GPS-Aided Odometry Navigation for IAVs: An Assessment of Integration Topologies and Odometry Mounting Configurations

Nowadays, navigation of agricultural vehicles is almost exclusively dominated by Global Navigation Satellite Systems (GNSSs), in particular, the Global Positioning System (GPS). Nevertheless, when precise positioning requirements are imposed, which are typical of the so-called “Precision Agriculture” (PA), such technology becomes financially impracticable to small and medium-sized farmers, which correspond to more than 80% of farmers in Brazil (family farming). This is mainly due to the high cost of the employed GPS receivers, which are multi-frequency and based on advanced differential and relative techniques, and also to the frequent unavailability of internet access in the field as a mean to provide positioning corrections. Odometry, in turn, can serve as a low-cost reference (and complementary) navigation technology to GPS, which, unlike the latter, presents good short-term positioning precision, despite the recognized error accumulation in the long term. The objective of this work, hence, is to propose, test and evaluate the performance of a low-cost Single-Frequency (SF) GPS-aided odometry navigation solution, aiming at improving positioning accuracy for Intelligent Agricultural Vehicles (IAVs), As the main contribution of this work, two of the main integration techniques employed in sensory fusion applications are compared, namely, the Loosely Coupled (LC) and Tightly Coupled (TC) topologies, which are implemented based on an Extended Kalman Filter (EKF) framework. As an additional contribution of the work, we evaluate the performance degradation of the proposed integrated solutions when odometry is used on the IAV's driven wheels, instead of the non-driven ones. Experimental results shows the superiority of TC solution, especially when odometry is employed on the IAV's driven wheels.