Towards an Integrated Service-oriented Energy Management Platform for Plugin Hybrid Electric Vehicles W. Ait-Cheik-Bihi*. M. Bakhouya**. J. Gaber***. R. Outbib****. E. Coatanea**. X. Z. Gao**. K. Zenger** * University of Strasbourg 7, rue René Descartes, F - 67084 Strasbourg, France E-mail: [email protected] ** Aalto University, Otakaari 4, FIN-00076 Aalto, Finland E-mail: {mohamed.bakhouya, eric.coatanea, xiao-zhi.gao, kai.zenger}@aalto.fi *** Université de Technologie de Belfort-Montbéliard 90010 BELFORT cedex, France E-mail : [email protected] ****Université d’Aix-Marseille Av. Escadrille Normandie Niemen, 13397 Marseille cedex 20, France E-mail : [email protected] Abstract: This paper introduces the architecture of a platform that integrates an intelligent energy management system (IEMS) with on-board components and ICT services for efficient energy use in plug-in hybrid electric vehicles (PHEV). The emphasis is put on energy management (when and how each energy source is used) based on conditions and different modes of operation of a PHEV. The relevant use of an information system (data gathered from drivers’ actions and preferences such as driving behavior and itinerary, road traffic information (RTI), and road profile data) as an input for vehicle will improve the global management and will provide an optimized use of energy by selecting the best suitable energy source(s) according to the context (geographical location, traffic state estimation, available energy sources). Preliminary simulations are conducted using a test scenario to show the usefulness of increasing the synergy between the ICT and PHEV for energy management. Keywords: Hybrid electric vehicles, energy management systems, Information and Communication Technologies, traffic limited zones, geofencing.

1. INTRODUCTION Nowadays, it is well known that a main solution for energetic system in transportation will be hybridization of sources. The hybridization can be Thermal/Electrical or/and Electrical/Electrical. However, the penetration of plug-in hybrid electric vehicles (PHEV) on transportation markets faces some obstacles that have not yet been tackled. Among them is the absence of associated services connected with the use of PHEV to mitigate environmental impact, energy use and reduction, and congestion in city centers. The way drivers are using these vehicles can also be optimized and supported in order to address emerging type of constraints in use of this category of vehicles such as the pollution aspects at city level, the risk for drivers not to proper charging stations, and the uncertainty linked to how PHEV will perform in different scenarios such as traffic situation, driver behavior, and road profile. Recently, new topologies of hybridization, methodologies for local control of converters and strategies for global management of energy were developed. Thus, in the literature a large number of studies on the subject are proposed [4]. These studies provide new topologies of hybridization or new strategies for the control and the

management based on driver behavior and environmental and vehicle conditions. Nevertheless, a new and interesting challenge consists in taking into account the environment in which the vehicle evolves. More precisely, it will be very relevant to take into account all available information (road information, position of vehicle in global context ...) in order to manage at best the energy aboard the vehicle. Thus, the relevant use of an information system as an input for vehicle will improve the global management and provides an optimized use of energy. In this paper, we first review and analyze the current situation and technologies used in plug-in hybrid Electric vehicles. It will survey hardware and software components used in PHEV in order to specify and introduce the future energy management system of PHEV under constraints provided by collected ICT data. A platform that interfaces with other services such as road status and geographical information is introduced. For example, how data gathered from these services will be processed to be either communicated to the on-board energy management system or as recommendations to be provided to the driver (see Fig. 1). The main objective is to increase the cooperation between ICT services such as wireless communication and embedded technologies (e.g., sensors, actuators, computing

hardware/software, web services) and vehicles to sense and gather their context information and allows in-vehicle EMS to make suitable decisions based on that information. Therefore, an intelligent energy management system has to be developed. It will allow, for example based on the current state of the battery, the intended itinerary, and the geographical information, to select the best suitable energy source to the current situation.

battery reaches its minimum state-of-charge, another source or an engine could be activated to propel and recharge batteries. HEV has an advantage over FEV since recharging the battery in a recharging point is not required and hybridization allow improving fuel economy by 50% [4]. PHEV technology is recently introduced into the market and batteries can be recharged by plugging into an electric power source. The aim is to use HEV mode operation for long distance journey (e.g., highway) and FEV mode operation for short journey (e.g., urban). PHEV has large battery pack that can be charged either by an onboard engine, regenerative breaking of motor or external electric supply [5]. For example, Opel has developed an energy management system to regulate the interaction between the electric motor, gasoline engine, generator and battery.

Figure 1: Scheme for energy optimization of hybrid vehicles

In Ampera/Opel, the driver can select one of four drive modes [1]: normal mode, sport mode, mountain mode, and hold charge mode. Under the first mode, the gasoline engine is automatically started to keep up the battery charging. This process starts when the energy level of the battery drops below a certain level or under breaking. In the sport mode, the vehicle becomes more responsive to movements of the accelerator pedal. The mountain mode will be selected by the driver to guarantee the necessary power for hilly roads by raising the battery’s charging state. More precisely, the driver should activate this mode immediately after a full charge or 10 to 15 minutes before getting to hilly roads. Hold charge mode is suitable in urban areas or restricted emission zones. Selecting this mode allows a pure electric driving using the energy saved in the battery.

A test scenario for the platform will be dealing with restricted TLZ (traffic limited zones), called also geofences. Geofences are virtual perimeters based on defined geographical areas such as forbidden areas for circulation, residential zones, areas with limited speed. In this scenario, if restricted areas are on driver’s itinerary, a best strategy, which allows crossing these areas by using an electric power mode only, will be recommended to be followed either manually by drivers or automatically by the EMS. If the battery level does not allow crossing these areas, recommendations, using dynamically re-routing strategies to bypass them, will be provided to the driver. It is worth noting that some public authorities in European cities have considered the necessity to define TLZ in city centers: areas forbidden to vehicles powered by internal combustion engines but allowed for vehicles using electric power. The remainder of this paper is structured as follows. Section 2 presents a state of the art review of exiting technologies and on-board energy management solutions. The architecture of the platform and a brief description of main components will be presented in Section 3. A test scenario is presented in Section 4 together with preliminary results. Conclusions and future work are given in Section 5. 2. RELATED WORK Recently, industry is making very great research effort to further develop the power engine of electric vehicles and the batteries. Three major types of vehicles are available in the market: full electric vehicles (FEV), called also pure batterypowered vehicles, hybrid electric vehicles (HEV), and plugin hybrid electric vehicles (PHEV) [1, 4]. Because batteries have a relatively low energy density, FEV are primarily suited for urban areas, e.g., short journey with limited ranges. Depending on the batteries used the ranges are from 50km to 80km, but for long journeys (e.g., weekend trips) heavy and bulky batteries that need several hours for charging are required. Furthermore, batteries have to be efficiently used since the propulsion of the vehicle depends on their energy storage capacity. HEV technologies have been developed to overcome the limitations of FEV in order to extend range capability. If the

Switching manually to the suitable mode requires a great attention from drivers and full knowledge about driving context. Furthermore, the synchronization of multiple energy sources under uncertain factors and constraints, e.g., environment conditions and driver behavior is the most important issue. To overcome this issue, intelligent energy management systems are developed to allow automatic switching between sources according to current driving conditions. Two distinguished families of control strategies have been proposed: rule-based control and optimal control [4,5,6,7,8,9]. Rule-based controls are heuristic-based strategies, such as fuzzy logic and neuro-fuzzy, and based on predefined driving cycle. Optimal control strategies are based on optimizing a cost function of the system using techniques such as linear programming, neural networks, and genetic algorithms. However, little attention has been paid to integrate other information into the switching mechanism. Examples are data gathered from drivers’ actions and preferences (e.g., driving behaviour, itinerary), traffic information, and road profile. For example, in [10] it was stated that the road condition, speed limit, and traffic light distribution, real-time monitoring using GIS (Geographical Information System) and GPS (Global Positioning System) data can be used to generate better decisions. In this paper, a framework for energy management system that interfaces with on-board components and ICT services for efficient energy use is introduced. The framework architecture is presented and a

scenario is used to show its usefulness in avoiding TLZ; areas forbidden to vehicles powered by internal combustion engines but allowed for vehicles using electric power. 3. THE PLATFORM ARCHITECTURE

addition to those provided by road service providers, in order to generate recommendations and decisions related to the driver context. The platform provides, then, the driver with this information to make decision. All these components are detailed in the following sub-sections.

The architecture of the platform is depicted in Fig. 2 and shows the different components required to allow, based on the intended itinerary, the driving behaviour, the current traffic situation, and other geographical information (e.g., restricted areas), selecting the best suitable energy source to the current situation. The platform makes usage of available road information to provide IEMS with relevant data to make a decision in choosing the suitable energy depending on different parameters (e.g. green zone, battery level). To get the information provided by road entities (traffic information centre or management centre) the platform is based on Web services and SOA architecture [13]. This emerged architecture has shown, in the past few years, great benefits for the development and integration of systems over Internet, where functionalities are packaged as independent interoperable services. Furthermore, SOA splits and packages features in distinct entities, i.e. services that developers make accessible on a network, so that they can be reused and combined to develop new services over the Web. A Web service is a W3C standard defined as a software system designed to support interoperable machine-to-machine interaction over a network [2]. Web services use a standard XML format to exchange information and they are not tied to a particular operating system or programming language. Web service architecture is composed of three main entities: the service provider, the service broker and the service requester [13]. The service provider hosts a Web service, which can be invoked by the service requester (i.e. client) through requests formatted in SOAP (Simple Object Access Protocol). To satisfy the need for Web service discovery, the service provider should register its service in the service broker. The service broker is actually a service repository usually UDDI (Universal Description Discovery and Integration), which can be interrogated by service requester to find services that match their requests. The platform is accessible via an Internet connection by using wireless communication technologies [3] associated with GNSS (Global Navigation Satellite System), which allow now individuals and professionals to access information services over Internet wherever they are and whenever they want. This architecture consists of the PHEV in-vehicle embedded system, the HELECAR platform, and the top layer consisting in different road service providers. The in-vehicle embedded system interfaces with IEMS and the platform by sending periodically relevant data such the status of the battery, the driver itinerary, and the current geographical location. HELECAR platform uses all data sent by vehicles, in

Figure 2: The architecture of the HELECAR platform 3.1 HELECAR platform The HELECAR platform allows orchestrating all the road participants (e.g. drivers, services, traffic data centers) to optimize the vehicle energy use depending on the crossed areas. In other words, the selected energy mode will be done in accordance to the crossed characteristic areas. This information is provided by the road services that provide the platform with the traffic information, the road status, and the real time area behavior. All these data are stored in database and frequently updated. The platform allows tracking and monitoring vehicles in order to inform them about the current status of the TLZ in order to use the suitable energy in accordance to the zone characteristics. To do so, different data are gathered from invehicle embedded system and then are sent to the platform. In [11], a data-logger software was mainly developed and uses a diagnostic interface to plug into the vehicle existing network (i.e. CAN bus) to extract relevant data. Data are temporarily stored in an embedded computer memory and analyzed to be used for warning the driver about dangerous situation. Data also uploaded into the infrastructure using V2I for filtering, processing and storage for eventual usage. In this paper, data to be gathered for the CAN bus are for example battery status and the driver behavior. Other data such as the geographical location and the itinerary can be extracted from GPS devise and embedded computer. The system could use the driver’s itinerary to anticipate the energy power to be used. For example, if the current battery level is low, then the energy mode will be switched to the combustion power and a re-routing decision will be made to avoid green zones and all areas that do not allow the use of this kind of power.

3.1.1 Exchanging message format

3.2 Intelligent EMS

The exchanged message format used by the platform is depicted in Fig. 3. As illustrated, the message consists of five fields. The message type defines the type of the sent message; e.g. ITI for itinerary message, or BAT for battery status. The sender or the destination ID is used to identify the vehicle in both cases V2I and I2V communications. The GPS location is the real time location of the vehicle when the message is sent. The timestamp is the date of the message. Finally, the payload data contains the data (e.g. itinerary or a status of the battery).

A hybrid vehicle is equipped by a control system or IEMS. This system depends on the hybridization considering the sources of energy, the converters and the storage system. A suitable management must take into account the constraints on the sources and the kind of storage system. Figure 5 shows an example of control for a hybrid system composed of battery and Supercapacitor. Here, the load can be interpreted as the needs of an electrical engine. Notice that the command of converters can be done by heuristic approach or is based on methodologies of automatic control (PID, sliding mode, adaptive control ...). In the second case, the constraints have to be formulated mathematically.

Figure 3 : The message format The exchanging messages between the platform and vehicles have to be formatted as illustrated in Fig. 3. It is worth noting that, the vehicle gets its assigned itinerary and send it to the platform to be recorded on the database, and periodically, the current location and battery status are sent to the platform. 3.1.2 The TLZ service provider To show the effectiveness of the use of HELECAR platform, we have used TLZ Web service, called also Geofencing service [2], which provides information related to the nearest zones to the location of tracked vehicles. So, we have implemented a web application (see Fig. 4) to define different characteristics of the monitored areas (i.e. geofences).

Figure 5: Example of hybrid system with control In this work, new constraints, coming from information system, must be considered. For instance, for a hybrid vehicle of type thermal/electrical, the information concerning the position of the vehicle with the nearest petrol station, road profile, and driver behavior, must be taking into account in the strategy of management. Thus, automatic control methodologies can be used to manage the energy aboard vehicle based on references and constraints. These constraints and references can change in real time according to available information. Figure 4: TLZ Web application All the TLZ areas are defined and registered in MySQL database. The stored data are then provided when the TLZ service is invoked. This service takes as input the current location of the vehicle and its battery level and type.

Among the possible approaches based on mathematical models, one can cite discrete event methodology that can be suitable to tackle to situations with sudden changes in the road (e.g., traffic jam, accidents). Indeed, in this approach, the mathematical models take into account the occurrence of events and switch to the suitable models corresponding to the situation. Besides, another approach should be adaptive control suitable to handle unknown parameters that can reflect new situations. In fact, in this approach the controller evolves according to the current information.

4. SIMULATION RESULTS Recently, public authorities in European cities have considered the necessity to define TLZ in city centers: areas forbidden to vehicles powered by internal combustion engines but allowed for vehicles using electric power. The scenario tested is as follows. If restricted areas are on driver’s itinerary, a best strategy, which allows crossing these areas with only an electric power mode, will be recommended to be followed either manually by drivers or automatically by the IEMS. Outside these areas, different cases can be considered depending of the type of vehicle. For example, a PHEV vehicle can be operated by the internal combustion engine to charge the battery, and if the autonomy is sufficient inside these areas, the electric motor is operated by battery. If the battery level does not allow crossing these areas, recommendations, using dynamically re-routing strategies, to bypass them will be provided to the driver. To show the effectiveness of the use of HELECAR platform, a prototype is developed and implemented using Java EE5 and runs on Glassfish application server. The development was made using Netbeans 6 IDE, which includes SOA development tools to develop both Web service and Web service clients using JAX-WS API. These Web services can then be deployed on the Glassfish application server for testing and evaluation. The Google Map is used to display the real time location of vehicles and different TLZ areas.

It is worth noting that we have developed different services [2] (e.g. Geofencing softwares, snow clearance software, etc.) in the European ASSET 1 project. We have used this initial work and we had improved these services in order to meet the HELECAR platform’s requirements. The primary experiments of this platform have been conducted. Three vehicles were considered and their itineraries are provided and defined. These vehicles send to HELECAR platform their current geographical location and their battery’s level. We note that at this current implementation of the software, we are focusing on generating and sending information about the nearest TLZ areas according to the received data from vehicles. This information is sent then to the related vehicle. Fig. 6 illustrates the obtained results and shows the real time positioning and exchanging information between the platform and vehicles. The provided information will allow the IEMS to choose the optimized energy to be used with respect to the vehicle environment. An adaptive control will be developed and will use this information to select the best suitable energy source according to the current context. 5. CONCLUSIONS AND PERSPECTIVES In this paper, the architecture of the HELECAR platform is introduced to highlight the usefulness of increasing the synergy between ICT services and PHEV for automatic management of the energy. An IEMS control strategy is briefly described to show how other external information could be used depending on the driver context and road information to select the best suitable energy source. The needs and the requirements of such platform are clarified and provided. Future work deals with including a complete prototype that will be tested and evaluated by using the automatic control and commands to allow an automatic switching of the energy mode taking into account different environmental constraints and driver behaviour, and by experimenting and simulating the platform in real environment. Furthermore, re-routing process will be taken into account in case the access to a TLZ needs the use of electrical battery and the vehicle has not enough autonomy. Other scenarios will be considered, for example city busses that operate on same itineraries with fixed time timetables. However, several unknown parameters could be considered such as possible stops based on passengers’ requests and itineraries profile data. The IEMS could adapt to different operating situations to minimize fuel consumption while taking into account traffic limited zones. Different control strategies that allow the selection of the best suitable energy source will be studied and virtual test runs will be carried out to test their suitability for day-to-day use with a multitude of vehicles and requests to the platform.

Figure 6: Real time positioning of PHEV vehicles 1

ASSET project website: http://www.project-asset.com/

REFERENCES [1] Electrifying the Automobile–Variations, Strategy for the Electro-era, Magazine ADAC, April 2011, 19 pages. [2] W. Ait-Cheik-Bihi, A. Nait-Sidi-Moh, M. Bakhouya, J. Gaber, and M. Wack. TransportML platform for collaborative Location-based services. In Service Oriented Computing and Applications journal, 2012, DOI:10.1007/s11761-012-0114-2. [3] K. Dar, M. Bakhouya, J. Gaber, M. Wack, P. Lorenz, Wireless Communication Technologies for ITS Applications, IEEE Communications Magazine, 48:5, 156-162, 2010. [4] Hamid Khayyam, Abbas Kouzani, Saeid Nahavandi, Vincenzo Marano and Giorgio Rizzoni, Intelligent Energy Management in Hybrid Electric Vehicles, Energy Management, Edited by Francisco Macia Perez, ISBN 978-953-307-065-0, Publisher: InTech, 2010, pp. 147175. [5] Harpreetsingh Banvait, Xiao Lin, Sohel Anwar, Yaobin Chen, Plug-in Hybrid Electric Vehicle Energy Management System using Particle Swarm Optimization, World Electric Vehicle Journal Vol. 3 - ISSN 20326653, 2009. [6] Y. Gao, M. Ehsani, Design and Control Methodology of Plug-in Hybrid Electric Vehicles, EEE Vehicle Power and Propulsion Conference (VPPC), September 3-5, 2008, Harbin, China. [7] L. Sun, R. Liang, Q. Wang, The Control Strategy and System Preference of Plug-in HEV, IEEE Vehicle power and Propulsion Conference (VPPC), September 3-5, Harbin, China. [8] S. Moura, D. Callaway, H. Fath, J. Stein, Impact of Battery Sizing on Stochastic Optimal Power Management in Plug-in Hybrid Electric Vehicles, IEEE International Conference on Vehicular Electronics and Safety, Sept 22-24, 2008, Columbus, OH. [9] Qiuming Gong, Yaoyu Li, and Zhongren Peng, Power Management of Plug-in Hybrid Electric Vehicles Using Neural Network Based Trip Modeling, 2009 American Control Conference, pp. 4601-4606. [10] Q. Gong, Y. Li, Z. Peng, Optimal Power Management of Plug-in HEV with Intelligent Transportation System, Advanced Intelligent mechatronics 2007 IEEE/ASME international conference, 4-7 Sept, 2007. [11] S. Cai, M. Becherif, M. Bakhouya, M. Wack, A Context Aware Embedded System for Intelligent Vehicles and Smart Roads In: ACM ICPS, UPC Workshop, 2009. [12] M. Bakhouya and J. Gaber. Service composition approaches for ubiquitous and pervasive computing environments : A survey. In Eldon Li and Soe-Tsyr Yuan, editors, Agent Systems in Electronic Business, pages 323–350, Information Science Reference/IGI Publishing, 2008. [13] Wu C, Chang E, State-of-the-art web services architectural styles, Proceedings of the Third IEEE European Conference on Web Services (ECOWS), Vaxjo, Sweden, 2005.