Ultra-tightly coupled GLIO aided by GNSS multipath correction
Date:
Main challenge of the tightly-coupled GNSS/IMU/LiDAR integration
The GNSS is almost everywhere but the raw GNSS measurements are hard to use in urban canyons due to the strong GNSS signal reflections and blockage. The the tightly-coupled GNSS/IMU/LiDAR integration is promising but is essentially limited by the quality of the raw GNSS measurements, such as the code (pseudorange) and phase measurements. I was attracted by the carrier-phase for a while due to its high accuracy in positioning or even the velocity estimation. In particular, the observation model of the carrier-phase measurements can be denoted as below:
\[\varphi_{r,t}^{s}=r_{r,t}^{s}+c(\delta_{r,t}-\delta_{r,t}^{s}) - I_{r,t}^{s}+T_{r,t}^{s} + \lambda *N_{r,t}^{s}+\epsilon_{r,t}^{s}\]where the $\varphi_{r,t}^{s}$ denotes the carrier-phase measurement (in the unit of meters) received from the satellite $s$ by the receiver $r$. The $r_{r}^{s}$ denotes the real range between the GNSS receiver and satellite. The $c$ denotes the speed of light. The $\delta_{r,t}$ and $\delta_{r,t}^{s}$ denotes the receiver and satellite clock bias. The $I_{r,t}^{s}$ and $T_{r,t}^{s}$ denotes the innospheric and tropospheric errors. The $N_{r,t}^{s}$ denotes the unknown integer ambiguity which need to be reliably resolved before its usage in absolute positioning. The $\epsilon_{r,t}^{s}$ denotes the Gaussian noise associated with the carrier-phase measurements.
References
- [1] Using latex to input equations in Overleaf Link.