Agricultural robot fully autonomous independent of the GPS signal.

Autonomous navigation through agronomic terrains represents an enormous challenge due to the great variability of the operating conditions. The available technologies have traditionally trusted on GPS-based autonomous guidance and many developments have been achieved in this area so far. However, the GPS-based autonomous guidance presents important problems when it is applied in agronomic environments with dense or tall vegetation, such as speciality crops canopies or inside greenhouses. The limitations of GPS-based solutions in agricultural environments derivates mainly from the uncertain or discontinued reception of the GPS signal in some locations, from multipath errors and from the lack of detection of objects under complex, dynamic and uncontrolled scenarios, such as the presence of structures, unattended canopies, farmers, tools, branches, boxes, etc. The present invention provides a method, a system and a complete robot for the autonomous navigation through rows of plants or trees grown in vertical trellises, such as vineyards or orchard-type plantations. This new system is based on the combined use of local perception sensors. Three sensing devices are incorporated into the robot’s front. Two of them allow the system to define and navigate a bi-dimensional grid (Image 1). This makes the robot capable of operating without human interaction. The system allows the robot to keep on moving from one row to another, also to avoid obstacles that may appear on the path (Image 2). The system was developed by a Spanish university research team specialised in robotics and automation of agricultural equipment. It is the result of an H2020 / Fast Track to Innovation granted project. The navigation system has been implemented in a robot whose main objective is collecting crop and environmental data from the plantations to assess plant water status. A fully functional prototype has been tested under real conditions in vineyards placed in Spain and Portugal (Image 3). The input signal and the actual path travelled by the robot can be compared (Image 4). The prototype was able to navigate the crops indistinctly during day and night while collecting data related to hydric and thermal stress. During the trials, it operated for 3 days with no battery recharging, so it may well work non-stop during 12-hour sessions, and under ambient temperatures over 30ºC. Further agricultural applications are possible, i.e. the automation of crop harvesting, the provision of agricultural formulations (plant protection, biostimulants, plant nutrition), as well as the clearance or maintenance of plantations, by integrating the navigation system in other autonomous or semiautonomous agricultural equipment, such as medium-size orchard harvesters, weed removal machinery or plant sprayers. The technology was specifically developed for agricultural environments characterised by tall rows of plants, such as vineyards. Therefore, although not tested yet, it may be applicable to other similarly-shaped working environments, such big logistic areas or warehouses made up of relatively tall aisles. The prototype robot devised to capture hydric and thermal stress data is a turnkey solution, which is ready for a technology transfer through a license agreement. Anyway, the research team will also provide technical cooperation if required by the licensee. For the development of further applications (crop harvesting, plantation maintenance, plant care, etc), the research team offers a technical cooperation agreement. The type of partner sought in any case is an agricultural automation machinery/equipment developer or constructor, a field robotics builder or a farm machinery manufacturer. A European patent application was submitted in July 2020 to protect the whole method, system and robot. Public demonstrations of the prototype have been conducted and media material has been released to show the technology working in real operations.