RAPP Deliverables
The PDF reports available on this page are all the public deliverables of the RAPP European funded project (FP7). They contain lots of resources that may be valuable for various projects and research in the field. These reports concern both the technical aspects of RAPP and the user perspective, e.g. RAPP Platform Final Requirements and KPIs, User requirements, Ethical legislations and guidelines, Image processing module, evaluation of RApps, etc.
A collection and analysis of requirements from the users’ perspective with emphasis on elderly end‐users. These requirements have refined the focus of RAPP and the specifications of the Platform
The current document is dedicated to the presentation of the adopted ontology schemes that will be utilized in the RAPP platform. One of RAPP’s central aspects (and aspirations) being to contain a means of storing, utilizing and transferring information among different robotic devices.
A description of the requirements and specifications necessary for the RAPP implementation as a software platform. A clear distinction is made between the different layers of the multi-tier architecture of the system, focusing primarily on service architecture.
The ethical issues considered in RAPP to ensure that users’ rights are fully guaranteed. The report includes ethical and regulatory concerns regarding research at European level while it also identifies important national Greek and Spanish legislation related to processing of personal data and Ethics Committee.
A report on the Artificial Intelligence methods developed in RAPP, aiming at providing the RAPP supported robots with means to expand their computational and functional capabilities. This approach is beneficial, as robots do not have to be equipped with expensive sensors or powerful computational units, since the RApps (Robotic Applications) can be executed in a distributed fashion both on the robot and the cloud.
The document presents the interaction between the RAPP platform and the robot based software. Usually robot controllers do not interact with external sources of software and have a fixed structure. This document presents an alternative approach: the architecture of a robot control system with exchangeable modules related to specific applications.
A description of the design and implementation of localization and mapping algorithms based on sensorial data, as well as efficient path planning algorithms.
This deliverable presents how to allow robots to perceive the indoor environment with an onboard RGB camera, supported by an off-line creation of a 3-D occupancy map using an external depth-map scanner (e.g. like the MS-Kinect device). The goal is to design efficient and reliable image pre-processing and segmentation functions, and symbolic concept detection and model-based object recognition algorithms, that can be used in real-world scenarios. A particular implementation of such RAPP API for the NAO robot is reported. The image processing module (and corresponding RAPP API) is structured into three layers, related to the API-s of the Core Agent, the Dynamic Agent and the RAPP Platform. Additionally, speech capture and speech synthesis functions, necessary for human-robot communication, are also implemented in the Core Agent API.
The document discusses requirements and proposes solutions for the integration of robotic and web application frameworks, the use of the HOP framework to support dynamic downloading and execution of RAPP applications on the robot, and the security model for the applications. The first section of the document lists technical requirements (network requirements, application distribution with an embedded part and a cloud part, dynamic installation of additional applications on the robot, integration of web communications with ROS communications, security). The other sections detail proposed solutions to these requirements. Several API have been specified to implement the solutions. Also, proof of concept code has been developed to demonstrate distributed communications and application downloading from the RAPP store. References to the API and POC code are given in the annex.