002), was not found any description about the raw data layout file. By the way, this data can be represented by a matrix using as a first column a index with time instant, it represented by microseconds in floating point format, followed by other columns representing each sensor data (in table 1, a exemple of sensor data structure).
The WIM data recorded per vehicle are typically (Enright and OBrien, 2011):
In (Qu, Lee and Huang, 1997) is showed a layout structure used in their research. Where, in the file name standard, these rules were used:
About the ASCII data file format, these rules were used:
Piezo-ceramic and piezo-polymer sensors are sensible to temperature influence and, for this reason, a temperature sensors is needed to apply some correction method (Burnos, 2008).
Some data cleaning criteria can be used to eliminate some undesireble data in statistical analysis. For example, in (Elkins and Higgins, 2008) is used these criteria:
In [28]:
from IPython.display import display, HTML
import pandas as pd
import numpy as np
import io
data = pd.DataFrame({
'A1_v': np.random.random(10),
'A2_v': np.random.random(10),
'D1_v': np.random.random(10)
})
data.index /= 5000
data.index.name = 'TIME_ss'
buffer = io.StringIO()
data.to_csv(buffer, sep=';')
buffer.seek(0)
print(buffer.read())
print('20151231')
[^](#ref-3) [^](#ref-5) [^](#ref-6) [^](#ref-8) Jacob, Bernard and OBrien, EJ and Jehaes, S. 2002. Weigh-in-Motion of Road Vehicles: Final Report of the COST 323 Action.
[^](#ref-4) [^](#ref-9) Enright, Bernard and OBrien, Eugene J. 2011. Cleaning weigh-in-motion data: Techniques and recommendations.
[^](#ref-7) Jiang, Xiaohua. 2009. Improvements in Piezoelectric Sensors and WIM Data Collection Technology.
[^](#ref-10) Qu, Tongbin and Lee, Clyde E and Huang, Liren. 1997. Traffic-load forecasting using weigh-in-motion data.
[^](#ref-11) Burnos, Piotr. 2008. Auto-calibration and temperature correction of WIM systems.
[^](#ref-12) Elkins, Lori and Higgins, Christopher. 2008. Development of truck axle Spectra from Oregon Weigh-in-motion Data for Use in Pavement Design and Analysis.