Experimental Study of Two-channel UWB-OFDM Radar for Indoor Navigation with INS Integration

There is a great need to develop alternative navigation sensors for situations where GPS is not available. In our previous work, we developed a set of navigation algorithms for a novel indoor/outdoor navigation sensor, a UWB-OFDM radar. The previously developed experimental prototype system was mounted on an indoor vehicle and used in combination with an HG1700 tactical grade INS to provide a position and velocity solution. The algorithms extracted stationary radar features from the environment and used the relative motion between the features and the platform to correct the inertial solution errors. This test experimentally validated the use of a UWB radar as a navigation sensor. However, due to technical limitations the radar was only able to operate in single channel mode. With a single channel available, the radar would intermittently stop tracking features and therefore stop providing error corrections. The accuracy of the error corrections applied was also greatly limited by the intermittently available features. In this paper, the radar has been upgraded to operate in two-channel mode to allow for more robust and reliable navigation. In addition, the transmitted OFDM symbol uses a random subcarrier modulation. This spreads out the transmitted energy to the full 1GHz spectrum and allows the waveform to be a discrete approximation to a noise waveform, enabling the radar to operate in LPI/LPD and jamming scenarios. Finally, a truth reference system is developed to allow for better analysis of navigation solution errors. The previous single-channel navigation experiment is repeated, moving the platform through a narrow passage containing metallic reflective surfaces. In order to have a GPS truth reference available, the experiment was performed outdoors next to a building. The range and Doppler measurements for each tracked reflector are used as updates to the navigation filter, implemented as a modified EKF. The EKF combines the radar measurements with INS measurements to compute an optimal estimate of the platform position error. The navigation solution is compared for two scenarios: INS-only position and INS position corrected by the EKF error estimate. These two navigation solutions are compared to the true trajectory. In the end, this paper demonstrates that a UWB-OFDM radar combined with an inertial system can be a viable alternative navigation sensor in GPS-denied environments.