Embedded Sensor Fusion System for Unmanned Vehicle Navigation

This paper presents the hardware design as well as the implemented sensor fusion algorithms used for navigation purposes in unmanned systems. A design concept is illustrated that has advantages when limited space and changing demands of different unmanned systems are of particular interest. The embedded system demonstrated in this work consists of a Field Programmable gate array (FPGA) board in conjunction with a digital signal processor (DSP) board. By exploiting the individual properties of FPGAs and DSPs, a highly efficient system can be achieved, which is able to process more computationally complex sensor fusion algorithms like iterative sigma point Kalman filters in real time as well as more complex sensor signal processing algorithms used for pulse radar target detection. To validate the system's performance, experimental results are presented by using an unmanned aerial vehicle (UAV).