As robots become more ubiquitous and more intelligent, more people will interact with them. But how do you design these complex systems in such a way that humans can comfortably interact with those machines? Preceding a workshop on this theme at ERF 2024, Rocking Robots discussed human-robot interaction with Valeria Villani, a researcher at the University of Modena.
“My focus is on integrating human factors into the broad concept of human-robot interaction”, says Valeria Villani, Assistant Professor at the Department of Science and Methods for Engineering, University of Modena and Reggio Emilia. “This multidisciplinary topic can be approached from various perspectives. The technological perspective involves making robots and complex systems work efficiently, focusing on technical aspects like accuracy, precision, and performance metrics. However, there’s much more to consider when interacting with humans.”
That aspect was remained largely unexplored until recently, but is now gaining attention. “This means focusing on making the interaction intuitive, sustainable, and acceptable for the user. It’s not just about performance metrics. While we can now create devices that are highly accurate and perform well, this doesn’t necessarily mean they are easy to interact with. There is a gap that needs to be bridged.”
Perceptions
She says the key to achieving good human-robot interaction and team performance is to consider the interaction from the user’s perspective. “It’s not sufficient to focus only on the technical aspects. It’s also crucial to understand how people interact with these systems, their preferences, and their perceptions when they face a complex technological system, whether it’s a robot or another interactive device. The challenge remains the same across different types of systems.”
“For example, in our lab, we have a robot dog. From an engineering standpoint, it’s a quadruped with advanced locomotion skills. It’s essentially a robot – a collection of hardware designed for efficient movement. But when we present this robot during public demos or open days, people’s reactions vary widely. Some are frightened by it, perhaps because it resembles a dog but behaves unlike any natural animal. Others, including both children and adults, attempt to pet it, expecting animal-like behavior, despite it being just a machine. They might try to establish an emotional connection, attributing to the robot characteristics it doesn’t possess.”
AI
This is a clear example of how the perception of technology can differ vastly from its intended design or functionality. This phenomenon extends to artificial intelligence as well. “There’s a lot of hype around AI, with portrayals in media suggesting that these machines can feel or generate emotions. However, from a technical standpoint, this isn’t accurate. They don’t experience emotions; they are programmed to respond in ways that mimic emotional responses, based on their training. The public’s perception, influenced by how AI is portrayed in the media, often overestimates its capabilities.”
This disparity in perception is a critical aspect of human-technology interaction. It highlights the importance of understanding and managing user expectations and perceptions in the development and presentation of technological systems. “My research primarily focuses on measuring the physiological condition of the user while interacting with the system. The goal is to enable robots to respond appropriately to the user’s physical and emotional states without explicit communication.”
Adaptive behavior
“For instance, if two people are working together on a task and one is feeling tired or unwell, it’s usually easy for the other to recognize this and adjust their effort accordingly. In my research, I aim to replicate this kind of adaptive behavior in human-robot interactions.”
Her background in biomedical engineering helps in applying knowledge of physiological responses to human-robot interaction. “With the advent of wearable devices like smartwatches and chest bands that can monitor physical activity, it’s possible to record the user’s physiological activity during interaction with robots. Medical literature indicates that changes in physiological activity can signify fatigue, stress, high workload, etc. By detecting these changes, the robot can adjust its behavior to accommodate the user’s current state, mirroring the adaptive nature of human-human collaboration.”
“This information becomes a quantifiable variable that can be communicated to the robot. For instance, if I become fatigued, as indicated by physiological monitoring, this information can be transferred to the robot. The robot can then adapt its behavior, similar to what would occur in human-human collaboration. Depending on the type of robot and the nature of the interaction, the robot could slow down or become more autonomous in its tasks, allowing me time to relax and restore my normal physiological condition.”
Privacy
Understanding the human condition during interaction tasks can be utilized both during the design of the interaction system, and while using the system, where the robot implements adaptive strategies or behaviors based on the user’s current state.
Of course, privacy is an important aspect. “First, it’s important to clarify that it’s not the robot itself but the system, which might include a software running on a laptop, that is monitoring the user’s condition. We must ensure that we don’t mislead people into believing that robots can autonomously monitor or understand our emotions, as this aligns with the broader AI concerns we discussed earlier.”
“Nevertheless, the issue of comfort and privacy remains complex. Gaining user trust and consent is critical, especially when dealing with sensitive data like physiological responses or brain activity. Users need to be aware of what data is being collected, how it will be used, and the benefits of such monitoring in terms of improved interaction and system responsiveness. Addressing these ethical and privacy concerns is as crucial as the technological development itself in the field of human-robot interaction.”
Challenge
The biggest challenge in human-robot interaction is to create robots that are not only technically efficient but also socially acceptable and user-friendly. “This means integrating aspects of design that make the robot feel less threatening and more approachable, ensuring that it aligns with human needs and societal norms. It involves a multidisciplinary approach, incorporating insights from psychology, user experience design, and ethical considerations, alongside the technical aspects of robotics.”
The goal is to develop robots that not only perform tasks efficiently but also fit seamlessly into human environments and interactions, enhancing the user’s experience and contributing positively to society. “The dilemma is whether the gap between humans and technology should be fully bridged. The current trend is to make technological agents like robots behave and be perceived as human-like as possible. However, it’s worth considering whether maintaining a certain gap between humans and technology might be more appropriate. This is a profound question that touches on the essence of social acceptance and the role of technology in our lives.”
At ERF 2024 this topic will be discussed at a workshop:
