gps-wizard

TODO

Core Functionality

File Readers

Thoughts relating to the FIT reader:

• Add support for Timex Ironman tracks where speed and position are split into separate data records.
• Consider re-writing the reader to avoid the fitdecode library?
  • This could be a lot faster but will require a significant amount of development effort.

Thoughts relating to the GPX reader:

• Add support for track segments

File Writers

Thoughts relating to the GPX writer:

• Add support for GPX 1.1 and the Garmin TrackPointExtension V2

Consider adding the following writers:

• Consider adding an SBP writer

Processing

• Filtering - acceleration, satellites, HDOP, SDOP
• GP3S results - 2s, 10s, 100m, 250m, 500m, 1852m, 1/2 hour, 1 hour, alpha, distance + time

General Utilities

• COROS simulator
  • Implement improved run detection
• Charts
  • Speed charts - Line for estimated speed + scatter for positional speeds
  • Application of central limit theorem and “sampling distribution of the sample the mean”
• Format conversion
  • Most useful being to convert the variety of FIT files to GPX
  • Should allow SBP output where sat, COG and HDOP are present (e.g. COROS FIT)
• Smart reduction
  • Reduce tracks to minimal number of points for KML export.
  • Needs to be smart enough to avoid corners being cut, etc.
• Track collation
  • e.g. South West Coast Path walks combined into a single KML and / or GPX that provides an overall summary
• Locosys Tools
  • SBP splitter - split multiple days / sessions into separate files
  • SBP / SBN username updater

Investigations

• Locosys checks
  • Compare filtered and non-filtered speeds in SBN files
  • Check for correlations using scatter charts
    • Accuracy information
      • Sats vs HDOP
      • EHPE vs HDOP
      • SDOP vs HDOP
      • SDOP vs EHPE
      • SDOP vs VSDOP
      • EVPE vs VSDOP
      • Explore fixType from u-blox - e.g. dead reckoning during crashes
    • Unfiltered data
      • SOG vs USOG
      • COG vs UCOG
    • Coloured scatter plots
      • HDOP vs position / speed / speed error
    • Horizontal vs vertical
      • Speed vs climb rate
      • SDOP vs VSDOP - expect vsdop to be worse
      • ehpe vs evpe (SiRF) - expect evpe to be worse
  • Search for Doppler spikes - e.g. 50 knots+ in historic GT-11 / GT-31 files
• u-blox checks
  • Equivalent of SiRF checks, where possible / applicable
  • hacc vs vacc - expect vacc to be worse
  • cacc
  • “tunnel mode” - clever detection?
• Chart types

  • Linear Correlation showing linear trends and R².

  • 2D scatter plots using speed + COG and using colors for SDOP / sAcc, etc.

• Sony
  • repeated speeds prior to crash then zero
• Check if Estimated Horizontal Position Error (SiRF) or Horizontal accuracy estimate (u-blox) are useful to identify spikes
  • https://logiqx.github.io/gps-wizard/formats/sbn.html
  • https://logiqx.github.io/gps-wizard/formats/ubx.html
• Check fix type during crash - e.g. use of dead reckoning on u-blox
• Determine time shifts using RMSE
  • Estimated speed vs positional speed - same GPS device
  • Estimated speeds - different GPS devices
• 5 Hz + 10 Hz acceleration filter
  • Create project that implements proposal on Seabreeze
  • n.b. Stuey’s WK track had a 3.7 m/s² spike which slips through existing filters

Articles

• Move doppler-speed-resolution from gps-guides to gps-wizard
• Filtering
  • Acceleration
  • Crash detection
• Variations in reported results
  • Across devices
  • Doppler vs Positional
• Chipset summaries - page per chip
  • Note Sony UDS

Check Links

https://validator.w3.org/checklink

This site is open source. Improve this page.