GPS-based navigation systems are widely used today and their benefits are ... StreetMaps (OSM), which is a project that provides free geographical data that .... and driving distance with respect to a Point Of Interest (POI) like a biogas plant.
CRUV - A Collaborative Route-Planning System for Utility Vehicles Christopher Tuot1,2 , Darko Obradovic1,2 , Fabian Fichter1 , and Andreas A. Dengel1,2 1
2
University of Kaiserslautern, Germany German Research Center for Artificial Intelligence (DFKI), Kaiserslautern, Germany
Abstract. In this paper we first present the Collaborative Route-planning system for Utility Vehicles (CRUV), a Spatial Decision Support System (SDSS) using the geographical information provided by OpenStreetMap (OSM) to deliver more adapted routing for heavy and large vehicles (e.g., lorries, tractors, . . . ). We further describe a second SDSS: the biomassplanner, a mashup of maps and tabular information developed at the German Research Center for Artificial Intelligence (DFKI) and providing an intuitive platform for efficient biomass planning on the scale of a single production plant. Finally we explain how our collaborative routeplanning system for utility vehicles is beeing integrated with the DFKI biomass-planner, currently suited for single production plants, to enable efficient biomass planning and logistic planning on a larger scale.
1
Introduction
GPS-based navigation systems are widely used today and their benefits are highly appreciated. Utility vehicles however have more special requirements than a standard car, as their physical characteristics impose new routing constraints. There is a very small number of UV-specific navigation systems on the market, but due to low competition and high prices, most UV drivers use standard offthe-shelf navigation systems instead. This regularly leads to troublesome situations and accidents in traffic, as confirmed by a report of the German automobile club ADAC. We suggest a new approach for a collaborative navigation system that could lead to better datasets and lower costs. As demonstrated by Wikipedia and similar Web 2.0 projects, collaborative approaches can provide qualitatively and quantitatively good information. We based our CRUV system on OpenStreetMaps (OSM), which is a project that provides free geographical data that is collected in a collaborative way and stored in a central database.
2 2.1
Collaborative open-source Routing system for Utility Vehicles (CRUV) System Description
In [6] Goodchild introduces the term Volunteered Geographic Information (VGI) describing the widespread engagement of large numbers of private citizens in the creation of geographic information. In [5] Goodchild further describes one of the biggest and most powerful sensor network that could be used for the acquisition of geographic information: “the six billion humans constantly moving about the planet collectively possess an incredible rich store of knowledge about the surface of the Earth and its properties”. OpenStreetMap (OSM) is a project aiming at providing free geographical data that is acquired in a collaborative way and stored in a central database. The data is acquired using a GPS receiver and can then be annotated with additional relevant meta- information before being uploaded and integrated into OSM. OSM was founded in 2004 by Steve Coast whose motivation was to find an alternative to the expensive maps, required work with laptops and GPS technology. OSM has now more than 50,000 participants from around the world, its community is still growing. In terms of quality, the geographical information of OSM follows the one of its competitors like Google, Navteq or Tele-Atlas. In some cases, the OSM data might even be more accurate [9]. However OSM still lacks detailed geographical information for some sparsely populated areas but this will for sure be fixed with time and taking into consideration the growing OSM community. OSM has also a very flexible framework to store additional meta-information that is relevant for routing, e.g., speed limits or maximum weight. This explains why we decided to use OSM for our CRUV system to store all the relevant geographical information for routing. No changes have been realized to the OSM user management system. A user can login and add, modify or delete geographical information and its meta-information. Additionally, we introduced the possibility for a user to score the data-quality and also define different levels of trust regarding the input of other users or groups. Routing itself is realized using the pgRouting module for the PostgreSQL database with PostGIS to support spatial operations. pgRouting first builds a dedicated topological representation of the route network to enable fast routing. pgRouting supports different routing strategies for the shortest path like A* [8] or Dijkstra [2]. Out of the box, our system can handle meta-information like the maximum weight, maximum length, maximum height, maximum width, maximum and minimum speed limits, bridges, tunnels and road types. Thanks to a straightforward rulebased language, any additional attribute can also be taken into consideration for routing. 2.2
Comparison and User Study
We did not yet perform an evaluation of our CRUV system in terms of routing quality but as a first result we realized some simple comparisons with other
routing systems. Fig. 1 (left) depicts a comparison between CRUV and Map24 for traditional route planning, showing no quality difference. Fig. 1 (right) presents a comparison between CRUV and GoogleMaps where bridges below a specific maximum weight should be forbidden. GoogleMaps simply does not support such query and this explains the differences in the results. CRUV proposed a 2.8 km route, instead of 1.5 km for GoogleMaps, but preferring another bridge while respecting the given maximum weight of the vehicle in question.
Fig. 1: Left: A comparison between Map24 and CRUV. Right:A comparison between GoogleMaps and CRUV.
Furthermore, we have realized a contentment survey with agricultural technicians to check whether or not the current functionalities of the CRUV route planner were actually meeting their requirements. The questionary is depicted in Fig. 2 and the results are summarized in the graph of Fig. 3. The questionary consists of a list of 22 features of the CRUV route planner organized in 3 categories: Pertinence of a navigation system for utility vehicles This category actually only consist of a single feature that is the general demand for a navigation system for utility vehicles. Model complexity This category is the most important one in the survey and regroups all features regarding the complexity of the model in terms of routing. For example, participants have to evaluate the current parameters available for the routing. System usability Features in this category are related to the graphical user interface and especially to the usability of the system. The aim of the survey is to let participants give their opinions on how important those features are and how satisfied they are with the current implementation of those features in the system. There are four different levels of importance for a feature: very low, low, important and very important. The satisfaction about the implementation of a feature can also be defined in a similar way by very low, low, good or very good. Five agricultural technicians accepted to take part to the evaluation of the CRUV route planner. Before taking the survey, the system was presented to the
t an po r ve tant ry im po rt
low
im
od low
go
ry
ry
ve
od
ve
go
ry l low ow
ve
Satisfaction A. Pertinence of a navigation system for utility vehicles: 1. General demand for a navigation system for utility vehicles 5 B. Model complexity: 1. Support for different strategies of driving directions (shortest vs. fastest) 5 2. Support for average speeds 4 1 3. Support for intermediate stages 4 1 4. Support for different road types (motorway, highway, country road, country lane, private, private road): 4.1 General demand 2 3 4.2 without priorities 3 2 4.3 with priorities (none, few, normal, prefered, strong) 3 2 5. Support for bridges 5.1 General 1 4 5.2 without priorities 3 2 5.3 with priorities (none, few, normal, prefered, strong) 3 2 6. Support for tunnels 6.1 General 1 4 6.2 without priorities 3 2 6.3 with priorities (none, few, normal, prefered, strong) 3 2 7. Support for different vehicles types 7.1 Support for turn-off restrictions 5 7.2 Support for weight, height and width 3 2 7.3 Support for different fuel types 2 1 2 7.4 Support for maximal speed 3 2 8. Possibility to ignore lane direction 4 1 9. Possibility to use OpenStreetMap meta-informations 3 2 C. System usability: 1. Zoom 3 2 2. Move pane 1 2 2 3. Resize map 1 2 2
Importance 1
3
1
3 2
1 1 1
4
1 1
4
1 1 2
1
4
2
1
2
2
1
2
2
1 5
1
3
1
2
2
1 4
2
1
2
1
2
1 2 2 2
2 1 2 1
1
3
1
1 1 1
2 1 1
1 4 4 4
Fig. 2: Questionary for contentment regarding the CRUV route planner. 5 agricultural technicians took part to the survey. The questionary consists of a list of 22 features that have to be evaluated in terms of importance and satisfaction, i.e., how important a feature is and how good this feature has been implemented in the current system.
participants. Therefore the participants were given some background information about the system and also a live demo. Then each participant had the chance to work with the system for 30 minutes and ask all remaining questions. Finally we went with each participant through the survey making sure that they understood correctly which features they were asked to evaluate. The results of the survey are summarized in the graph depicted in Fig. 3. The idea of such a graph is to identify potential features that are important and not yet correctly implemented since not satisfying the users. The graph shows clearly that most of the implemented features were considered as important and that the participants were quite satisfied with current implementation of those features in the system. Please notice that features B.4.2, B.5.2 and B.6.2, considered as not being important, were presented in a negative form, e.g., 6.2 Support for tunnels without priorities. This was done to check the fact that those features expressed in positive form B.4.3, B.5.3 and B.6.3 are actually of importance since those 3 features are actually the key features of the system. B.4.3, B.5.3 and B.6.3 were indeed defined as being important and well implemented.
3
Biomass-planner
The biomass-planner, developed at the German Research Center for Artificial Intelligence (DFKI), is a prototype for a Web-based Spatial Decision Support System (WSDSS) demonstrating the benefits of location-based decision making using digitalized geographic information about ground allocation and soil quality [11]. The biomass-planner uses an extended version of the Biomas Yield Model (bym) that has been developed at the University of applied sciences Eberswalde [10,1]. The biomass-planner can compute the yield for 16 different crops depending on the soil quality and the precipitation levels. Different kinds of scenarios are taken into consideration, e.g., conventional vs. ecological farming and three different levels of precipitation (low, normal, high). On the whole, six different scenarios are computed at once i.e., for conventional or ecological model each with 3 levels of rainfall.
4
Combining Biomass-Planning with Route-Planning
The biomass-planner is currently suited for single production plants but could be combined with the collaborative route-planning system for utility vehicles to enable efficient biomass planning and logistic-planning on a larger scale. Indeed, experimental research in the field of resource planning, like for example for a biogas plant, show that most of the costs are related to transport and can only be optimized by relying on efficient route-planning [7]. Fig. 4 is a snapshot of a mock-up for a mash-up of maps and table information providing an intuitive platform for biomass planning and logistic-planning. The system takes as input a production plan established with the biomass-planner, i.e., a list of fields with their crop definition and the yield prognosis. The aim of the tool is to help deciding which fields should be taken into consideration depending on the yield
high
High Priority for improvements
High Priority for continuation ●
B.1 ● B.4.1
● B.7.2
B.6.3 ● B.5.3
Importance
C.3 ● C.2
● B.7.3
B.4.3 ● C.1 B.7.4
● B.6.1 ●
●
B.5.1
A.1
● B.3 ● B.9 ● B.2
B.7.1 ●
● B.8
B.5.2 ● B.4.2
low
Low Priority for improvements ● B.6.2
Low Priority for continuation
low
high Satisfaction
Fig. 3: Graph showing the contentment regarding the CRUV route planner
and driving distance with respect to a Point Of Interest (POI) like a biogas plant. The user can specify the transport strategy by defining the number of available trucks and or trailers. A very important feature of the system is being able to specify the characteristics of the vehicle(s) (e.g. length, width, height, and fuel consumption. . . ) that are used for route-planning. The real driving distance can therefore be taken into consideration instead of the typical flight distance.
Fig. 4: A mash-up of maps and tabular information providing an intuitive platform for biomass-planning and logistic-planning (mock-up).
The maximal distance between the POI and the field can be defined using a simple slider. To help optimizing the maximal driving distance, the user has access to all relevant information. All field relative information like the size, the driving distance between the POI and the field, the kind of crop, the yield and the transportation cost are summarized in the table on the left. All the fields can be visualized using the map on the right (currently based on GoogleMaps). The fields in green are within the maximum driving distance, the fields in gray are not. A graphical representation of the field repartition with respect to the driving distance is also available above the table. By selecting a field, using the table or the map, the user can explicitly chose to include or exclude this field disregarding the maximum driving distance. The specific route between the POI and this field is drawn in blue on the map. The information about the slope of the route can also be visualized in the graphic below the table. Finally, an overview of the total available biomass and transportation costs is available at the top of the application.
References 1. S. Brozio, D. Mueller, and P. H-P. Biogaspotenzial des Landes Brandenburg. Agrartechnische Forschung (Agricultural Engineering Research), 12(6), 2006. 2. E. Dijkstra. A note on two problems in connexion with graphs. numerische Mathematik, 1(1):269–271, 1959. 3. R. Earles. Sustainable Agriculture: An Introduction. Summary of ATTRA (Appropriate Technology Transfer for Rural Areas), pages 1–16, 2002. 4. G. Feenstra, C. Ingels, and D. Campbell. What is sustainable agriculture. University of California–Davis. Available at: http://www pbs.org/ktca/farmhouses/sustainable agriculture.html, 2005. 5. M. Goodchild. Citizens as voluntary sensors: spatial data infrastructure in the world of Web 2.0. International Journal, 2. 6. M. Goodchild. Citizens as sensors: the world of volunteered geography. GeoJournal, 69(4):211–221, 2007. 7. C. Gunnarsson, L. V˚ agstr¨ om, and P. Hansson. Logistics for forage harvest to biogas production?Timeliness, capacities and costs in a Swedish case study. Biomass and Bioenergy, 32(12):1263–1273, 2008. 8. P. Hart, N. Nilsson, and B. Raphael. A formal basis for the heuristic determination of minimum cost paths. IEEE transactions on Systems Science and Cybernetics, 4(2):100–107, 1968. 9. heise.de. Offene datensammlung schlaegt googl maps, 2008. http://www.heise.de/newsticker/Offene-Datensammlung-schlaegt-Google-Maps– /meldung/115120. 10. H. Piorr, K. Kersebaum, and A. Koch. Die Bedeutung von Extensivierung und o ¨kologischem Landbau f¨ ur Strukturwandel, Umweltentlastung und Ressourcenschonung in der Agrarlandschaft. Eberswalder Wissenschaftliche Schriften, 3:99–114, 1999. 11. C. Tuot, M. Sintek, and A. Dengel. IVIP–A Scientific Workflow System to Support Experts in Spatial Planning of Crop Production. Lecture Notes in Computer Science, 5069:586–591, 2008.
