Article Preview
TopIntroduction
Nowadays, more and more agricultural machinery mount sensors for collecting ðne raw data (temperature, humidity, GPS, speed, etc.) in the ðeld. These data are then exploited by decision-making algorithms to improve agricultural technical operations such as labour, spreading, monitoring, scouting, etc. These data are usually integrated in a Farm Management Information System (FMIS), which has been deðned as a “planned system for collecting, processing, storing, and disseminating data in the form needed to carry out a farm’s operations and functions” (Sørensen et al., 2010). One major functionality of a FMIS is ”machinery management”, which ”includes the details of equipment usage, the average cost per hour per unit. It also includes ñeet management and logistics” (Fountas et al., 2015a). In addition, we are witnessing a signiðcant development of autonomous vehicles making it possible to carry out increasingly complex agricultural operations for environmental and social purposes such as chemicals reduction and painful jobs and accident reduction.
In this context, as stated by (Fountas et al., 2015b), the role of the farmer will move towards decision-making in order to monitor and manage tasks of autonomous robot ñeet (RF). In particular, the scheduling and recovery activities of a RF (S-RF) is a crucial functionality that must be offered by an FMIS. As described in (Sørensen et al., 2010b), S-RF in agriculture is strongly characterized by uncertainties due to contextual agricultural data such as weather, machine performance, obstacles present in plots, etc. Therefore, maximizing the efficiency of the equipment when small autonomous robots are present is a key issue for smart agriculture (Sørensen et al., 2010b). Several works have been proposed to handle the scheduling of agricultural equipment at different spatial scales from plot to farm, but contrary to other industrial activities (Rossit et al., 2019), few works take into account the recovery phase of the scheduling step (called online scheduling in (Sørensen et al., 2010b)). The seminal paper (Sørensen et al., 2010b) lists the requirements of a ñeet management system for agriculture and proposes a conceptual framework based on sensor networks. Authors in (Gonzalez-de Soto et al., 2014) also propose such kind of sensor-based conceptual framework for ñeet management.