Applications of Robotics in Social Robotics

[Date of submission and author affiliation] Applications of Robotics in Social Robotics First A. Author, Fellow, IEEE, Second B. Author, and Third C. Author, Jr., Member, IEEE Abstract—The field of social robots deals with the design and development of robots that can interact with humans, thus advancing robots from being tools used in automation to partners with humans in vast areas including education, healthcare, entertainment, and scientific research. The area of social robotics depends largely on robotic technologies; another vast domain depended on other technologies such as big data, machine learning, artificial intelligence, and cognitive science. The growing needs of robots at home, industrial, and workplace companion has led to the design of smart robots that can interact with people using the natural language. However, challenges are designing and building such robots are still to be resolved. This paper discusses some of the applications of robotics in enabling the design of social robots such as human-robot interaction technologies, cognitive robotics, and behavioral dynamics and automation technologies. Index Terms—Artificial intelligence, cognition, cognitive robotics, robotics, social robotics INTRODUCTION R obotics research has advanced to the level where scientists can begin designing robots acting not only as tools, but also as partners with humans in different domains including healthcare, education, entertainment, household tasks, and scientific research (1). To act as a partner and not as a tool means that robots are configured to interact with humans in a way that is in line with human social psychology to understand human commands in accomplishing a task, assist humans when they need a helping hand, and give directions in learning novel ideas (2). In a study involving staff members, participants were asked if they preferred working with fellow humans to having a robot as a companion to assist them with tasks (3). Overwhelming answers pointed out that social robots are better companions when employees are unwell, and this poses a challenge to consider the role of social robots in the workplace as a new area of focus (4). This new breed of robots are referred to as sociable robots/socially interactive robots, or social robots and the long-term goal of social robotics in building competent robots acting as capable partners for humans is a daunting task (5). Social robots will need to communicate naturally with people through both verbal and nonverbal communication signals (6). The field of robotics has to ensure that social robots engage humans both at cognitive and emotional levels to provide appropriate social and task-related support to humans (7). In addition, social robots require an array of social-cognitive skills and mind theories to appreciate human behavior to be subliminally understood by humans (8). Therefore, the area of social robotics needs an understanding of human intelligence and how people behave across manifold dimensions such as social, physical, cognitive, and physical aspects is critical in building robots that can play important roles in people’s daily lives (2). To achieve these goals, the field of social robots needs a multidisciplinary approach where the development of social robots technologies will depend on robotics, artificial intelligence (AI), design, neuroscience, psychology, anthropology, and many more disciplines [9]. The contributions of robotic technologies in the design of social robots are the scope of this paper, although other interrelated fields in the area of robotics will also form an important discussion to appreciate the need of a multidisciplinary approach in designing effective technologies to address the complexities in social robotics. social robotics Social robotics is considered a disruptive technology projected to have profound impacts on healthcare, society, business, and the global economy (10). Social robots such as Sophia, created by Hanson Robotics, is designed purposely for education, research, and entertainment (11). She helps bolster public discussions about the ethics of AI and the future of robotics. The World Bank has supported many countries in the world harness the power of AI and other disruptive technologies (12). Sophia is one of the most popular lifelike robots globally and has been an ambassador in the application of AI for social and economic development. The design of social robots such as Sophia that accomplishes complex tasks requires a thorough understanding of robotic technologies such as artificial intelligence and automation (13). Robotic technologies help in the design, building, operation, and applications of robots and their computer systems to guide, provide sensory feedback, and process information (14). These technologies are used in developing machines capable of substituting humans or replicate people’s actions. A working definition of social robots refers to them as AI systems such as an android built to interact with humans and other machines (15). These machines can be controlled remotely either serving as a telepresence representative in homes or business meetings or as a health facility companion — a robot (16). The field of robotics involves the design, building, and use of machines used to execute functions that are traditionally carried out by humans. The rising cost of healthcare in the United States has seen robots being used in healthcare settings to care for the elderly (17). Such machines (robots) can also be used in some industries, including the manufacturing industry to carry out simple tasks that are often repetitive or perform functions in environments that are hazardous to humans (18). Many areas of robotics usually involve artificial intelligence, and robots may be built with human senses such as touch, vision, as well as the ability to sense a variation in temperature (19). Some robots are also capable of making simple decisions, but current developments in robotics are aimed at devising machines with some degree of self-sufficiency, which will allow them to move and make decisions in unstructured environments (20). While modern industrial robots do not look like humans, private investors and enthusiasts have created androids, which are the human forms of robots (21). Evolution in social robotics brought by machine learning has made robots to interact in sophisticated ways, thus gaining vast applications in society [3]. There is a growing demand for robots, which can interact safely with humans in daily activities and situations. Some robots need to anticipate the impact of their actions on humans and the actions and demands of the humans surrounding them (22). As an appreciation of social robots in solving societal challenges, the social consequences of robotics will depend on the degree on how humans employ robotics and the evolution of robots from a technical standpoint (4). It is, therefore, necessary that engineers fascinated in cooperating with sociologists become aware of social work and social service problems as well as sociologists interested in robotics have a closer look at technical issues of this new aspect of robotics. An enabling technology of AI has been integrated into robotics for carrying out the tasks of perception. With the modern robots able to sense the environment through computer vision or integrated sensors, the perception has not only become a critical area of research in robotics to create an artificial sense of self-awareness among robots but also support interactions of robots with their immediate environment. With the field of robotics undergoing rapid evolution, scholars project that it will soon demand a systematic collaboration between sociologists and engineers. Two overarching concepts in the field of social robotics are persona and scenario, which are already commonplace design tools in the area of Human-Computer Interaction (HCI) and cognitive robotics (23). HCI aims to improve the perception of humans in robots such as understanding emotions, activities, non-verbal communications and is capable of navigating the environment alongside humans. Currently, an approach based on social robotics has been integrated into Human-Robot Interactions (HRI), and as robots become more and more complex, engineers will require the help of trained psychologists and sociologists to create scenarios and personas, and train humanoids different behaviors based on settings. Other types of social robots can represent autonomous systems with in-built local AI allowing them to be autonomous following different cues from humans and stimuli in their settings (24). Such autonomous robots are sometimes referred to as smart robots, which are based on cognitive computing frameworks to simulate human thought processes. The field of cognitive robotics/computing gains from the latest advancements in machine learning that use pattern recognition, data science, and natural language processing in mimicking the working of the human brain. In an online survey seeking to identify different applications for social robots, participants provided responses based on the outlook of four social robots: AIBO, BARTHOC, BIRON, and iCat, as well as the information they had gathered in regards to what they could accomplish (4). Human-Robot Interaction (HRI) Technologies in Social Robotics Human-robot interaction (HRI) technologies in robotics continue to make progressive advances in ensuring the design of better robots that can socially interact with humans. For the robots to develop the necessary social skills, they will need to engage various interaction dynamics conveying social meanings. Durantin, Heath, and Wiles (25) term the events that occur within interaction dynamics as social moments. Social moments in social robots occur between social agents within multiple timescales and conversations as well as other consciously considered social interactions may span ranging from seconds to minutes or even longer. One of the problems that HRI technology can solve is the management of social exchanges that also depends on the manipulation and interpretation of fast timescales upon which such interactions are constructed. These timescales may be about 102ms (25). Designers of social robots have to understand the social importance of such fast interaction dynamics when robots are participating in any social interaction. For robots, the social moments need to be grounded in both personality and culture of the one interacting with the machine and in the characteristics of the interaction or the environment, the interaction task, as well as the role of the participants and the robot (25). Social robots will need special skills in detecting, treating, and understanding the meanings of the social moments with the interactant. Although the field of HRI and in general, social robotics have looked into the issues of social interaction, language, and social motion, the area of social moments during the interaction between robots and other agents as well as the short timescales typified by these interactions is an area that has little been explored (25). Social moments are known to be brief events occurring during the interactions between two or more agents having a potential to affect social dynamics. These moments have the possibilities of conveying semantic and pragmatic information at the time of interaction that needs to be a conscious or deliberate action. In a practical case scenario, if a lecturer is instructing a group of students and happens to take a glance on one of the students, but finds that the student is not paying attention perhaps out of boredom, the behavior of the lecturer will be affected by such an event. The lecturer will either decide to assume the bored student and focus his attention to the rest of the students or try to make the session even more interesting to capture the attention of the bored student. An alternative measure the lecturer can make is to decide and ignore completely anything to do with the bored student and avoid looking at the student any more for the remaining session of the lecture. In practice, the time is taken between the lecturer looking at the student and the student looking away can take a very short time, but it bears a social meaning, which can significantly affect the remaining part of the lecture (25). These sets of events are known as social moments, and it is critically important for social robots to be built with this challenge in mind. That is where the area of HRI technology comes in to solve the short social moments of interaction. Social robots used in healthcare or to serving as pets or education assistants will have to be built while considering this aspect. Social moments often induce performative meanings and can communicate either negative or positive valences. Communication intentions is a prerequisite of language abilities acquisition in humans, with performance meanings communicated in both verbal manners of non-verbal ways (25). For instance, the response speed in social exchange will define if one speaker is being answered or there is a generation of a new comment, and any small delays in responses can point to the state of the respondent being reluctant or eager to engage with the other agent. Longer delays are considered a violation of the social norms and may result in the interpretation that the interactant is anti-social, distracted, or offensive. Also, patterns of movement, especially those in peripersonal space, can relay social meanings. Different motions can also denote different intentions as well as changes in the body posture can imply different engagement levels. In their groundwork in social robotics, Dautenhahn Ogden and Quick hypothesized that as robot agents continue to become socially embedded, they need to have the capacity to observe their surroundings, learn from it, and find ways of adapting to their social environment (26). Importantly, social robots have to have the capacity to influence their immediate social environment. The authors specified the number of micro-behaviors to hand-notate the influence of robots on the humans that surround them at a temporal resolution of about 500 ms (25). In summary, the work of Dautenhahn Ogden and Quick hypothesized that social robots need to have the ability to sense, interpret, and construct social moments, but on smaller timescales compared to those of micro-behaviors. Cognitive Robotic Technologies in Social Robotics The modern robotic design aims to not only build robots to aid in manufacturing sectors but also introduce them as social partners. In light of the functions that modern robots have adopted such as RIKEN Japan’s Ri-Man in healthcare, iRobot’s Roomba in domestic helping, Sony’s AIBO as household pets, AIST’s PARO as emotional companions, and MIT’s Kismet and Leo as educational aids, appropriate social behavior is critically important for humans to develop a personal relationship with robots (27). Actors in robotic industries have called for the better design of agents that are capable of engaging humans in meaningful social interactions. Cognitive robotics is an area of robotics that comprises the control of physical systems designed to interact with unrestricted environments and control architectures, which explicitly consider the need to learn and use experiences. It is a multidisciplinary approach drawing on research in cognitive science, artificial intelligence, and adaptive robotics and exploits models centered on biological cognition. Social robots built with cognitive robotic technology achieve their objectives through the perception of the environment while paying attention to those events that only matter. The smart robots then plan the next action, while anticipating outcomes resulting from their actions and those of other agents. Next, social robots will learn from the entire interaction with environmental stimuli. To design a socially interactive robot, it must be able to deal with the obvious uncertainties of the natural environments through a continuous learning process, reasoning, a...