Periodic Reporting for period 2 - ENDORSE (Safe, Efficient and Integrated Indoor Robotic Fleet for Logistic Applications in Healthcare and Commercial Spaces) Periodo di rendicontazione: 2020-10-01 al 2022-09-30 Sintesi del contesto e degli obiettivi generali del progetto Commercial indoor spaces including hospitals, hotels, warehouses, offices, etc. offer great potential for commercial exploitation of logistic robotics. Such spaces provide inherent advantages for robotics solutions deployment, since they are required by law to meet stringent building codes, and therefore the navigation space exhibits some structure, which favours autonomous navigation. In addition, they offer reliable communication infrastructures, a prerequisite for normal business operation. Thus, commercial spaces are rightfully considered the next great field of logistic robotics deployment. Despite these advantages, today, only a few solutions exist, which have not triggered widespread acceptance by the market. This is because existing systems require costly infrastructure installation (arrays of peripheral sensors, mapping, etc.); they do not easily integrate to corporate IT solutions and as a result, they do not fully automate procedures and traceability; they are limited to a single type of service, i.e. transfer of goods. Through transfer of knowledge, multidisciplinary research and cross-fertilization between academia and industry, ENDORSE has addressed the aforementioned technical hurdles. Four innovation pillars were pursued: (i) infrastructure-less multi-robot navigation, i.e. minimum (if any) installation of sensors and communications buses inside the building for the localization of robots, targets and docking stations; (ii) advanced HRI for resolving deadlocks and achieving efficient sharing of space resources in crowded spaces; (iii) deployment of the ENDORSE software as a cloud-based service facilitating its integration with corporate software solutions such as ERP, CRM, etc.; (iv) reconfigurable and modular hardware architectures so that diverse modules can be easily swapped.The modular architecture was successfully demonstrated and validated by the integration of an e-diagnostic support module, equipped with non-invasive sensors/devices, and the Electronic Health Records (EHR) interfacing, thus serving as an autonomous mobile e-health diagnostic robotic module.The latter mobile e-health diagnostic module allows remote patient monitoring and vital signs acquisition and facilitates full connectivity to a dedicated inpatient-monitoring platform. In addition, the data produced by the e-diagnostic module diagnostic exam report are securely forwarded to update and complement the patient’s cloud-based EHR. The integrated ENDORSE solution comprising a fleet of autonomous mobile robots and the ENDORSE Fleet Management System has been deployed and successfully tested in full-scale lab-controlled environment replicating typical health centres. The evaluation conducted emphasised the good performance of the ENDORSE solution as regarding time-to-deployment and operational dependability. Lavoro eseguito dall’inizio del progetto fino alla fine del periodo coperto dalla relazione e principali risultati finora ottenuti The consortium partners first defined the user requirements, the use-cases and the technical specifications of the ENDORSE solution. They then adopted an iterative development-integration-prototyping-testing strategy to produce finalized versions for the software and hardware subcomponents of the overall ENDORSE solution towards achieving the objectives of the action. Ethics requirements were thoroughly addressed.The consortium developed the innovative mobile e-health diagnostic module (both hardware and software components) to allow remote patient monitoring and vital signs acquisition, as well as the integration software to access the Electronic Health Record (EHR) from the mobile e-diagnostic module. The EHR builds on the concept of the Patient Summary (PS), an EU wide initiative to harmonize the sharing of medical data across member states and thus to facilitate cross-border healthcare provision. Extending the range of facilitated logistic applications, the consortium adapted an autonomous mobile robot by including a lifting unit and an elevation platform that allows the transport of trolleys. Moreover, to optimize the range of services offered, ENDORSE partners developed a cloud-based fleet management system for multi-robot task allocation and planning as well as cost-effective fleet operation supervision. In terms of enhancing Human-Robot Interaction, algorithms for intuitive and transparent human-aware navigation capabilities were developed, leveraging deep-learning techniques for object detection and model predictive control, as well as infrastructure-less solutions for robust global localization for multi-robot navigation using square fiducial markers detection and the integration of the Anyplace Indoor Navigation Service into the Robot Operating System (ROS).The results achieved were documented and published in scientific publications. Besides, prospective exploitation plans were elaborated for mature ENDORSE components. Progressi oltre lo stato dell’arte e potenziale impatto previsto (incluso l’impatto socioeconomico e le implicazioni sociali più ampie del progetto fino ad ora) A questionnaire was elaborated to elicit the requirements of the stakeholders regarding a logistic robotics system for healthcare and commercial spaces. It served as a basis to understand the opportunities and challenges of such robotics solutions from a nursing professionals’ perspective. The findings went beyond the state of the art and were documented in an article published in a peer-reviewed international journal. Likewise, the fleet management system uses several scalable optimization algorithms for robot task allocation that advance the state of the art, which were also published in a peer-reviewed international journal. In that sequence, innovative algorithms enhancing the robustness and stability of indoor robot navigation in dynamic environments were elaborated using interval analysis or the evidential grids framework. Besides, state-of-the-art deep-learning methods were used to endow the robots with the capability to detect the presence and identify the type of dynamic obstacles inside their visual field and adapt their navigation accordingly. In addition, an adaptive video delivery methodology was studied to enhance quality of experience and aid towards the wider adoption of wireless medical video communication systems in standard clinical practice. All these technical results go beyond the state-of-the-art; they were published in the proceedings of international peer-reviewed scientific conferences.The ENDORSE mobile e-health diagnostic robotic module minimizes physical interaction. Therefore, it allows risk-free remote patient monitoring and vital signs acquisition that are securely forwarded to update and complement the patient’s EHR. The ENDORSE mobile e-health robot may also relieve hospital staff as it facilitates their daily tasks pertaining to physiological measurements monitoring. Finally, the ENDORSE solution offers a cost-effective multi-robot solution for logistics and health services in the healthcare and commercial spaces. The three types of ENDORSE robots The ENDORSE mobile e-health module The ENDORSE Electronic Health Record GUI
