GNSS vs Dead Reckoning accuracy

[jg_spitfire]

Hi, which of these solutions have better accuracy for positioning? and why? do you have any recommended documents to learn more about it? thanks

[Tesla DeLorean]

Unaided GNSS is good to a few metres, best to use RTK for any degree of accuracy and repeatability. Establishes location vs governmental references.

INS/IMU provides secondary perspective, can help identify smaller motions GNSS might see as measurement noise.

With sensors and wheel tick type input you can extend usage into tunnels and parking structures where direct measurements of GNSS signals are impractical/unhelpful. You're measuring distance to sub-mm levels, bouncing off things adding metres, or tens of metres, increase the circle/sphere of uncertainty.

Accelerometers and Gyros tend to drift, this can vary over time/temperature, GNSS can help calibrate, quantify and correct.

Compasses and magnetometers are affected by metal and ore deposits, may have local/situational issues, ie driving by a metal shed, pylons, etc.

https://www.amazon.com/Global-Positioning-System-Inertial-Navigation/dp/007022045X

https://www.amazon.com/Aided-Navigation-High-Rate-Sensors/dp/0071493298