Which are the main challenges faced by the project?

Continued population growth requires an increase in agricultural production, which is threatened by many challenges such as climate change, labor shortages and productivity performance costs. In recent years, the use of Smart Farming Technologies (SFTs) is increasing rapidly and appears to be a reliable solution to the above challenges. SFTs include Farm Management Information Systems (FMIS), Precision Agriculture (PA) Systems, agricultural automation and robotic applications that (a) increase yield, (b) improve product quality, (c) make rational and efficient use of inputs, (d) reduce energy consumption and (e) protect soil and water resources.

While SFTs for FMIS, PA methods, and automation operations have been developed commercially, only a few cases of robotic systems are at the commercial level. However, robotics is making great strides as autonomous Unmanned Aerial Vehicles (UAVs) and Unmanned Ground Vehicles (UGVs) are increasingly incorporating technological equipment to monitor crops and implement agricultural practices. Their use can increase productivity, profitability, work safety and environmental sustainability.

But even commercially available UAVs and UGVs for agricultural use work as stand-alone solutions rather than as a combined system where UAVs and UGVs help each other.

Main project goal

The main goal of AgroSys is the development of an autonomous collaborative robotic system for precision viticulture integrating electric UAVs and UGVs and small-scale solar charging stations

Why to implement this system on vineyards?

  • vineyards have a special feature: it is necessary to simultaneously monitor the upper and lateral growth of the vine-stocks in order to fully diagnose the state of the vineyard
  • therefore, the combination of UAV and UGV can provide the solution for monitoring the vines by designing 3D maps of vineyard vegetation characteristics, based on aerial and ground data
  • AUA’s vineyard (35 ha) will be monitored by flights (UAV) and ground observation (UGV) in selected phenological stages of the vine (bud burst, bloom, fruit set, cluster closure, véraison and fruit maturity (harvest)).

UAVs, UGVs and their sensors

The system consists of a heterogeneous fleet of three levels:

  • operational environment [ground-air]
  • power requirements [small – medium – large robotic vehicle]
  • mission [awareness, mapping – sampling – global scan]

Sensors mounted on the UAVs

  • multispectral sensor – calculation of vegetation indices
  • thermal camera (FLIR Vue Pro) – determination of water stress – irrigation

Sensors mounted on the UGVs

  • hyperspectral camera (Cubert S185SE) for the lateral vine growth
  • thermal camera same as UAVs – water stress
  • laser scanner (Velodyne puck Hi-Res)
  • Micro Electromechanical Systems (MEMS)

Other systems

  • weather station with wireless network
  • central server that will host FMIS

Information System Characteristics

  • autonomy to achieve a goal or task without any human intervention
  • independent navigation of the UGV/UAV fleet
  • Robotic Operating System (ROS)
  • UGVs and UAVs capable of autonomously exploring, mapping and navigating an unknown environment in an efficient and safe way
  • full development of optimal path planning algorithms to minimize unnecessary movements of UGVs and UAVs
  • specially designed controller to determine the optimal operating point of UGVs and UAVs based on computational intelligence techniques (Fuzzy Cognitive Maps)
  • monitor the UAV/UGV battery discharge rate and decide their speed

Charging of electric UGVs & UAVs

  • charging electric robotic vehicles through autonomous charging stations with photovoltaic arrays and batteries
  • microgrid topology (small scale, low voltage, distributed generation, isolated areas)
  • promotion of smart grids and decentralized energy production
  • Charging Stations: – advanced energy management and control systems for optimal energy management – appropriate geographical dispersion based on the optimal routes of UGVs and UAVs
  • Vehicle – Charging Stations communication (exchange of information about their stored energy and the charge level of the batteries)
  • maintenance of a history report for the attributed energy and use of algorithms for forecasting the energy produced based on the Machine Learning theory

Robotic arms for charging UAVs & UGVs

  • the system can work continuously (24/7)
  • the design is based on commercial solutions for electric car charging stations
  • implementation of 2 robotic arms at each station within the vineyard: the first will connect the charger to the UAVs and second to the UGVs)
  • installation of optical sensors to detect the exact position of the vehicle socket/plug
  • each station will have 2 parking spaces for UGVs and UAVs

Smart Operating System

  • universal supervision
  • communication between the charging stations and the autonomous vehicles
  • information transfer (e.g., battery charge level of vehicles and stations)
  • decision: which station will charge the respective vehicle and at what time?
  • based on Game Theory: competitive game between players
  • use of Nash’s equilibrium (available information of each player – no leader-follower relationship between players)
  • non-cooperative game: dynamically variable game due to changes in weather conditions and the charge level of each player’s storage capacity (an important innovation in agricultural robotics)