Face recognition in Service robotics: Analysis of the padding effect according to people age / Reconhecimento facial em robótica de serviço: análise do efeito de preenchimento de acordo com a idade das pessoas

Service robots usually perform repetitive tasks such as collecting garbage, cleaning the house, among others. This kind of robot needs different skills to perform its daily tasks, being people´s recognition a critical skill. One of the techniques used to improve face recognition is padding. The padding technique increases, by a given scale factor, the bounding box of a detected face. In previous work, we had presented a comparative analysis of the influence of the padding in the algorithm used for face recognition. This paper extends the previous analysis by considering the effect of various padding scale factors among different life stages (i.e., toddler, children, teenager, adult, senior, and golden oldie). The result of this analysis shows that increasing the bounding box of detected faces is less efficient for middle-aged people than for younger and elderly people.

Inter-Task Similarity for Lifelong Reinforcement Learning in Heterogeneous Tasks

Reinforcement learning (RL) is a learning paradigm in which an agent interacts with the environment it inhabits to learn in a trial-and-error way. By letting the agent acquire knowledge from its own experience, RL has been successfully applied to complex domains such as robotics. However, for non-trivial problems, training an RL agent can take very long periods of time. Lifelong machine learning (LML) is a learning setting in which the agent learns to solve tasks sequentially, by leveraging knowledge accumulated from previously solved tasks to learn better/faster in a new one. Most LML works heavily rely on the assumption that tasks are similar to each other. However, this may not be true for some domains with a high degree of task-diversity that could benefit from adopting a lifelong learning approach, e.g., service robotics. Therefore, in this research we will address the problem of learning to solve a sequence of RL heterogeneous tasks (i.e., tasks that differ in their state-action space).

Combining Reinforcement Learning and Causal Models for Robotics Applications

The relation between Reinforcement learning (RL) and Causal Modeling(CM) is an underexplored area with untapped potential for any learning task. In this extended abstract of our Ph.D. research proposal, we present a way to combine both areas to improve their respective learning processes, especially in the context of our application area (service robotics). The preliminary results obtained so far are a good starting point for thinking about the success of our research project.

Analyzing the features of service robotics development and distribution during the coronavirus pandemic

Subject. The work analyzes the features of service robotics development and distribution during the coronavirus pandemic, using the case of delivery drones and UVC disinfection robots. Objectives. The aim is to identify the main trends in the development of certain types of service robots during the COVID-19 period, analyze technical and economic characteristics, assess assistive technologies for the future development of robotics. Methods. The study draws on methods of correlation and regression analysis. To assess the development trends of certain types of service robots, we apply website data of manufacturers of the equipment under consideration. Results. We unveil major trends in the development and production of certain types of service robots during the fight against coronavirus, using the case of unmanned aerial vehicles and disinfection robots. The development of delivery drones is aimed at solving the problems of flight efficiency, increasing the flight time, reducing the energy use and noise, and ensuring the safety of delivered goods. Medical robotics developers focus on streamlining the management interface, increasing the speed of task execution, enhancing the reporting abilities. Conclusions. Our analysis shows that the coronavirus pandemic affected the speed of development and implementation of robots in medicine. The market is shifting towards the production of UV-disinfection robots. The most important technical and economic characteristics of the main types of disinfection robots are the surface treatment cycle to achieve the maximum disinfecting effect, maximum operating time (without recharging), charging time, wavelength of radiation, weight of device.

Maritime Robotics and Autonomous Systems Operations: Exploring Pathways for Overcoming International Techno-Regulatory Data Barriers

The current regulatory landscape that applies to maritime service robotics, aptly termed as robotics and autonomous systems (RAS), is quite complex. When it comes to patents, there are multifarious considerations in relation to vessel survey, inspection, and maintenance processes under national and international law. Adherence is challenging, given that the traditional delivery methods are viewed as unsafe, strenuous, and laborious. Service robotics, namely micro aerial vehicles (MAVs) or drones, magnetic-wheeled crawlers (crawlers), and remotely operated vehicles (ROVs), function by relying on the architecture of the Internet of Robotic Things. The aforementioned are being introduced as time-saving apparatuses, accompanied by the promise to acquire concrete and sufficient data for the identification of vessel structural weaknesses with the highest level of accuracy to facilitate decision-making processes upon which temporary and permanent measures are contingent. Nonetheless, a noticeable critical issue associated with RAS effective deployment revolves around non-personal data governance, which comprises the main analytical focus of this research effort. The impetus behind this study stems from the need to enquire whether “data” provisions within the realm of international technological regulatory (techno-regulatory) framework is sufficient, well organized, and harmonized so that there are no current or future conflicts with promulgated theoretical dimensions of data that drive all subject matter-oriented actions. As is noted from the relevant expository research, the challenges are many. Engineering RAS to perfection is not the end-all and be-all. Collateral impediments must be avoided. A safety net needs to be devised to protect non-personal data. The results here indicate that established data decision dimensions call for data security and protection, as well as a consideration of ownership and liability details. An analysis of the state-of-the-art and the comparative results assert that the abovementioned remain neglected in the current international setting. The findings reveal specific data barriers within the existing international framework. The ways forward include strategic actions to remove data barriers towards overall efficacy of maritime RAS operations. The overall findings indicate that an effective transition to RAS operations requires optimizing the international regulatory framework for opening the pathways for effective RAS operations. Conclusions were drawn based on the premise that policy reform is inevitable in order to push the RAS agenda forward before the emanation of 6G and the era of the Internet of Everything, with harmonization and further standardization being very high priority issues.

Service Robotics in Nursing Care. The Preliminary Results of a Scoping Review

A scoping review was prepared in the first study phase of the REsPonSe project. The objective was to gain an overview of existing literature and available evidence on the subject of service robotics in nursing care. The systematic literature search took place using the following databases: PubMed, CINAHL, Cochrane, Web of Science and IEEE Xplore. The titles and abstracts of 2.761 references were screened based on the inclusion criteria. A total of 31 articles were identified as relevant for the study.