Hubert Schumacher

Most of the times when we talk about sensors, we mean technical devices, measuring some relevant metrics and transmitting them to some data collection or management entity.


Sensing data

Sensors are more or less dumb and most of the time very specialized. Not to minimize the importance of technical sensors in the preparedness context, but a flood sensor may only tell you that the water level of a river exceeded a predefined threshold. It won't tell you that the floods damaged a road making it unusable for rescue operations.

Using his mobile phone, a person might notify some central crisis management entity about the damaged road. But this solution of sensing and assessing complex events, lets another problem pop up: How to consolidate the transmitted data, verbally communicated in a non structured way by a person to a crisis operator? Beyond the contextual information and knowledge the operator needs in this case, this communication isn't very efficient and prone to errors.


Consolidating data

The big advantage of the communication with the technical sensor is that it only understands one question "What is the water level?" and gives precise answers like "2,11 m (above some defined reference)". The transmission of this data can be stored locally and transmitted at regular interval or on demand, or can be automatically pushed to a central location, by mobile or fixed telecommunication means.

Consolidating human data is somewhat more complex. You can imagine that a trained call center operator will transcode human provided data into structured data usable by crisis management systems. But you can also imagine that a local application on the human sensors' mobile phone or PDA pops up and ask the relevant questions, e.g. "Can you confirm that the bridge is destroyed? YES/NO", "How many vaccines have you left in the Flu Pandemic Vaccination Center? Enter the number".

Of course there are some prerequisites and constraints to this approach:

• Central management of the application's dialogs (questions and answer)
• Users' awareness of the data sensing application. Minimal training might be required
  
• Users' ability to use their own mobile phone or PDA beyond pure telephony
• Device management and User-device associations: which user has which device, which devices have J2ME support, what screen resolution and communication facilities are available

The usability of the transmitted data could or should be augmented by time and location and presence data provided by the users device and it's application.

Collecting, structuring, parsing and validating user provided data is only part of the story. Consolidating them in a useful way with technical sensor data and using them for procedure and alarm triggers is the most challenging part.

Part of these features are already reality in my company's solution (AlarmTILT), the remaining parts are under development. Stay tuned ...


Assessing data

A good model of the crisis environment is of course the prerequisite for human sensors to provide added value. Detailed planning of the applications and training of the users is the key to successfully exploit the valuable human sensor data. Just recall Tsunami or the London Bombing in your memory to understand that a lot of highly useful "field" information is available but is left unused most of the time.