High resolution spectral estimation for radar in a pipe
MetadataShow full item record
In this project I have studied and evaluated several high resolution spectral estimation algorithms to improve accuracy and robustness in a FMCW radar tank gauge (RTG). The system parameters are taken from KONGSBERG's GLA-300 LNG-radar. I have in particular investigated how subspace based spectral estimation algorithms can improve liquid level estimation in the situation when the liquid surface is close to a reflector. A radar tracking algorithm have also been implemented to increase accuracy of measurements and follow the time variability of the measurements.The project includes data from simulation of a waveguide, as well as experiments using physical measurements from the actual RTG. All signal processing have been implemented in MATLAB. The simulation results show that one can not achieve measurement accuracy that satisfies the accuracy criterion of $\pm 5$mm of maximal estimation error with the proposed signal processing, but the accuracy in measurements are approximately doubled compared to the existing setup. Using a radar tracking in addition, the accuracy satisfies the accuracy criterion. The liquid level estimation using physical measurements confirms the results from simulation and a more stable liquid level estimation is achieved. Compared to the existing method, the high resolution methods are to a lesser extent affected by resolution problems close to reflectors.