Towards a Reliable Monitoring Robot for Mountain Vineyards Filipe Neves dos Santos * , Heber Sobreira * , Daniel Campos * , Raul Morais † , Ant´ onio Paulo Moreira * and Olga Contente ‡ * INESC TEC - INESC Technology and Science (formerly INESC Porto) and Faculty of Engineering, University of Porto , Porto, Portugal Email: {fbsantos, heber.m.sobreira, daniel.f.campos, antonio.p.moreira }@inesctec.pt † INESC TEC - INESC Technology and Science (formerly INESC Porto) and Universidade de Tr´ as-os-Montes e Alto Douro, UTAD, Vila Real, Portugal Email: rmorais@utad.pt ‡ Instituto Politcnico de Viseu, ESTGV, Viseu, Portugal Email: ocont@estv.ipv.pt Abstract—Crop monitoring and harvesting by ground robots on mountain vineyards is an intrinsically complex challenge, due to two main reasons: harsh conditions of the terrain and reduced time availability and unstable localization accuracy of the GPS system. In this paper is presented a cost effective robot that can be used on these mountain vineyards for crop monitoring tasks. Also it is explored a natural vineyard feature as the input of a standard 2D simultaneous localization and mapping approach (SLAM) for feature-based map extraction. In order to be possible to evaluate these natural features for mapping and localization purposes, a virtual scenario under ROS/Gazebo has been built and described. A low cost artificial landmark and an hybrid SLAM is proposed to increase the localization accuracy, robustness and redundancy on these mountain vineyards. The obtained results, on the simulation framework, validates the use of a localization system based on natural mountain vineyard features. I. I NTRODUCTION Crop monitoring and harvesting by robots remains a com- plex challenge, particularly due its low efficiency, accuracy, and robustness on sensing, perception and interpretation of the agricultural environment [1]. The strategic European research agenda for robotics [8] states that robots can improve agriculture efficiency and com- petitiveness. But there are still very few available commercial robots for agricultural applications [1]. Fig. 1. A typical Mountain Terraced Vineyard in the Douro region of Portugal. In Europe space we can identify two ongoing research and development projects to deploy monitoring robots on flat vineyards: the VineRobot [16], and Vinbot [17]. However, there are other kind of vineyards that are not built on flat terrains but on mountain hills which is complex environment for the machinery and for the robotic algorithms (such as lo- calization, mapping and path planning). These called mountain vineyards exist in Portugal in the Douro region - an UNESCO heritage place - Fig. 1, and in another regions of five European countries. As these crops are not built on flat terrain but in steep hills, the robotic challenge becomes even more complex because: • the GPS system availability and accuracy are largely reduced due to the signal blockage or multi-reflection; • the dead-reckoning systems (for example odometry and inertial measurement systems) accuracy is drasti- cally reduced due to the harsh conditions of the terrain; and • the path planning and control system requires an accu- rate map of the vineyard and an accurate information about the robot posture (localization and attitude) for a safe robot motion planning and control. This paper deals with localization and mapping on moun- tain vineyards problem which is constrained by the GPS availability. In other publications we deal with other problems, such as path planning and control on steep hills terrains. In these vineyards, the GPS system availability and accu- racy is largely reduced due to the reduced number of satellites. This number of satellites in view is limited by the high elevation mask imposed by the hills. The tool available at http://satpredictor.navcomtech.com/ can estimate the number of satellites on the receiver field of view for a specific coor- dinates and elevation mask. Considering mountain vineyard at 7.55250734W 41.1756008N (at R´ egua city) with an elevation mask of 35 degrees, it is verified that the number of satellites is very week and sometimes not enough for the position estimation, as reported in Fig. 2. This will reflect in a high value for the dilution of precision (DOP) during a large amount of time. Besides these problems a GPS receiver is always exposed to spoofing attacks which can be a safety issue. In order to get a full time operational, efficient and safe robot is required to 2015 IEEE International Conference on Autonomous Robot Systems and Competitions 978-1-4673-6991-6/15 $31.00 © 2015 IEEE DOI 10.1109/ICARSC.2015.21 37