Real-Time Implementation of Relative Positioning Approaches Using Low-Cost Single-Frequency GPS Receivers and Raspberry Pi Platform for Agriculture Applications

Autonomous agricultural vehicles rely on accurate real-time positioning information to perform precise farming operations, including seeding, fertilizing and harvesting. Although many positioning solutions exist on the market, their cost is out of many farmers' budgets. Relative positioning can be used for accurate positioning purposes in agriculture applications. Despite the fact that several research attempts have addressed the issue of relative positioning, no agriculture-specific study appears to exist in the literature at the time of writing. In this paper, the opportunity of using real-time GPS-based relative positioning for agriculture applications instead of the typical absolute positioning solutions has been investigated. To this end, we propose to implement and evaluate different relative positioning approaches, namely absolute position differencing, pseudorange single differencing as well as pseudorange double differencing (PDD). Real-world experiments, including static and dynamic scenarios, have been conducted using low-cost single-frequency GPS receivers in an open sky environment and an agricultural field. The obtained results show the superiority of the PDD approach with about 76% of estimation errors less than 1 m in the open sky environment and more than 97% of estimation errors less than 3 m in the agricultural field, representing a considerable accuracy enhancement compared to GPS standalone positioning.