EcoPDA

From CENS Urban Sensing

Contents 1 Current Issues 2 Discussions 3 Problem Statement 3.1 Stakeholders, Users, and Their Needs 4 Vision of the Solution 4.1 Timeline 4.2 List of Features to be Developed 4.3 List of Features Considered, But Not Being Developed at this Time 5 Tasks (Nokia N80)

Current Issues

Discussions

Ecopda Sensor Log Format

Ecopda Barcode

Ecopda Possibly Low Hanging Fruit

Problem Statement

Biodiversity and ecological research often depend on paper-based surveys that have a prescribed protocol for recording information. There are several problems with the current methodology:

• Missing data -- For various reasons, the data yield often falls below 100%. Data can be dropped at the time of collection (out in the field) or during transcription from paper to electronic.

• Incorrect data -- At the time of collection, numerous observations need to be recorded by the data gatherer. Errors can be introduced when the data gather is asked to note species, note location (trap #18234), make readings (temperature, humidity), make counts (number of captured specimens, etc.). During transcription, handwritten notes may be incorrectly interpreted.

• Wrong data -- Protocols often require that the observer conditionally record certain kinds of information. An error here may result in the wrong data being gathered.

• Scalability -- The current method may require a lot of time for each site, and so large scale surveys require a lot of people, making scalability an issue.

Stakeholders, Users, and Their Needs

Folks who can answer, "What's the problem, and what do you need?"

• Jeff G. - Needs...
• Deborah - Needs...
• Jeff B. - Needs...
• Mark - Needs...

• Eric Graham
  • Affordable hardware.
  • Ruggedized hardware.
  • New protocols to be easily scriptable.

• [Conservation International]
  • Affordable hardware.
  • Ruggedized hardware.
  • New protocols to be easily scriptable.

Vision of the Solution

The approach is to use a PDA (Nokia 770) to:

• Electronically record observations.
  • This removes the transcription stage, which may introduce errors or miss data.

• Guide and prompt users through a protocol.
  • This helps higher data yield, and assists in gathering the correct data.

• Verify data at the time of collection.
  • This helps to reduce incorrect input by doing range or other simple validity checks.

• Automate or Augment some observations, via GPS tagging, barcode reading, voice recording, and imaging.
  • This helps scalability by speeding certain observations.

Timeline

• Prototype by Sept. 1, 2006
• Demonstratable by Nov. 1, 2006

List of Features to be Developed

• Ruggedized PDA

• SW
  • Validity Checks and QA/QC
  • Easy Upload
  • Alerts (What kind of alerts?)
  • Automated backup (What is this exactly?)

• Scripting Wizard
  • Authoring Tool

• Data Acquisition
  • Text
  • Voice Annotations
  • GPS
  • Images
  • Bar codes or visual tags

• Data Transport and Management
  • Integrate with ESP

List of Features Considered, But Not Being Developed at this Time

• We are not assuming or using connectivity out in the field.
  • Data will be uploaded when the user specifically says to.

• Display or play back
• Image contrast
• Noisy Acoustics
• Null/Binary text (What's this?)
• User -> tag or replace annotate data (What's this?)
• OCR
• Speech Recognition
• Ruggedized PDAs
• Field guide, other sensors

Tasks (Nokia N80)

• Overall Coordination
  • Core funcationality / usage: Jeff G.
  • Implementation tasks: Ben G.

• Ruggedization plan (Lead: Jeff G.)
  • HW
    • Gather requirements from Bill/Dave Caron
  • SW
    • Validity Checks
    • Automated upload
    • Alerts
    • QA/QC
    • Automated backup

• Scripting Wizard (Lead: Chris M.)
  • GUI
    • Web browser on device
      • 08/02/06 Comment: When directed to connect to 127.0.0.1 or localhost, both the default web browser and Opera browser attempt to connect to the Gateway before connecting to itself. If there's a Gateway available, then it works. If there's no Gateway, then it doesn't work. :( We're looking into the source code of the default browser to see if it's the browser's problem or something lower level.
    • Web server on device
      • 08/02/06 Comment: Apache httpd is in the process of being ported to S60 3rd edition, which is what the N80s run. We are looking into adopting a lighter-wieght, python-based web server, since our needs for the server are rather modest--we just need to be able to invoke python on the server side to retrieve images and other sensor data.
    • Implement butterfly protocol - Ver 0. by 7/5/06
    • Add features for Ver 1.0
    • Barcode vs. other ID
  • Backend
    • Authoring tool/interactive webpages

• Data Acquisition (Leads: Andrew P. and Ben G.)
  • Text
  • Voice Annotations
  • Images
  • Bar codes or other IDs

• Data Transport and Management (Lead: Andrew P.)
  • Integrate with ESP

• Additional Protocols (Leads: Eric G, Eric Y.)
  • Identify an other existing CENS protocol that could benefit from EcoPDA.