The evolution of assistive social robots

evolución de los robots sociales asistenciales

Robotics is transforming the way we interact with the world around us and with each other. The concept of a robot, though it may vary in interpretation among scientists, is generally defined as a programmable machine capable of manipulating objects and performing tasks that were previously exclusive to humans. This definition encompasses everything from physical electromechanical mechanisms to virtual software systems, all designed to simulate a capacity for thought or problem-solving. Although, for the most part, they execute commands dictated by humans. Within this vast field of robotics, emerge social robotics and the evolution of assistive social robots, whose primary purpose is to interact with humans and assist them in a natural and intuitive manner, as if they were living beings.

Principles to understand the evolution of social robots?

Obviously, social robotics is gradually integrating into society, allowing for an increasing scale of interaction between humans and robots. But what exactly does it mean for a robot to be social? Sociability implies that there are interactive relationships, where an agent is considered social if it can interact and exhibit communicative behaviors. Suppose two autonomous robots coexist in the same environment, whether directly or indirectly. This already generates aspects of social contact.

In this sense, the hypothesis of social intelligence postulates that to achieve intelligent behavior in an agent, it must be immersed in a physical and social environment. This agent undergoes dynamic and complex social interactions in the real world to develop intelligent behavior.

Therefore, the definition of social robots is a delicate issue, as it requires the conjunction of various factors linked to intelligence. In the current stage of the evolution of assistive social robots, their communication and interaction are still dependent on artificial intelligence.

In addition to being autonomous and having a pleasant appearance, the programming of these units should be easy for non-experts. They would even have to be tolerant of imprecise perceptions and actions and inherently safe. And, above all, they should be capable of learning and adapting to undefined and dynamic environments. In summary, a social robot is defined as one that interacts and communicates with people in a simple and pleasant manner, assuming social behaviors, patterns, and norms.

1940-1990, the first period of the evolution of social robots inspired by nature

In its beginnings, social robotics found inspiration in research on animal biological communities and the interactions among their individuals. Researchers’ fascination with nature led them to try to artificially recreate the interactions of biological communities, thus giving rise to social robotics in a primordial state.

Indeed, William Grey Walter‘s work was one of the first notable experiments in achieving robot interaction with the environment. In the late 1940s, Walter designed and built robotic tortoises that, through headlights at the front of the shell and positive phototaxis, interacted similarly to social beings. However, they did not show explicit communication or mutual recognition.

As the development of artificial life emerged, researchers applied principles such as stigmergy to achieve collective behaviors in robots. Stigmergy involves indirect communication between individuals based on modifications in the shared environment. In 1990, Jean Louis Deneubourg, a pioneer in stigmergy experiments, created robots similar to ants, thus establishing the foundations for studying the social behavior of insects through robotic models.

These principles were also applied in research on multi-robot systems or distributed robotic systems. They involved the use of interaction mechanisms such as communication, interference, and aggressive competition.

Kismet and humanoid robotics in the evolution of assistive social robots

Between 1990 and 2005, there were rapid and significant technological advances in the evolution of assistive social robots. One of these innovations was the creation of Kismet, one of the first social androids developed at the Massachusetts Institute of Technology (MIT). Designed by Dr. Cynthia Breazeal, Kismet was a pioneering experiment in affective computing, endowed with the ability to recognize and simulate emotions. This android head not only had input devices that provided it with hearing, vision, and proprioceptive abilities, but also simulated emotions through facial expressions, vocalizations, and movements. The interaction aimed to replicate the relationship between a caregiver and a child.

Meanwhile, in Japan, Honda developed a series of humanoid robots, from P1 to P4, culminating in the famous ASIMO. All this took place between 1996 and 2001. While the first four robots focused on humanoid robotics research, ASIMO stood out for exploring human-robot interaction, especially in the field of companion robots. ASIMO could recognize postures and gestures, react to them, distinguish sounds, and respond to questions through body movements or verbal responses. These advances marked a milestone in the integration of social robotics into everyday life.

In 2006, Aldebaran (belonging to the United Robotic Group) created the robot NAO, a 53 cm humanoid. NAO began to be used as an assistant in companies and healthcare centers to receive, inform, and entertain visitors. It also became a reference in the fields of education and research.

From 2009 to the present in the evolution of assistive social robots

The period from 2009 to the present represents a time of significant advances in social robotics. Technology available to researchers has far exceeded the limitations of the early decades, providing more advanced tools to research and develop more complex social robots.

Representing emotions has become a central aspect of human-robot interaction. Studies show that robots can mimic recognizable human expressions, establishing an emotional connection with people. Experiments with robots displaying expressions such as joy, sadness, anger, surprise, and fear have explored how emotional interaction can improve communication and relationships between humans and robots.

In the field of elderly and child care, social robotics has achieved significant milestones. From robots helping people with degenerative diseases take their medication to experiments in which social robots interact with autistic children to improve their imitation skills, the practical application of these advances is evident. A notable example is the experiment conducted in Germany, where a social robot guided an autistic child, correcting them until satisfactory results were achieved.

Examples of evolution in social robotics

Among the notable social robots, Pepper stands out, developed in 2014 by SoftBank Robotics (United Robotic Group). Its ability to perceive emotions and adapt its behavior to the user’s mood has made it a commercial project used in various countries to serve customers. Other robots, such as MAGGIE, Mini Maggie, and MBot from the Robotics Lab at Carlos III University of Madrid, have also contributed to the advancement of autonomy and intelligence in robots, focusing on human-robot interaction. All of them demonstrate a solid evolution of assistive social robots.

In 2018, the sixth version of NAO was presented, with a built-in CPU to improve its performance. NAO is the robot we use at Inrobics as part of our social robotics and AI-based rehabilitation solution.

Prospects for ongoing projects

Current advances in social robotics have led to the creation of fully developed and technologically advanced robots. Examples include assistive robots that play a crucial role in medical settings and homes for disabled or sick people. These robots facilitate healthcare tasks, accelerate emergency activities. However, in the near future, they could assist doctors in diagnosing by quickly accessing medical databases.

In homes of disabled individuals, social robots will not only administer drugs and therapies according to medical prescriptions but also assist in daily tasks, providing valuable help. Within environments such as nursing homes and pediatric units, social robots offer companionship, entertainment, and emotional support. Depending on this, they offer playful and conversational interactions, improving the quality of life for residents.

In terms of artificial intelligence, all future social robots will be equipped with learning capabilities to adapt to their environment and react appropriately in various situations. These capabilities will not only improve the autonomy of social robots but also allow their application in more specialized fields. For example, as surgical assistants, where they can advise and share the burden of the procedure with surgeons. At Inrobics, we are proud to be part of the evolution of assistive social robots, thanks to the continuous improvement and development of our social robotics and AI-based rehabilitation solution, the only one certified as a medical device in Europe. This solution is accessible to clinics, hospitals, and for private therapies in the comfort of home. Request a demo!

Picture of José Carlos Pulido

José Carlos Pulido

PhD in Computer Science and Technology, with Cum Laude honors from UC3M. MBA in Digital Health Management with a dual degree from OBS School and the International University of Catalonia. Over 10 years of experience in high-tech projects. He is an expert in Artificial Intelligence and Social Assistive Robots with a strong interest in digital health entrepreneurship. Associate professor at Carlos III University of Madrid. His studies and experience have given him a broad strategic vision of the sector. International career in the United States and Germany. Noteworthy for his commitment to innovation in the field of social assistive robotics. His motivation lies in his firm belief in the potential of technology to improve people’s quality of life and his desire to drive solutions that generate a positive impact on society.