Semantic Traffic Data Analysis for a Local Leader Election Algorithm (LLEA) Roua Elchamaa ∗ , † , Rima Kilany Chamoun† , Baudouin Dafflon‡ , and Yacine Ouzrout∗ ∗ Univ. Lyon, Lumière University Lyon 2, DISP, F-69676, Bron, France Email: {roua.elchamaa, yacine.ouzrout}@univ-lyon2.fr † Saint-Joseph University (ESIB) , Lebanon Email: rima.kilany@usj.edu.lb ‡ Univ. Lyon, University Claude Bernard Lyon 1, DISP, F-69676, Bron, France Email: baudouin.dafflon@univ-lyon1.fr Abstract—MASCAT is a research-based road traffic simulator. II. R ELATED W ORK In this paper, we propose a plugin for MASCAT. The aim of this proposition is to provide a semantic data analysis for the V2V (Vehicle-to-Vehicle) communications need to be tested simulators Multi-Agent System (MAS). The plugin introduce an interpretation phase during vehicle to vehicle communication through intensive experiments. Simulation models should in- (V2V). It will allow connected vehicles to make the most ef- clude mobility models for providing accurate simulation of ficient behavioral decisions based on the state of surrounding real time vehicular networking environment. The simulation environment. We designed a semantic web ontology to describe tools should be selected based on their compatibility with traffic data and ameliorate behavioral decisions. Our semantic application requirements and similarity to real-time traffic. We plugin can link, structure, analyze MASCATs traffic data and can also optimize the Local Leader Election Protocol. Indeed, analyzed some existing traffic simulators. For the needs of our we demonstrate that a small percentage of connected cars can work, the chosen platform will not necessarily have to offer a ensure traffic regulation specially in a shifting environment. very detailed definition of a mesh network. Index Terms—Ontologies, Semantic Rules, Multi-agent System, Our affinities with free and community software push us to Connected Vehicles, Local Leader Election Algorithm (LLEA) retain two main candidates: SUMO and MovSim, which offer in addition a lot of functionalities by their stage of maturity I. I NTRODUCTION compared to the previous platforms. SUMO is widely used The progressive growth of the number of vehicles in our and well represented in research, and MovSim is a newer and cities is considered as a cause of traffic congestion, but it is not less represented platform at the moment. the only reason for traffic jam. In order to study this problem Simulation of Urban MObility (SUMO) is a road traffic thoroughly, it is necessary to simulate real traffic conditions simulator [12]. With SUMO, traffic demand consists of single by using an appropriate traffic simulator that would help us vehicles moving though a given road network. Real-world study the behavior of the vehicles. networks are modeled as graphs, where the roads and intersec- Connected vehicles are able to sense and adapt their be- tions are respectively represented as a graphs. In SUMO, each havior according to their environment. How these vehicles vehicle’s speed is computed using a Car-Following Model. will change the way we deal with traffic regulation? What This model usually compute a targeted vehicle’s speed by percentage of connected vehicles is enough to ensure traffic looking at its own speed, its distance to the group leader, and regulation? In this paper, we will answer these questions, the leader’s speed. SUMO is widely used and well represented by describing the details of the framework we implemented in research, but MovSim was recently developed based on the as a semantic plugin for the MASCAT simulator. Semantic main recent concepts in traffic theory [28] [30]. description provides the means to store information while Regarding V2V inter-vehicle communication, an extension of giving it a logical meaning and a richer context. SUMO is under development and aims to study the effects This also means that we can construct advanced and intel- of on-board applications on driver behavior [13]. In addition, ligent queries over an ontology. The objective is to obtain the VEINS project aims to offer a set of models dedicated to inter- information inferred (by a reasoner) from the ontologys set of vehicle communications (IVC) in SUMO. But in MovSim pre-defined relations and rules. Semantic rules are simply a set side, the developers imagined the integration of these forms of of IF-THEN statements for structuring complex axioms about communication from the start of the project [23] [26]. MAS- a specific domain. We will start in section II with exploring CAT [8] is a research-based road traffic regulation simulator the related work. In section IIIwe will describe our proposed developed using the already existing Movsim simulator [29] by solution. The solution design and implementation are detailed transforming it into a Multi-Agent System where each vehicle in section IV, while results are shown and analyzed in section is modeled as an intelligent separate entity which can run V. Finally, conclusions and future work are exposed in section according to an algorithm of its own. Each instance of the VI . Vehicle entity is simulated independently either by respecting Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). the well-known IDM mobility model, thus representing the ranking and on a comparison of different GDBMS done by unconnected vehicles or by respecting a Local Leader Election the university of Leipzig [11]. Algorithm (LLEA) to represent connected vehicles (CVs). We came up with a list of the most popular and free to Multi-Agent System (MAS) is a loosely coupled network use GDBMS that support RDF/SPARQL as their data model: of problem-solving entities called agents that collaborate in Virtuoso, allegroGraph, StarDog, GraphDB, BlazeGraph. resolving problems that are beyond the single individual’s And for rule-based inference [1], there are two principle rea- scope of capabilities and knowledge [7]. In the traffic simulator soning strategies: Forward-chaining and Backward chaining. MASCAT, connected vehicles are modeled as agents to find The first one is the Forward-chaining, this type of reasoning the best behavioral decision for the CVs that are moving in strategy involves applying the inference rules to explicit state- a constantly shifting environment. So, to ensure vehicle to ments in order to produce new facts. The second one is the vehicle communication, the vehicles must plan their actions Backward-chaining strategy that which require to start with a jointly to allow better cooperation between them. So the key fact illustrating or a query answering. issue in a MAS is to formalize the coordination between agents In our case, we are interested in forward-chaining reasoning [24]. Many works have been proposed based on MAS and because query time is an essential factor that guarantees the ontologies. The intelligent levels of the system (Knowledge required performance to keep up with the constantly ”shifting Base and Multi-Agent System) in [6], use the knowledge environment” of the connected vehicle in the traffic simulator provided by the IOT devices and its semantic environment MASCAT. Therefore, we need to dynamically create the in order to reason and react to set theses devices. In [2] a Environment individuals and obtain the possible inferences model was implemented based on a multi-agent approach for before querying the semantic knowledge. Virtuoso and Star- urban freight transportation, a knowledge data model was used Dog GDBMS only support Backward-Chaining Reasoning, to represent the urban environment. A city logistics ontology which is not suitable for our case, so we are now only down was proposed. A method for the representation of knowledge to three Graph Database Management Systems: AllegroGraph, and reasoning in Agile Worker-Cobot manufacturing has also GraphDB and BlazeGraph. AllegroGraph [34] leads the largest been proposed in [21]. A similar system based on ontology deployment with loading and querying 1 Trillion triples. and multi-agent in [4] has been proposed for the Construction However AllegroGraph is available on a Windows platform and Cooperation Mechanism of Logistics Vehicle. Inspired by through a Linux Virtual Machine, which could degrade the these related works, we decided to increment the MASCAT performance factor as directly declared on AllegroGraphs web- Multi-Agent-System with semantic knowledge in order to site. Therefore, we considered the use of Ontotexts GraphDB have more realistic simulations, and auto-adapting agents that since it is more popular than BlazeGraph. are able to infer new traffic data such as changing weather Graph database [31] makes the application of operations pos- conditions, accidents, roadworks, A Relational Database may sible based on a graph of data and metadata. Such operations have been the answer to store such knowledge, but a relational can improve performance and greatly increase the speed of database [15] fails if provided with fragmentary or incomplete recovery operations while maintaining the same precision of knowledge because it works upon closed-world assumptions. pure ontology-based approaches and reasoning. Since with The traditional relational database stores concepts in the form Graph database there is no schema, each entity can contain of tables without containing information about the meaning different data attributes, and there is no need to perform of these stored concepts, and how they are related to other join operations on multiple tables in order to obtain the concepts. Ontologies [10] have also been adopted as a possible needed information. In this way, we will eliminate the need representation for complex concepts and domains. In fact, it for saving redundant data and sending complex queries to is a powerful model for mapping and describing information retrieve woven datasets. But, what Graph Database would related to real-world knowledge areas. [25]. Ontologies have we adopt? GraphDB [17], [22] is built on OWL. It uses advanced ways to provide automated classification for different ontologies that allow the repository to automatically reason types of data. However, the processing of the ontology’s about the data [19]. It offers OWL inference allowing to individuals is one of the costliest computational operations create new semantic facts from existing facts. Massive loads, within ontology reasoning. To query this data model, we can queries and inferencing can be handled in real time. And use two query languages SPARQL [3], [9] and SQWRL [18]. to augment the expressiveness of our designed ontology, we Furthermore, the reasoning ability of ontologies can also lead defined a set of Semantic Rules. These rules for GraphDB to problems in processing time. Several types of research have are called ”Entailment Rules”. In conclusion, we decided to been done [14] to analyze the performance of the available use an approach based on the GraphDB Graph database to reasoners. It shows that the structure and size of the ontology provide our vehicles with data interpretation capability in case and the complex queries submitted to the reasoner have a high of disturbance without having a high consumption of time. We influence on the performance of these ontologies. managed to bypass our exclusive need for the standard OWL Therefore, we studied yet other solutions to find a technology Ontology and benefited from a Graph Database Structure that that can provide similar capabilities with better performance. takes an Ontology file as its input. In our research, we are interested in the GDBMS that support RDF triple stores and inferences. So based on the mentioned III. P ROPOSED S OLUTION The Second Decision D2 is a collective decision that happens in a certain zone. It consists of three steps: A. Global Overview • Each CV receives the D1 Decisions of the surrounding CVs • The votes are counted for each CV in this zone • The CV that received the most votes is elected to be the local leader for this zone and all CVs must adapt their speed to follow the speed of the local leader. In our previous work, the implementation of this election protocol improved traffic flow in an optimal environment with no disturbances, by providing speed recommendations to the driver. This kind of recommendation will be offered to a con- nected vehicle after a negotiation with the surrounding vehicles based on a V2V communication. It allows a reduction of traffic congestion through an election of a local leader selected from Fig. 1. Vehicle behavior in the LLEA the neighboring vehicles. These will adapt their behavior to follow the elected local leaders speed recommendation. Since Based on our studies, we know that the vehicle driving this protocol is based on the concept of multi-agent system, behavior is one of the main factors that can disturb traffic flow. each vehicle follows the cycle of Perception- Decision-Action Each strange condition on the road can affect the surrounding However, many disturbances can affect the flow in real life vehicles. In case of an accident for example because of snowy like weather conditions, roadworks, unsafe and unexpected weather, near surround- ing vehicles will be aware of the events, obstacles on the road, vehicle stopping, stuck vehicle, situation and will try to control their behavior accordingly. no protected accident area, emergency braking, unexpected However, this reaction can have diverse effects on the flow, queue end, ... because other vehicles on the road might not be aware of what In order to be able to achieve traffic regulation in any is happening. They only know that the road is blocked and they context, it is strongly needed to provide a common framework cannot react appropriately. This is not the case of connected that allows Traffic data to be shared and reused between vehicles which can react to regulate congested traffic in normal entities. Therefore, it is necessary to Link and structure road traffic conditions as we demonstrated in our last work [5], traffic data exchanged between vehicles (agents) to simply where we succeeded to regulate traffic using our proposed access to the knowledge that it already contains and express Local Leader Election Algorithm (LLEA) (Fig.1) with only accurately the real road traffic situation. 10% of connected vehicles on the road. This algorithm is In other terms, we need to give our connected vehicles computed for each CV every 10 seconds. The CVs speed is modeled as agents in the LLEA a particular knowledge of the programmed to vary in a speed interval I = [0;130[ (KM/h). environment to get them to: first understand the situation, then We try in this case to mimic realistic traffic state variations, inform other connected vehicles on the road and finally emerge this interval was originally decomposed into 4 smaller, static with a correct behavior whatever the conditions, without and fixed intervals based on a certain logical traffic state: causing massive traffic congestion. As we showed in the state • I1 = [0;30[ : I1 includes vehicles in traffic congestion of art section, an ontology can provide a complete description • I2 = [25;50[ : I2 includes vehicles in critical state of traffic data, and thus can allow an accurate interpretation • I3 = [40;90[ : I3 includes vehicles in normal traffic of traffic context, which will give the connected vehicles • I4 = [80;130[ : I4 includes vehicles in fluid traffic adequate reactions. This ontology will represent the real road traffic road environment with the definition of all entities and These intervals will be modified in our actual solution and will properties that can affect the Connected Vehicles behavior. By be dynamically change based on each CVs Environment. This augmenting the ontology with pre-defined traffic rules which LLEA consists also of two consecutive decisions. The First respect to the European Traffic Norms and Regulations, the Decision D1 is a local decision that each CV has to make. It connected vehicles will adapt their behavior according to the consists of three steps: knowledge inferred, at the interpretation level. • The CV must determine the interval Ii whose speed limits We choose to begin with inferring the adequate speed intervals are respected by the highest number of surrounding CVs needed for different situations on the roads, in order to provide • The CV must compute the optimal speed of the interval CVs a with flexible intervals as we see in table (Fig.2). This Ii which happens to be the median of the speed values will enhance the local leader election algorithm, since the of the surround- ing CVs in Ii local leader speed will be calculated based on dynamically • The CV must choose the neighbor CV that has the closest inferred speed intervals. The framework we implemented speed to the computed median and send this decision to is extensible, so in order to cover any environmental and the surrounding CVs contextual knowledge inference, we only need to add related concepts to the ontology and the corresponding rules. MASCAT do not take into account the environment in the ideal world under disturbances like the meteorological conditions such as the visibility, weather We will describe how we were able to alter the behavior of the MASCAT connected vehicles so that they would consider their surroundings, and this, without deteriorating the simulators performance while always ensuring traffic regulation. A. Semantic Plugin The fact that the interactions between agents are highly information-dense raises many problems. Because of that our research was oriented to use a semantic method. This method can achieve a step of data interpretation by each agent on the multi-agent system. This goal can be achieved after implementing the Semantic Traffic Data Analysis plugin. Its Fig. 2. Intervals according to the European Traffic Norms and Regulations main objectives are to Link, structure and analyze Traffic Data Fig.3 is a UML activity diagram that shows the control flow in order to optimize the Local Leader Election Protocol imple- of the processing steps that enable a connected vehicle to mented in MASCAT. MASCAT Connected Vehicules would understand its environment based on a realistic inferred knowl- elect a local leader in a certain radius around them, based on edge. Individuals of our semantic data model, representing the the computed median speed value in the speed interval having environmental conditions, will be created dynamically on load the highest number of vehicles. The vehicle with the closest time. Each connected vehicles can check its environment to speed to this median is selected to become the local leader, so know if there is a need to adapt its behavior. In our approach, its speed will be adopted by surrounding connected vehicles. the environment is controlled by three main criteria: the The Plugin will provide each connected vehicle with the road type, weather, and visibility. As already explained, other adequate ”Speed Intervals” relatively to a certain context. The criteria can be simply added further on. If the environment LLEA will be then executed, based on the inferred intervals, didn0 t change, it would load the behavior of the connected in order to compute an adequate recommended speed. Doing vehicle in the LLEA directly without a level of interpretation. so will make the simulation more realistic by ensuring that If it did change, a query is prepared with the Road Type and the CVs behavior would adapt to a continuously shifting weather as well as the visibility detected by the corresponding environment. In the following sub-sections we will start by sensor, as input parameters. The defined ontology is then describing the ontology based approach and the connection queried with these aggregated values in order to infer new between the plugin and the simulator, then we will describe knowledge related to this correlation of values. If the query the graph database approach which we adopted for its high returns no result, this would mean that this is an occurrence of performance compared to the ontology based approach. a new context, so new individuals will be created and added 1) Ontology Based Approach: to the the ontology representing these detected conditions. If • Firstly, as shown in the Fig. 4, we started to implement successful, the new inferred speed intervals will be returned our solution in Protégé with a preliminary prototype by the query at the interpretation level in order to choose a including an initial ontology representing the context of Local Leader in the LLEA respecting the particular conditions road traffic. This ontology includes main classes (Context, at this particular time. The election protocol will be launched Weather) and properties (hasWeather, hasSpeedLimit, every 10 sec by every connected car, and based on the results hasRecommendedSpeed) of the election the Local Leader, vehicle adapt his speed in • Secondly, in order to obtain a recommended speed based order to follow his Leader. on knowledge about meteorological conditions, this solu- tion was augmented with semantic rules using semantic IV. S OLUTION D ESIGN AND I MPLEMENTATION web rule language (SWRL) to respond to each traffic MASCAT is a research simulator based on a Multi-Agent context (for each individual). This ontology uses the System of vehicles, implemented in Java, over the MOVSIM rule engine Drools and the Pellet reasoner in order to simulator. In our previous research work on MASCAT we generate the axioms of the defined rules, inferring facts proposed and implemented a local leader election protocol. and checking the consistency of the ontology. The purpose of this protocol was to ensure traffic fluidity • Thirdly, based on this solution, we succeeded in returning and regulation on a highway. This solution has shown its a value of ”the recommended speed” for a certain context effectiveness in tackling the congestion problem with a small by using a SPARQL query. [22] percentage of connected vehicles. However, the MASCAT 2) Connecting a semantic module to the MASCAT simula- simulator does not include a very important feature which tor: In order to connect MASCAT to the semantic plugin, we is the auto-adaption of connected vehicles under several implemented a Java module that uses the Jena API to query disturbances in the environment. The connected vehicles in the OWL Ontology. The communication with the simulator Fig. 3. Vehicle behavior in case of a semantic traffic data analysis with RDF and Linked Data (Section II). GraphDB can process an input OWL Ontology file and has a built-in reasoner (TRREE) that automatically can make inferences at load time (Forward-Chaining Reasoning), with a very good performance level, enabling us to connect to MASCAT and run our simulations appropriately. After en- suring that the semantic plugin is highly functional at the Fig. 4. First Prototype Ontology on Protégé technical level, we went through enhancing it at the semantic level, by extending the input ontology in order to be able to infer speed intervals according to the vehicule’s context. was done through a dedicated Singleton Java class that we We explored many proposed existing ontologies [16], [20], named the SemanticDecision class. A simple variable was [27] which include many entities and concepts that can affect used to toggle the semantic behavior On and Off, which will the behavior of the CV and their perception to discover their consequently toggle the interpretation phase On and Off, thus surroundings and respond to perturbations (entities such as granting the vehicles with two possible behaviors: the Weather property of the entity Environment). We used a • The First behavior where the CVs are following the combined subset of these concepts, nevertheless the ontology steps local leader election algorithm (LLEA) without a could be extended and refined furthermore in future works. semantic data analysis (without interpretation level) As for the speed intervals used in our last work for the • The second one where the CVs have an altered behavior: implementation of the local leader election algorithm, they the Semantic Decision Behavior in which they adapt were defined in the protocol in a static way. But if we want according to the inferred semantic knowledge. to give the connected vehicles the possibility to adapt their This approach was tested in order to communicate an behavior according to their context, the speed intervals should inferred recommended speed to the connected vehicles. The now be inferred from the GraphDB’s Ontology. test showed that the ontology lacked the performance needed Thus we added the following classes to the ontology Fig.5 to keep up with the response delay of the simulator. Therefore, : ”Environment, RoadType, Weather, Visibility” and the fol- we opted to consider as an alternative approach the use of lowing object properties ”hasRoadType, hasWeather, hasVisi- a graph database, which according to ref can perform much bility” and the 4 inferred speed Intervals which are represented better than the ontology in terms of response time. by 8 Data Properties which are the minimum and maximum 3) Graph Database approach: As we found when we bounds of each of these four intervals (hasInterval1Min, has- explored the related work, and after severals tests and com- Interval1Max, ...) parisons, we concluded that the use of the GraphDB Graph We also augmented our Ontology with semantic rules to database would be the best choice to enhance the performance obtain the speed intervals when needed. In GraphDB these of our semantic plugin. We used Eclipse RDF4J (formerly rules are called entailment rules and can be added in a custom known as Sesame) to connect GraphDB to the MASCAT code, ruleset .pie file. The main goal is to offer a connected vehicle and to process and handle RDF data. RDF4J also supports with speed intervals (Fig.2) that correspond to its specific creating, parsing, scalable storage, reasoning and querying in MASCAT were adapting their behaviors as a reaction to a specific disturbance (a specific weather condition) according to a deep interpretation realized by each agent in order to get correct intervals for this situation. In consequence, elect a local leader and respect the offered speed recommendation in the regulation strategy. 1) Particular Weather Condition Scenario: Initial traffic density in this scenario is set to 37 vehicles per kilometer. This setting tends to model a critical regime that can cause Fig. 5. Final Solution Ontology a network capacity drop due to the heterogeneities in the flow of numerous vehicles. We explained previously in Fig. Environment individual, and thus respect road type, weather ?? that connected vehicles should be able to adapt their conditions, visibility and regulations. speed after detecting the actual weather condition based on vehicle knowledge. For example, in case of snowy weather V. R ESULTS and low visibility, on a Highway Road, our semantic query As the simulations done in our latest work show, [5] a will return the following inferred intervals: [0;30[ [15;50[ small percentage of connected vehicles can improve traffic [17;30[ [26;40[ flow, when adopting V2V (Vehicle-to-Vehicle) communication approach combined with a traffic regulation scheme based on a decentralized election protocol. In this model, the elected Local Leader plays a key role in the regulation of the traffic. What was missing in the simulations already undertaken, is the study of the behavior of the connected vehicles in critical weather conditions. This is exactly what the semantic plugin will help us test, since the LLEA will be based on the knowl- edge (inferred intervals corresponding to the specific weather condition) offered by our GraphDB semantic approach. A. Set Up In this section, we describe the common experimentation parameters to run our scenarios: Since Highways suffer from an enormous daily amount of vehicles, the road type studied in our tests is the Highway type. The speed limitations are those defined by European Laws (Fig.2). The baseline scenario consists of a dense traffic state generated on a 3-lane straight highway. The input flow is maintained to 1800 vehicles per hour during the 1200 seconds simulation. Then, we gradually introduce CVs (from 0% to 30% ) which will execute the modified LLEA which takes into consideration the inferred semantic knowledge. For now, Fig. 6. Space-time diagram (vehicles trajectories and speed on the right-most this knowledge consists of the speed intervals, needed for the lane) for the baseline simulation (with 0 to 30% of CV) in Snowy Scenario computation of the recommended speed, in each environment and thus the election of the local leader. These scenarios Figure 6 depicts the trajectories (and speeds) of all the correspond to early CVs deployment phases that will occur vehicles in the right-most lane, for the baseline scenario (0% soon in real-life scenarios. Traffic flow is expected to be of CVs). Then, we gradually introduce CVs (from 0% to 30% ) gradually improved when we increase the number of CVs, executing the LLEA enabled with semantic knowledge. Traffic but we need to verify vehicles behaviors in critical weather flow is expected to be gradually improved with the increasing conditions based on the knowledge offered by our imple- number of CVs . mented semantic plugin. Simulations results presented in this The baseline Scenario (0% CV) models a dense traffic state paper were given by the mentioned Multi-Agent Simulator for where congestion waves (in red color) appear spontaneously Connected and Automated Traffic (MASCAT) [8], augmented and grow, leading to the formation of massive traffic jams. with our semantic plugin. The absence of connected vehicles in this baseline scenario will not allow the detection of environmental condition. The B. Simulations only possible way to control IDM vehicles speed in MASCAT After the implementation of the GraphDB based Solution, simulator is to set their maximum speed to 40 km/h (equal to it was time to test our overall solution to validate if the CVs upper interval limit value). But, this maximum speed limit was not useful. Vehicles didnt respect this maximum value and no Environment (Notice the yellow and orange colors of the regulation strategy of the LLEA was used. Traffic congestion CVs) was mainly due to the speed heterogeneities between individ- uals, and it is accentuated by lane-changes. The purpose of our approach is the self-adaptation in any weather condition using our knowledge-based electoral pro- tocol thus maintaining traffic fluidity even in case of low speed (snowy weather). We can observe on the same figure, the results for the 10% of the connected vehicle which show how connected vehicles were able to adapt their speed based on our proposed semantic approach in LLEA. The local leaders tend to stabilize the flow in very difficult weather conditions. We notice after simulating our scenario with an increased percentage of connected vehicles (up to 50% of CV), connected vehicles were able to behave correctly as we see in the space-time diagrams. We also check our approach with all weather conditions mentioned in Fig. ?? for the road type Fig. 8. Space time Diagram for 100 percent of CVs During Shifting Scenario Highway. We were satisfied by the results, our knowledge- based LLEA shows its effectiveness and reactivity in case of a weather disturbance. The space-time diagram shown in Fig.8 was obtained at the 2) Shifting between several Weather Conditions Scenario: end of this scenario. Notice the middle road segment that is Initial traffic density in this scenario is set to 15 vehicles per between the positions 4000 and 5000. As we can see, the cars kilometer in order to detect the connected vehicles behaviors are immediately slowing down when crossing from the Sunny changes from one weather condition to another on a Highway to the Snowy Environment (Notice the red color of the CVs Road. The aim of this scenario is to validate the concept of on the position 4500). This means that the CVs are actually auto-adaptation of our approach and analyze what happens in abiding to the semantic LLEP result that was executed using the step of shifting between two extreme weather conditions. the queried Speed Intervals. Each time a CV enters a new We tested this scenario in semantically enabled MASCAT Environment, it queries the GraphDB for these Speed Intervals version, with a Road length of 9000 m. This Road is divided and will keep using them until this environment changes, into 9 road segments 1000 m each. From position 0 to 4500, therefore optimizing MASCATs performance. These results we choose to have the first Environment (HighWay - Sunny validates the overall behavior expected from our projects - High) while the second Environment (HighWay - Snowy- outcomes (the plugin and the ontology) that we detailed in Low) will cover the position 4501 to 9000.In 7, we show the previous sections. Our solution was tested and validated how the vehicles started behaving at the beginning of the after several simulations in MASCAT simulator. We can safely simulation. say that the Graph Database approach seems to be way more convenient than its semantic web ontology counterpart by effectively querying the Graph Database structure without having to rely on a less - performant exclusive ontology approach. We switched to the GraphDB GDBMS because we found out that the standard ontology could be easily inserted into a GraphDB repository. Doing so, we accomplished our projectsǵoals by implementing a plugin that would make the simulations more realistic. Being aware of the requirements and constraints, we ultimately did not deteriorate MASCATs pace while performing the necessary computations and tasks. As we have seen in this paper, we have used many tools Fig. 7. Vehicles behaviors adaptation in Shifting scenario in order to develop this project and make it work. We design and build a working Semantic Web Ontology on Protégé and As expected, we can see in Figure 7 that the highway is GraphDB, and successfully linked it to the MASCAT simulator divided into two parts: by implementing a rigorous plugin. We also extended and • The first part (between the positions 0 and 4500) in improved our solution by dynamically creating the Ontolo- which the CVs are moving noticeably fast in a Sunny gys individuals and activating the Visibility Sensor class in Environment (Notice the green and blue colors of the MASCAT. Our plugin helped achieve traffic regulation, by CVs). enabling the connected vehicles to sense and adapt to changing • The second part (between the positions 4501 and 9000) in environmental conditions without deteriorating the simulator’s which the CVs are moving noticeably slower in a Snowy performance. VI. C ONCLUSIONS AND F UTURE W ORK [10] S. Jain. Intelligent decision support for unconventional emergencies. In Exploring Intelligent Decision Support Systems, pages 199–219. To summarize, results indicate that our approach can have Springer, 2018. a positive impact on the peoples lives where the driver of a [11] M. Junghanns, A. Petermann, M. Neumann, and E. Rahm. Management and analysis of big graph data: current systems and open challenges. In CV could obtain a message in which the recommended speed Handbook of Big Data Technologies, pages 457–505. Springer, 2017. of his vehicle is specified. He would then adapt to this speed [12] D. Krajzewicz, J. Erdmann, M. Behrisch, and L. Bieker. Recent knowing that its the most optimal, safe and law-abiding way development and applications of sumo-simulation of urban mobility. International Journal On Advances in Systems and Measurements, to go. This way, he would make sure that the decision he takes 5(3&4), 2012. would contribute in boosting the overall traffic conditions on [13] D. Krajzewicz, J. Erdmann, M. Behrisch, and L. Bieker. Recent the road and help in maintaining an efficient traffic situation development and applications of sumo-simulation of urban mobility. International Journal On Advances in Systems and Measurements, that is conform to traffic regulations and laws. Additionally, 5(3&4), 2012. our plugin made MASCATs simulations more realistic, and [14] T. J. Lampoltshammer and S. Wiegand. Improving the computational therefore more reliable by helping road traffic experts expand performance of ontology-based classification using graph databases. Remote Sensing, 7(7):9473–9491, 2015. their studies and analyze the simulators results after introduc- [15] C. Martinez-Cruz, I. J. Blanco, and M. A. Vila. Ontologies versus ing new factors that can now be taken into consideration thanks relational databases: are they so different? a comparison. Artificial to the ever-expanding Semantic Web knowledge. The designed Intelligence Review, 38(4):271–290, 2012. [16] P. Morignot and F. Nashashibi. An ontology-based approach to relax Ontology can be tailored to the experts needs in order to try traffic regulation for autonomous vehicle assistance. arXiv preprint to replicate any simulation environment thus contributing in arXiv:1212.0768, 2012. finding solutions to the daily traffic congestions in shifting [17] S. Mouzakitis, D. Papaspyros, M. Petychakis, S. Koussouris, A. Zafeiropoulos, E. Fotopoulou, L. Farid, F. Orlandi, J. Attard, and scenarios. J. Psarras. Challenges and opportunities in renovating public sector We propose now some of our recommendations for future information by enabling linked data and analytics. Information Systems works and implementations. We understand that this work is Frontiers, 19(2):321–336, 2017. [18] M. O’Connor and A. Das. Sqwrl: a query language for owl. In not fully complete in what concerns Semantic Traffic Data Proceedings of the 6th International Conference on OWL: Experiences Analysis. The first noticeable thing that could be done is and Directions-Volume 529, pages 208–215. CEUR-WS. org, 2009. expanding the Ontologys designs in order to include more [19] J. Pokornỳ. Graph databases: their power and limitations. In IFIP Inter- national Conference on Computer Information Systems and Industrial parameters and entities that can influence the behavior of the Management, pages 58–69. Springer, 2015. CVs in MASCAT such as the Drivers skills & abilities, traffic [20] E. Pollard, P. Morignot, and F. Nashashibi. An ontology-based model to lights, road intersections, obstacles & road works, pedestrians, determine the automation level of an automated vehicle for co-driving. In Information Fusion (FUSION), 2013 16th International Conference etc. Ultimately, the CVs sensors could be linked to a stream on, pages 596–603. IEEE, 2013. processing software platform like Apaches Kafka in order to [21] A. R. Sadik and B. Urban. An ontology-based approach to enable knowl- simulate real-time environment detection in MASCAT. edge representation and reasoning in worker–cobot agile manufacturing. Future Internet, 9(4):90, 2017. R EFERENCES [22] D. R. Schlegel, J. P. Bona, and P. L. Elkin. Comparing small graph retrieval performance for ontology concepts in medical texts. In [1] A. Al-Ajlan. The comparison between forward and backward chaining. Biomedical Data Management and Graph Online Querying, pages 32– International Journal of Machine Learning and Computing, 5(2):106, 44. Springer, 2015. 2015. [23] M. Schönhof, M. Treiber, A. Kesting, and D. Helbing. Autonomous [2] N. Anand, R. van Duin, and L. Tavasszy. Ontology-based multi-agent detection and anticipation of jam fronts from messages propagated by in- system for urban freight transportation. International Journal of Urban tervehicle communication. Transportation Research Record, 1999(1):3– Sciences, 18(2):133–153, 2014. 12, 2007. [3] D. Calvanese, B. Cogrel, S. Komla-Ebri, R. Kontchakov, D. Lanti, [24] P.-O. Siebers and U. Aickelin. Introduction to multi-agent simulation. M. Rezk, M. Rodriguez-Muro, and G. Xiao. Ontop: Answering sparql 2008. queries over relational databases. Semantic Web, 8(3):471–487, 2017. [25] M. Sir, Z. Bradac, and P. Fiedler. Ontology versus database. IFAC- [4] L. Y. Chu, S. Dang, G. P. Xie, and C. Meng. Construction and PapersOnLine, 48(4):220–225, 2015. cooperation mechanism of logistics vehicle scheduling system based on [26] C. Thiemann, M. Treiber, and A. Kesting. Longitudinal hopping in ontology and multi-agent. In Applied Mechanics and Materials, volume intervehicle communication: Theory and simulations on modeled and 865, pages 681–688. Trans Tech Publ, 2017. empirical trajectory data. Physical Review E, 78(3):036102, 2008. [5] R. Elchamaa, M. Guériau, B. Dafflon, R. K. Chamoun, and Y. Ouzrout. [27] C. Thierry, M. D. Hina, A. Soukane, and A. Ramdane-Cherif. Cognition A local leader election protocol applied to decentralized traffic regu- of driving context for driving assistance. World Academy of Science, lation. In Tools with Artificial Intelligence (ICTAI), 2017 IEEE 29th Engineering and Technology, International Journal of Computer and International Conference on, pages 1013–1020. IEEE, 2017. Information Engineering, 5(2), 2018. [6] B. Gâteau, Y. Naudet, and J. Rykowski. Ontology-based smart iot engine [28] M. Treiber and A. Kesting. Traffic flow dynamics. Traffic Flow Dynam- for personal comfort management. 2016 11th International Workshop ics: Data, Models and Simulation, Springer-Verlag Berlin Heidelberg, on Semantic and Social Media Adaptation and Personalization (SMAP), 2013. pages 35–40, 2016. [29] M. Treiber and A. Kesting. Traffic flow dynamics. Traffic Flow Dynam- [7] M. Glavic. Agents and multi-agent systems: a short introduction for ics: Data, Models and Simulation, Springer-Verlag Berlin Heidelberg, power engineers. 2006. 2013. [8] M. Guériau, R. Billot, N.-H. El Faouzi, J. Monteil, F. Armetta, and [30] A. M. Uhrmacher and D. Weyns. Multi-Agent systems: Simulation and S. Hassas. How to assess the benefits of connected vehicles? a simulation applications. CRC press, 2009. framework for the design of cooperative traffic management strategies. [31] L. Zheng, L. Zhou, X. Zhao, L. Liao, and W. Liu. The spatio- Transportation research part C: emerging technologies, 67:266–279, temporal data modeling and application based on graph database. In 2016. 2017 4th International Conference on Information Science and Control [9] O. Hartig and G. Pirrò. A context-based semantics for sparql property Engineering (ICISCE), pages 741–746. IEEE, 2017. paths over the web. In European Semantic Web Conference, pages 71– 87. Springer, 2015.