Real-Time INS/GNSS Integration of Low-Cost IMU and Multi-Element Antenna Array Receiver

Kurzfassung

Safety-critical applications such as autonomous cars and small UAVs have gained popularity over the last years. These require a good knowledge of their position and attitude at all times. Recent research has shown that multi-antenna GNSS receivers can provide an accurate attitude estimation in addition to the position. However, they are sensitive to interference and can not output a solution at high rates. Thus, these systems can benefit from the use of an inexpensive inertial measurement unit (IMU), as it provides measurements that are always available. In this work, an INS/GNSS integrated system has been implemented, which is able to take advantage of the multi-antenna receiver capabilities in order to obtain a robust and accurate vehicle attitude estimation. The design of the system architecture is presented, along with the different algorithms used for the navigation solution computation. In addition, several mathematical models have been developed for the designed Kalman Filter used in the integration. This Filter integrates position in a loosely-coupled way while the attitude is integrated from the one computed by the GNSS receiver or by directly using the direction of arrival (DOA) of each visible satellite, significantly improving the attitude accuracy. Additional layers of protection have been included into the Filter to prevent faulty masurements from corrupting the Final solution. The designed algorithms have been implemented in a C++ framework able to support both online and real-time operation modes. The algorithms and code have been validated and tested using a small platform with a low-cost IMU and GALANT, the GNSS receiver with multi-antenna support developed at DLR, through a measurement campaign carried out with the DLR MessBus