Oscar Guerrero Rosado

I am currently pursuing a Ph.D. in NeuroRobotics at the Donders Institute for Brain, Cognition, and Behaviour (Nijmegen, Netherlands). My research activity aims to understand how complex systems self-regulate by generating goal-oriented autonomous behaviors. To explore this, I adopt a convergent synthetic methodology where biologically-constrained computational models are designed and embedded in robots to simulate cognitive functioning and reproduce intelligent behavior.

I believe that for artificial agents to self-regulate, their adaptation to environmental demands and opportunities must be grounded in the interoceptive state of the self. To account for both the internal and the external milieu together with the individual’s actions that bridge both domains, I consider Cognitive Architectures as the correct approach. These architectures guide the organization of computational models and implement self-regulatory principles, enabling behaviors such as reactive orchestration of internal needs, prediction and anticipation of future homeostatic errors, and the formation of cognitive maps that associate state-dependent stimuli with rewards.

Understanding how living systems manage to ensure their fitness by self-regulating in a complex world has also strong implications for Artificial Autonomy, Intelligence and Life. My contributions to HORIZON 2020 EU projects can showcase the suitability of self-regulation in industrial, clinical and social contexts. For instance, in HR-Recycler, collaborative robots disassembling WEEE material modulate their task, speed, or safety distance according to workers’ psychological preferences and physiological metrics. In ReHyb, a self-regulatory cognitive architecture allows stroke patients to individualize their neurorehabilitation training, meaning a serious rehabilitation game, an exoskeleton and a functional electrostimulation system match their parameters to the current cognitive and physical state of the user. More recently, I am contributing to the CAVAA (Counterfactual Assessment and Valuation for Awareness Architecture) project where motivational states are critical for modeling artificial awareness.

Background

My interest in understanding intelligence led me to study Psychology at the University of Granada. There, under the supervision of Julio Santiago, I had the opportunity to start my research activity by working on my bachelor’s thesis with the Grounded Cognition Lab. As a result of this fruitful collaboration, my thesis “Can our focus on the past put it in front of us?” was awarded with the “Best bachelor’s thesis” prize from the Official College of Psychology of Eastern Andalusia. Eager to complement my cognitive science background with technical expertise, I studied the Multidisciplinary Master in Cognitive Systems and Interactive Media in Pompeu Fabra University (Barcelona). There, I encountered the Synthetic, Perceptive, Emotive, and Cognitive Systems (SPECS) Lab, led by Paul Verschure. I joined the group first as a research assistant and later as a predoctoral researcher — an experience that deeply shaped my scientific mindset and allowed me to grow both personally and professionally within a vibrant, interdisciplinary environment.

Today, my academic journey continues at Radboud University, where I am completing my Ph.D. in NeuroRobotics under the supervision of Martin Vinck. My work bridges neuroscience, robotics, and artificial intelligence in an effort to build autonomous systems that not only act intelligently but are also capable of self-regulation, adaptation, and learning. As I look ahead, I am excited to continue my research career by further investigating the foundations of artificial autonomy. At the same time, I am eager to apply this knowledge beyond academia — collaborating with industry and innovation partners to develop intelligent systems that are both technically robust and ethically grounded. I believe that combining scientific insight with real-world application is key to shaping the future of meaningful and trustworthy intelligent robotics.