arduinor package provides a simplified approach to get serial data from arduino to R. Right now it depends on Tod E. Kurt's arduinoSerial C library, which can only runs on POSIX-compatible systems, such as Mac OS X and Linux. As a result, for now this package can only runs on these systems as well.

I'm working on getting this libserialport C++ package to R. Once I finished the importing, I will try to move the dependencies of this package on top of that. So hopefully, in the future, we will be able to get this work on Windows as well.

devtools::install_github("r-arduino/arduinor")

For a more detailed walk-through, check out this blog-post.

This package offers a simplified workflow to work with arduino connections.

library(arduinor)

con <- ar_init("/dev/cu.SLAB_USBtoUART", baud = 57600)  

ar_flush_hard(con) # Flush the port to remove initial random readings
ar_read(con)
[1] "8510106,8976,-7748,-11848\r\n"

You can find the connection port at the footnote of your Arduino IDE window.

And you should have setuped your Baud rate somewhere in your arduino void setup

void setup(){
  ...
  Serial.begin(57600);
  ...
}

ar_monitor streams the arduino output into your R console, like the "Serial Monitor" feature in Arduino IDE. It will stop when you hit the STOP button.

ar_collect is similar with ar_monitor except the fact that it will save a limited amount of data into a R.

> dt <- ar_collect(con, size = 300)
Flushing Port...
Done  
> length(dt)
[1] 300

ar_plotter aims to provide you something similar with the "Serial Plotter" in Arduino IDE but give you more power to control:

• You have the ability to Pause/Start.
• You have the ability to select variable to display.
• You have the ability to name the variables in the way you provided.
• You can choose to save the observations in a csv file.

However, note that due to a limitation in plotly (? not confirmed yet), this plotter can't visualize signals sampling at a 25Hz or higher (on my computer) and it will freeze your R session if you do that. Therefore for now I added a reduce_freq option to ar_plotter with default ON. It will add a 40ms (0.04s) delay for every reading. If you are looking for a more precised way to collect data, you can either choose to turn this option off or use ar_collect, which is much simpler and can work in any condition.

Part of this work goes to the a study project running at Marcus Institute for Aging Research. It won't happen without the support from Brad Manor and his grant funding (1-K01-AG044543–01A1). This project was also supported by the leadership from Tom Travison and the funding support from the Boston Claude D. Pepper Older Americans Independence Center (P30-AG013679).