Automated Data Collection and Analysis

[The_Oceanography_Kid]

I am graduate student in Oceanography, and I have been tasked with trying to write MATLAB code for an automated system that measures seawater pH using an Agilent Spectrophotometer based off of a LabVIEW Program that performs a similar function. In addition, to the computer that will have all of the appropriate software for this, but will need to communicate with the spectrophotometer, as well as a syringe pump that controls the fluid flow for the entire system. I have a basic programming background my undergradaute work, but never dealt with anything this complicated. Can anyone offer any advice on where to start with this (i.e. helpful links, articles, books, or commands that would be useful). Thanks.

 

-The_Oceanography_Kid

[AtomicMaster]

Step 1:

Get model numbers off the devices, and figure out what the hell kind of a protocol is used to communicate with those devices, get full spec, how they work, what they can and can not do.

Step 2:

Get some sample data.

Step 3:

Try to find some libraries to help you communicate using the protocol.

Step 4:

Implement the solution/protocol in code (if no solution is found already)

Step 5:

Make sure that when you feed the sample data into your program that you see the expected outcomes

Step 6:

Write a protocol fuzzer

Step 7:

Fuzz your code with at least a few hundred million messages

Step 8:

Figure out what software your software is going to interface with

Step 9:

Figure out how to interface with that software

Step 10:

Follow step 2 through step 7 to develop that side of things

Step 11:

Test you solution on the software side with fake data

Step 12:

Test your solution on the hardware side with fake actions

Step 13:

Test the whole system and make sure that you get expected results both ways

Step 14:

Fuzz the whole system

Step 15:

Have a drink, play with your final solution.

[Sensei]

That's job for somebody who has a few years everyday experience in writing software..

 

Figuring out device models and checking manufacturers websites for SDK, if not found maybe writing e-mail to them is the first thing you need to do.

Such devices often comes with software to install, so maybe SDK is already on CD/DVD or installed?

Edited May 7, 2014 by Sensei

[AtomicMaster]

More correctly it's a job for somebody who has a lot of experience in protocol developement, computational linguistics or compilers... Implementing protocols correctly and safely is part science, part black magic.

[Greg H.]

...part black magic.

Black magic would be easier to learn.

[AtomicMaster]

Heh, indeed

[John Cuthber]

Other people have learned this black magic. Given time and teaching, so can the OP.
What's the timescale for the project?