The Web of Things (WoT) aims at interconnecting network-enabled appliances using Web standards. However, Web protocols and languages are not adapted to those connected objects. There is also an emerging need for usages of meaningful services relying on interoperable objects. The major challenge of the ASAWoO project is to enhance appliance integration into the Web. Our project builds an architecture to provide users with understandable functionalities under the form of WoT applications, while enabling collaboration between heterogeneous physical objects. To reach this objective and guarantee our solution verifies additional properties, we combine advances from complementary disciplines: Semantic Web to reason about knowledge models; Service-Oriented Architectures to enable interoperability and scalable deployment from home environments to big organizations; Context-Aware Computing to make situation-based multi-level decisions, Multi-Agent Systems to enable autonomous collaboration between objects; Delay-Tolerant Networks to enable disconnection-tolerance for mobile objects; and Cloud Computing to enable minimize global power consumption through Green-IT and Green-by-IT awareness. In addition, our solution protects sensitive information carried by appliances through Privacy-awareness mechanisms. Our project embeds each physical object into an avatar that exposes physical functionalities as semantic Web services, supports optimized communication and code deployment, proposes collaborative functionalities with other avatars, and dynamically adapts the preceding points to changing situations. The project addresses several scientific locks: - provide an energy-aware cloud infrastructure to manage the lifecycle of WoT architectures supporting WoT application deployment - enable autonomous, collaborative avatar behavior, semantic description, discovery and composition of object capabilities, and context adaptation at the object, communication and functionality levels - design and implement energy- and privacy-aware interaction protocols and disruption-tolerant routing protocols for WoT objects - enable avatar/object protocol negotiation The ASAWoO project is organized in 8 tasks. First, Task 1 validates the interfaces and tools to be used in the project. Then, in parallel, Tasks 2, 3 and 4 respectively define the object architecture, cloud infrastructure and context-aware reasoning mechanisms. Task 5 focuses on semantic functionalities. Finally, Tasks 6 and 7 devise communication mechanisms and develop collaborative behavior. All along the project, Task 8 handles management and dissemination activities. We envision several benefits from the ASAWoO project. Our WoT infrastructure will provide a scalable, out-of-the-box framework enabling high-level interaction with sets of heterogeneous objects via the deployment of Web of Things applications. Moreover, the tasks deliverables represent independent advances in their domains. Assistance robots and co-workers (co-Bots) will be a major market in a near future. This project is an opportunity for vendors like Génération Robots to occupy a strategic market position with the development of advanced applications for connected objects that shall attract customers with easy application deployment and attractive interaction possibilities, leading to the creation of WoT application marketplaces, where software vendors will monetize these apps. For the LIRIS lab, this project concretizes existing work from the Web service and semantic Web research fields. For the IRISA-CASA research team, it leads to context-aware solutions for dynamic code deployment on constrained devices and disruption-tolerant routing protocols. For the LCIS lab, it allows defining new agent interaction models for physical objects. The project is also an experimentation field to validate the work on energy-efficient cloud infrastructures.

The PANAME (French acronym of Augmented Perception for Outdoor Navigation) consortium associates the experienced and agile SME named GENERATION ROBOTS, which holds a strong track in the development and delivery of concrete and mature solutions for terrestrial mobile robotics, and the public Research institute ONERA, which has a well-recognized capability to contribute to autonomous robot technologies and algorithms beyond the current state of the art. The SME brings its expertise in sensor and hardware integration for outdoor environments as well as more than 12 years of expertise in developing autonomous navigation algorithms embodied through the market release of a security autonomous robot 2 years ago. Together with a cutting-edge Research Lab, the consortium has strong ambitions to address the MOBILEX Challenge. The main objectives of the project will be to materialize and fill the reality gap for new perception and robotics technologies in realistic environments as follows: • The exploitation of complementary sensor modalities and their fusion, so as to monitor and optimise the environment perception with a measure of confidence, and also to identify situations where perception gets degraded through the evaluation of data integrity. Several new-generation sensors will be tested and integrated in order to improve the perception and interpretation of the surrounding environment of the autonomous robot. • The integration of the sensor information in a global consistent representation of the geometry and semantics of the robot environment. This will be based on parallelized advanced preprocessing of the different sensor tracks that will extract relevant information and labels to be integrated in a multimodal environment model, which will interact with the Trajectory and Decision modules. • The generation and execution of dynamic trajectories based on the aforementioned cartography, in association with a robust vehicle localization. Several trajectories will be evaluated in parallel with respect to specific criteria related to the mission objectives and constraints, the current model of the environment and the vehicle capabilities, but also considering opportunity information or complementary sensors to improve mapping or reduce perception ambiguity. • A decision mechanism that will interact through a dedicated innovative user interface where the operator will only be involved in high-level actions. This process will evaluate the collected status of the sensors and subsystems and decide whether the operator needs to be solicited to tele-operate the robot or provide high-level information to help the system continue its mission autonomously. A particular care will be put on comprehension and user experience. The deployment of these interacting algorithms is designed to be evolutive so as to adapt to the various steps of the MOBILEX Challenge and the related raise in complexity, as follows: • The sensor architecture is distributed and evolutive, to easily incorporate and exploit new modalities depending on the expressed needs. • The data fusion architecture is designed from the beginning to handle and integrate several sources of information in common representations. • The monitoring architecture is generic to gather the sensor and subsystems status and connect them to the decision mechanisms so as to adapt the system's actions and provide feedback to the operator accordingly. The project will rely on a well-established and shared project methodology, mostly based on (field) test and learn. At the end of the 3-year project, the main objective is to reach a system solution with a significant raise in system readiness by the integration of technologies beyond the state of the art, and opening the path to the industrial and end-user exploitation of these novelties. The protection and valorisation of intellectual property will be considered with great care in the process.