vsimko / rroad

R Package: Road Condition Analysis

rroad R package

Set of tools for monitoring road condition.

Currently, the following features are supported:

• International Roughness Index (IRI) computation
  • continuously increasing segment
  • fixed length overlapping segments with an offset

Download and Install

To download the development version of the package, type the following at the R command line:

install.packages("devtools")
devtools::install_github("vsimko/rroad")

To download the release version of the package on CRAN, type the following at the R command line:

install.packages("rroad")

Examples

profile <- rnorm(10000)
iri <- CalculateIRIperSegments(profile, IRI_COEF_100, 20)
par(mfrow = c(1,2)) # space for two diagrams
plot(profile, type = "l",
  xlab = "Distance [dm]", ylab = "Profile [mm]",
  main = "Read profile (Laser measurement)")
plot(iri, type = "s",
  xlab = "Segment", ylab = "IRI [m/km]",
  main = "International Roughness Index (IRI)\nsample = 10cm, segment = 20m")

References

Sayers, Michael W; Gillespie, Thomas D; Queiroz, Cesar A.V. 1986 The International Road Roughness Experiment (IRRE) : establishing correlation and a calibration standard for measurements. World Bank technical paper ; no. WTP 45. Washington, DC : The World Bank. Link

rroad's Issues

Use input dataset from a smartphone mounted to a bicycle

Would your package be able to estimate surface roughness based on data from a smartphone on a bicycle do you think? Thanks for publishing it in the open!

Make the reference in description smaller

(suggested by the reviewer)

For your next version:
Please omit the space before the colons in your description
(e.g. (IRRE) :).
In general, authors, year, ISBN and the ULR should be sufficient for the reference.

CRAN reviewer's comment about `DESCRIPTION`

Is there some reference about the method you can add in the Description field in the form Authors (year) doi:.....?

Better names for the functions

Propose some better naming schema for the functions.
The current names seem to be unnecessarily long.

update `inst/CITATION`

• update the citation, e.g. use the new title ...

CRAN reviewer's comment about `Authors@R`

We want to see person() calls directly since we do checks on this
and we cannot know if you have specified something incorrectly
in case you use as.person.

As in your case:

         Maintainer: Viliam Simko <[email protected]
         <mailto:[email protected]>>

is wrong, and the reason is that you string is incorrect,

  Authors@R: as.person(c(
               "Viliam Simko <[email protected]> [aut, cre]",
               "Kevin Laubis <[email protected]> [aut]"))

would have been correct, but then we directly have the example why
we want person() calls where we can check that they can be correctly
converted to the Author and Maintainer fields.

Best,
Uwe Ligges

Create a ROADMAP

Proposed functions:

• generate additional features from input 3d-accel / gyroscope / gps-velocity time series, such as wavelet coef. at multiple bands.
  • automatically compute context size
• train a random forest model given training data + ground truth
• predict IRI given input signals

update `README.Rmd`

• update the introduction paragraph, the package is now called "Road Condition Analysis"
• add "How to cite" section
• add "Who do I talk to?" section
• add "How to contribute" section
• mention somewhere that the generated vignette is available as a HTML page here: https://vsimko.github.io/rroad/

you can use https://github.com/taiyun/corrplot as an inspiration

update `vignettes/RoadFeatures.Rmd`

• add "Introduction" section to the vignette
• update the vignette that it actually makes sense for the rroad package because currently there is just some signal analysis which is not directly related to the IRI computation.

use `gh-pages` for publishing vignettes

related links:

• https://gist.github.com/cobyism/4730490
• http://rmflight.github.io/posts/2014/11/travis_ci_gh_pages.html

Proper matching of ground truth data with accelerometer measurements

For building prediction models, a mapping of the ground truth data and the accelerometer data is necessary.

Thus, for bringing laser measurements (e.g. referenced by starting position and driving offset) and accelerometer measurements (referenced by gps fixes) together, we need to map them on unique (e.g. OSM based) segments (e.g. 100m segments).

Assuming an point to point approach (start and end point of an road link) with relative offsets for defining the segments in between is a good way to go.

Solve seewave dependency which hinders installation of rroad

installing rroad causes following error:
"installation of package ‘seewave’ had non-zero exit status"