Requirements

• A Cellphone with credit and its USB cable
• Java version of SMSLib (download of sources here)
• Apache Jakarta Commons.Net required by SMSLib
• Slf4J also required by SMSLib.

Introduction
It is possible to use a programming language to send Hayes Commands (aka AT commands) through the serial port to a cellphone modem, plugged into a PC . With these commands the cellphone can send SMS.

We could do the connection with the Serial port and write every AT command programatically. However, it is complex and would require learning the the AT Commands list.
In order to avoid this complexity you can use the SMSLib, intead of writting it manually.

Installing

• Configure the Java communication library according to the SMSLib Installation guide
• The installation guide tells how to build SMSLib with Ant. I prefer building it with Eclipse. Thue, I have imported the sources to a new Java Project inside of Eclipse.
• Add a reference to the Slf4j library and Commons.Net library in the recently created Eclipse project.
• Connect the cellphone in the computer. You can use USB, BlueTooth, IrD, etc.
• Find out what is the COM port associated to the new connection and its baud rate. Keep it in memory.
  On Windows: Contro Panel -> system -> Device Manager -> Modens. Right click on the modem, tab Modem and get the value of “port”. This value is similar to COM3.

Running a sample
The SMSLib comes with some good samples. The SendMessage class is the simple one. Open it and make the following modifications in the instantiation of SerialModemGateway:

- Change the cellphone number to a valid one.
- Change the COM port and the baud rate with the value you me memorized before.

The result is something like this:
new SerialModemGateway(“modem.com3″, “COM3″, 115200, “Sony Ericsson”, “W580″);

If you are looking for a SMS server, the class ReadMessages is a good beginning.

Limitation
A regular GSM modem is slow and can send about 6 messages per minute.

The SSerial2Mobile project is a library for Arduino, to stablish a serial connection with a mobile phone.
The project web site says that they could connect with a cheap Motorola C168i.

It costs around U$27,00 plus shipping on Ebay.
I bought an used one for U$14,00 (plus U$13,00 of the shipping)

It is pretty cheap compared to the EUR$77,00 GPRS Shield for Arduino from Libelium.

Looks like Jonathan Oxfer and his friend Marc Alexander have already implemented it. Check it out:

http://www.practicalarduino.com/news/id/171

I have previously discarded the option to integrate LaCrosse stations in my projet, due to its total cost. However, it is now back on the plan.
I still consider that this station is too expensive when compared to a DIY version, but the truth is that many people are currently using them. Altough, many are unhappy, due to the complex infrastructure needed to transmit data over the internet, as explained in the previous post.
By running a parallel project to integrate LaCrosse stations with Arduino, I could provide this “extensions” to these guys in a kind of partnership.
I would give them the ability to measure wind without their expensive computers and they would give me permission to use the anemometer measurements in my web-site.

Motivation
I have an on-line store to sell kitesurfing equipment and the most attractive page is the one that shows the wind speed for the main kitesurfing spots.
In addition, I have been playing with Arduino and remote device monitoring in the past few months just for fun and learning, so this project is pure fun and profit.

The problem
Weather monitoring is not an innovation. It is possible to buy a LaCrosse Weather Station for U$110,00 on Ebay. These stations are wireless and can be connected to a PC, which can easily transmit the wind speed through the web.
So, why do you want to build a new one from scratch?

For several reasons:

• Portability: A PC is too big to move and place on the top of a lamppost.
• Costs: A PC raises considerably (5+ times) the cost of the station.
• Connectivity: A regular internet connection is expensive and often unavailable in the most remote locations like lakes, rivers and beaches.

Alternative solutions

I have been studying some of the available alternatives, before starting to spend money on parts for this project. They can be divided three areas: the wind sensor, the cellphone interfacing and the Arduino interfacing.

The wind sensor

• I have spent a long time looking for an Anemometer Housing on the internet, without success. There are some cheap alternatives, but due to that size the shipping cost to Brazil is too expensive.
• Another thought was about buying a LaCrosse station and intercepting its RF signal with an Arduino RF receptor. Two disadvantages made me to put this idea on hold: 1. There is no documentation about the RF signal transmitted by the LaCrosse stations, so it would require reverse engineering, which is very complicated; 2. LaCrosse stations send their signals every 5 minutes, which is too slow for a $100 bucks sensor.
• There is a pretty good Anemometer sensor at Sparkfun web-site for U$69.95. It has an easy Arduino interfacing and a good coding library.  It’s Almost perfect for my purpose. Unfortunately, the shipping price is around U$40,00,  making a total of U$109,95, which can double if customs catch it.
• Do it yourself (DIY). It is possible to build one with around U$10,00 and some used parts. It is the cheapest option, but the disadvantage is obvious:  More labor.

By now I will focus on the last option.

Interfacing with the anemometer sensor
This problem might be easy or difficult to solve, depending on the sensor adopted.
The SparkFun and the DIY options can be easily interfaced with Arduino and a cable. The LaCrosse station would require a Arduino RF receptor and a lot of time reverse engineering its protocol.

Interfacing with the cellphone

That is the biggest problem, so far. There are too many variables involved like, usage of SMS or GPRS, the cheapest cellphone model, how connect arduino to the cellhpone pins and others.

SMS vs. GPRS: The data transmission over GPRS can be expensive, particularly in a country like Brazil where data plans are not so popular.
Sending data over SMS requires a complex SMS central receiver, built with another cellphone, integrated with a PC and internet. The data trasmission cost is cheaper compared to GPRS, but the required infrastructure is much more complex and expensive. It may be worth it, if there are several SMS clients communicating with the central receiver, but that is not my case.

At this moment I will try with the GPRS connection.

What is the best cellphone? The cheapest one!

A few years ago I took a look at my LowJack (GPS tracking system for Vehicles) and it had an old Motorola, which probably does not even support GPRS. After investigating a little more I found out that any cellphone with support for AT+ commands can be used for this.
After that I ended at sserial2mobile web page. It is an Arduino library to abstract the Motorola C168i At+ commands. This model can be found for around U$50,00 on Ebay, including the shipping to Brazil.

Alternatively, it would be very nice to have a GPRS shield easy to plug into Arduino. The cheapest one is not really cheap, it is again from SparkFun and its cost is around  U$100,00 plus shipping.

The solution
Currently the best solution seems to be the DIY anemometer sensor, an Arduino and a Motorola C168i.
I will be posting news about this project as soon as I have made some progress.