Ice scream, you scream...

Being a native California driver I rarely encounter icy road conditions unless I'm on my way to Tahoe to go snowboarding or skiing. I remember the day we had snow in the Bay Area and everyone ran outside to experience the phenomenon (although it only lasted on the ground for a few moments before melting away). In many parts of the world icy road conditions prevail and the ability to easily monitor and detect the danger is not easy due to the harsh environment. However, the idea of pavement condition monitoring would save many a spinning car and screaming driver from sliding on the ice into the side of the road.

Pavement maintenance is vital for travel safety. By using wireless sensor networks to monitor pavement temperature and moisture presence, icy road conditions can be detected. It is essential to provide warnings of dangerous traffic conditions in real-time. In a study done at University of Oklahoma, researchers determined to investigate a densely distributed sensor network and classify pavement conditions into certain categories - 1) dry 2) wet and 3) frozen.

This project was deployed with the MICA2 Motes from Crossbow while integrating them with various 3rd party sensing devices using the MDA100 prototyping board. The ability to integrate 'alien' sensors to the Mote platform gave the researchers the flexibility they needed to complete the task at hand. The sensors chosen to provide the data included a thermistor to gather temperature readings, a leaf sensor to detect the conductivity of a wet pavement to detect the existence of free moisture and an infrared sensor to detect ice by emitting a near infrared light that is reflected by the ice and detected by the infrared receiver (water is transparent to the receiver).

An integrated sensor and road button structure housed the 3 sensors as shown in the figure above. The top surface of the sensor road button contained the moisture and infrared sensors with the thermistor at the bottom. Due to the low power consumption of the sensors used, these devices were powered by the MICA2 Mote platform. The Mote platform was placed into a protective watertight aluminum casing with upgraded antenna doubling the Motes transmitting range.

When collecting readings from the sensors, the Mote transformed them into digitized data, sent them to the radio and waited until all data was sent before switching to sleep mode. In detail, the processor received sensor readings from the embedded 10-bit Analog-to-Digital Converter (ADC). If the data was taken correctly, the onboard light emitting diodes (or LEDs) lit up to signal the proper functioning of the mote. Analog to digital conversion was performed on the readings, after which the data was integrated into the packet to be transmitted. The default packet format was slightly modified to fit the size and format of the data to be transmitted. The packet was then sent to the radio and transmitted over the network until it reached the base station. The base station was connected to a laptop through a serial port. The data was then collected using a LabVIEW graphical user interface (GUI) developed for this project. Raw data from the serial port was collected, deciphered and displayed by the GUI.

Using the MICA2 Motes to monitor the pavement conditions is a unique application; therefore, the aforementioned features (directly applying time synchronization and embedding a pattern classification algorithm) further distinguish this study from existing research that utilizes Motes in real-world applications. A series of laboratory tests was conducted at the Asphalt Laboratory at the University of Oklahoma using an environmental chamber to study the effect of temperature and moisture on the sensors (and later, the motes). The environmental chamber was used to produce well-controlled temperature and humidity variations. The sensor-road button unit was tested to (1) test the full functionality when all the sensors were combined together, and (2) collect data to aid refinement and further development of the proposed ice detection algorithm proposed in this application. The entire lab test was completed in a four-hour time frame. Note that weather changes in reality could be much slower than this testing rate; thus such a test could be more stringent than a real-world situation.

A series of outdoor tests were conducted as well paying special attention to the packaging and survivability of fragile analog sensors in harsh roadway conditions and how they will be utilized in other applications of intelligent transportation systems (ITS) as well as structural health monitoring. These methods allowed the Mote wireless sensor network to be easily installed and provided a robust solution to environmental factors such as wind and rain. Imagine a day when roads are 'smart', when you are told exactly what conditions to expect before you encounter a patch of ice. It is this concept and future that we envision with the Mote platforms - a smarter safer future that we all can scream for!