=Paper=
{{Paper
|id=Vol-2030/HAICTA_2017_paper42
|storemode=property
|title=Agent Based Software Tool for Efficient Agricultural Logistics
|pdfUrl=https://ceur-ws.org/Vol-2030/HAICTA_2017_paper42.pdf
|volume=Vol-2030
|authors=Dimitrios Katikaridis,Dimitrios Bechtsis,Konstantinos Liakos,Dimitrios Vlachos,Dionysis Bochtis
|dblpUrl=https://dblp.org/rec/conf/haicta/KatikaridisBLVB17
}}
==Agent Based Software Tool for Efficient Agricultural Logistics==
https://ceur-ws.org/Vol-2030/HAICTA_2017_paper42.pdf
A Software Tool for Efficient Agricultural Logistics
Dimitrios Katikaridis1, Dimitrios Bechtsis2, Ioannis Menexes1, Konstantinos Liakos1,
Dimitrios Vlachos3, Dionysis Bochtis1
1
Institute for Bio-economy and Agri-technology (IBO), Centre for Research & Technology-
Hellas (CERTH), 6th km Charilaou-Thermi Rd, GR 57001 Thermi, Thessaloniki, Greece, e-
mail: dkatikaridis@ireteth.certh.gr
2
Department of Automation Engineering, Alexander Technological Educational Institute of
Thessaloniki, Greece
3
Aristotle University of Thessaloniki, Department of Mechanical Engineering, Greece
Abstract. The Internet of Things (IoT) revolution enables the integration of
advanced automation techniques to the intra-logistic agricultural activities.
Sensors and actuators seamlessly communicate in order to accomplish complex
everyday tasks. From a logistics perspective, Automated Guided Vehicles
(AGVs) pave the way to sustainable agriculture; working 24/7 with minimum
labor costs and high safety standards while optimizing energy consumption
and providing effective and efficient solutions for everyday activities. This
article proposes a software tool for daily field level agricultural activities. The
tool provides a cyber-physical interface to the physical entities (field
boundaries, loads and vehicles) located at field level. Map services are
integrated at our tool in order to provide the field layout, the absolute
coordinates of the entities and a real time monitoring tool. Sophisticated
algorithms can be used for navigation at simulation level while the absolute
coordinates are sent to the vehicle for the physical navigation.
Keywords: Software tool, Automated Guided Vehicles, Navigation
mechanism, Simulation tool, Agricultural Logistics.
1 Introduction
In this article, we develop a software tool that allows farmers to monitor agricultural
fields’ state and to effectively navigate an autonomous Inter-Agricultural Logistics
(IAL) system. The provided software tool provides a cyber-physical interface to
locate physical entities (e.g. field boundaries, loads and vehicles) and simulate a
desired navigation pathway for the actual tractor to accomplish specific tasks.
The Internet of Things enables integration of sophisticated software applications
and advanced automations to the cloud computing interface (Sarangi, 2016), hence
unleashing sustainable growth opportunities for the agricultural sector (Ruiz-Canales
and Ferrández-Villena, 2015). More specifically, Automated Guided Vehicles
(AGVs) with embedded sensors and actuators are used in IAL to inspect soil quality,
provide weed control by spaying pesticides, fertilize the field or automatically load
and transfer harvested goods at pick-up points for transportation. Agricultural AGVs
363
are usually free range vehicles and navigate on the field with the either individual or
combined use of optical devices like cameras and LiDAR sensors, absolute Global
Positioning System (GPS) enabled controllers (Gomez-gil et al., 2011), inertial
guidance sensors like accelerometers and gyroscopes, and teleoperation techniques.
Despite the advances in robotic solutions for agricultural field operations (Bechar
and Vigneault, 2016), their actual application is limited (Xiang, 2014) due to the
increased modelling requirements’ complexity stemming from the natural
components of the unstructured and dynamic fields’ environment (Bechar and
Vigneault, 2017). Such components include varied shape, size, fruits which are
characterized by high variability that affects robot behavior, many of which cannot
be determined a-priori. Fully-autonomous AGV system could tackle the
abovementioned challenge further promoting sustainable agriculture (Bechtsis et al.,
2017), but the main limiting factors in their adoption lies in production inefficiencies
and the lack of economic feasibility over the very short period of AGVs’ potential
utilization each year (Bechar and Vigneault, 2016).
In this regard, the present study proposes a software tool for daily field level
agricultural activities. More specifically, the tool provides a cyber-physical interface
to the physical entities (e.g. field boundaries, loads and vehicles) located at the field
level. The Google Maps service is integrated at our tool in order to provide the field
layout, the absolute coordinates of the entities and a real-time monitoring tool.
Sophisticated algorithms can be used for navigation at simulation level while the
absolute coordinates are sent to the vehicle for the physical navigation.
The remainder of the manuscript is organized as follows. In Section 2 we describe
the system's architecture, along with the required input and operational data, and the
field monitoring process. In Section 3, we discuss the key findings of the simulation
tool application in a conceptual field. Finally, in Section 4 we wrap-up with
conclusions and limitations.
2 Software Tool Design Architecture
In this paper we propose a stepwise simulation tool that combines GPS guidance
techniques, enabled by the Google Maps service, with routing algorithms used to
navigate an AGV on a field. The architecture of the system is demonstrated in Figure
1.
364
Fig. 1. System’s architecture.
The system's architecture involves both the simulation and the real world as the
real world coordinates are imported into the simulation environment; the motion
planning algorithms are tested at simulation level and can be transferred to the robot;
the robot can exchange information (coordinates, sensor information, robot's state)
with the simulation environment and inform about its current state. The proposed
software application is built with the C# framework using the Microsoft Visual
Studio 2013 platform. The software enables the import of specific entities into the
simulation environment (vehicle, obstacles and loads) in a grid layout. Each entity
that enters the grid is an object of a developed class in the object oriented
programming environment and is located on a specific cell of the grid. The software
tool embeds maps using the GreatMaps open source project (GMap.NET) that
belongs to the CodePlex hosting site supported by Microsoft. The GMap acts as a
middleware in order to promote the use of specific map providers (Google, ArcGIS,
Bing, OpenStreetMap, Yahoo and many more) and map types (political maps, terrain
maps, transportation maps, satellite maps) in windows applications. For working
with agricultural logistics it is necessary to implement satellite maps into the
software tool in order to select the appropriate provider and embed the map to the
application. Cell coordinates are calculated from the GMap middleware and create a
direct link to the real world providing predefined coordinates for each cell in the grid.
Finally routing algorithms are used at simulation level in order to navigate the
vehicle at the field. At each step of the simulation process the X,Y coordinates of the
cell are stored at the system's log. This log file can be provided as an input to the real
world vehicle in order to provide the navigation path in means of GPS coordinates.
365
3 Computational Study
3.1 Field Specification
Firstly, the provider's map is embedded to the developed simulation tool to
implement an absolute position oriented guidance system. It is worth mentioning that
the GMap.NET acts a middleware in order to promote maps from different providers
to windows applications. It enables the use of routing, geocoding, map directions and
map previews to developers. As a next step the GMapControl tool was incorporated
to the Toolbox for providing a handler to the MainForm of the application which
links the form to the map's service provider using the proper libraries. Inside the form
the GMap's menu is implemented in order to handle the maps properties and the user
can select the presentation details of the map. Furthermore the selection "Mark
Rectangle" provides an event handler for the creation of a dedicated form that
embeds the region of interest and metadata information to the application. The
region's coordinates and the map's scale are considered critical for the software tool.
The user manually locates the region of interest (Figure 2).
Fig. 2. The user locates the region of interest using the Google Maps Service.
Secondly, the user locates the specific agricultural field of interest (Figure 3) and
imports the captured field image into the simulation tool for further processing. The
absolute latitude and longitude coordinates of the selected area’s vertices are then
exported from GMap and are imported in the tool. The operator can save the region
with the required metadata for later use. As each field can have certain restrictions
(complex boundaries, areas with obstacles, water covered areas that alter the fields
landscape vary from year to year) the operator can capture a wider area of the map
366
and at a later stage set the exact boundaries of the field. The region of interest is fully
functional at application level for providing efficient inter-agricultural logistics.
Fig. 3. The user selects the specific agricultural field of interest.
Following that, the user provides feedback and interoperates with the system by
marking grid cells in the tool’s embedded layer applied on top of the retrieved
agricultural field’s image. The selected grid cells indicate the field’s boundaries,
mark the position of any obstacles in the field, specify exact location of goods that
should be transferred and point the start and exit positions of the AGV (Figure 4).
Fig. 4. The user defines the agricultural field’s characteristics.
367
Each grid cell can have different properties that describe the objects of the
physical world (e.g. barrier object for the outline of the field, obstacle object for
obstacles inside the field, start and final position object of the AGV). Furthermore, at
each grid cell correspond specific latitude and longitude coordinates formed by the
original absolute coordinates. This projection enables the absolute positioning of all
grid cells and provides an absolute localization mechanism for the AGV within the
field. Using these entities, the user creates a map of the field’s layout and establishes
the active operational zone for the AGV. Finally, the AGV is able to navigate within
the active zone following the motion planning algorithms that are implemented. The
proposed application can find the shortest path between two cells in the active zone
for either loading the vehicle or navigating to the exit point using the A* algorithm.
The A* algorithm can calculate the shortest path after examining all the possible
routes in the grid based environment while simultaneously avoiding all the possible
obstacles (other loads, boundaries etc). The program also offers a visualization of the
vehicle's movement in the field. Finally, for the evaluation of specific agrilogistics
tasks, desired operations and sustainability performance metrics (scheduled tasks,
vehicle state, direct and indirect emissions, etc.) can be reported from the software.
3.2 Exemplar Demonstrator
A common scenario is examined in the present paper as exemplar demonstrator. An
in-field transfer of goods is simulated. The load is allocated to specific cells at the
active field area and makes a transfer request to the AGV. The AGV begins from the
starting position and finds an optimum pathway using the A* algorithm for reaching
the user-specified goods. The AGV loads the goods and then transfers them to the
exit position (Figure 5).
Fig. 5. Simulation.
368
Users monitor the movement of the AGV while the absolute coordinates of the
grid cells are stored in a list to guide a corresponding physical AGV. The location of
the AGV can be monitored in real-time by the proposed software tool.
4 Conclusions
This paper contributes to the operations research field by allowing the application
of simulated results direct to the guidance of a physical vehicle. Unlike existing
sophisticated tools, the developed navigation mechanism is user-friendly and highly
customizable at outdoor navigation.
The tool provides a real time interface to the Google Maps engine in order to
handle the field's boundaries, to map the obstacles' coordinates and to exclude
specific areas from the vehicle's path. Real time positioning of the vehicle is used for
monitoring the vehicle's status and providing real time feedback to the motion
planning algorithm (Edwards et al., 2016; Zhe et al., 2015).
The proposed tool could lead to a commercial navigation system for agricultural
machines providing a friendly user interface and sophisticated algorithms for motion
planning while taking into consideration all the appropriate agronomical constraints.
References
1. Bechar, A. and Vigneault, C. (2016) Agricultural robots for field operations:
Concepts and components. Biosystems Engineering, 149, p.94-111.
2. Bechar, A. and Vigneault, C. (2017) Agricultural robots for field operations. Part
2: Operations and systems. Biosystems Engineering, 153, p.110-128.
3. Bechtsis, D., Tsolakis, N., Vlachos, D. and Iakovou, E. (2017) Sustainable supply
chain management in the digitalisation era: The impact of Automated Guided
Vehicles. Journal of Cleaner Production, 142, Part 4, p.3970-84.
4. Edwards, G.T.C., Hinge, J., Skou-Nielsen, N.,Villa-Henriksen, A., Sørensen,
C.A.C. and Green, O. (2016) Route planning evaluation of a prototype optimised
infield route planner for neutral material flow agricultural operations. Biosystems
Engineering, 153, p.149-57.
5. Gomez-Gil, J., Alonso-Garcia, S., Gómez-Gil, F.J. and Stombaugh, T. (2011) A
simple method to improve autonomous GPS positioning for Tractors. Sensors, 11,
p.5630-44.
6. Han, X.Z., Kim, H.J., Kim, J.Y., Yi, S.Y., Moon, H.C., Kim, J.H. and Kim, Y.J.
(2015) Path-tracking simulation and field tests for an auto-guidance tillage tractor
for a paddy field. Computers and Electronics in Agriculture, 112, p.161-71.
7. Ruiz-Canales, A. and Ferrández-Villena, M. (2015) New proposals in the
automation and remote control of water management in agriculture: Agromotic
systems. Agricultural Water Management, 151, p.1-3.
369
8. Sarangi, S., Umadikar, J. and Kar, S. (2016) Automation of Agriculture Support
Systems using Wisekar: Case study of a crop-disease advisory service.
Computers and Electronics in Agriculture, 122, p.200-210.
9. Xiang, R., Jiang, H. and Ying Y. (2014) Recognition of clustered tomatoes based
on binocular stereo vision. Computers and Electronics in Agriculture, 106, p.75-
90.
370