Advances in perceptual intelligence for autonomous robotic systems

Perceptual intelligence is a foundational pillar in robotics, enabling autonomous systems to interact safely and effectively with complex, dynamic environments. Recent advances in biological and engineered perception highlight the remarkable adaptability, efficiency, and resilience seen in nature and underscore the gains possible through interdisciplinary innovation. While bio-inspired models offer critical insights for embodied perception and action, robotics research today increasingly benefits from data-driven algorithms, hybrid methodologies, and novel sensor designs that extend beyond traditional paradigms. Ongoing research debates the relative merits of symbolic versus sub-symbolic learning, the integration of diverse sensory modalities, and strategies for achieving interpretability and transparency in robotic decisions—factors of growing importance as robots take on roles in mission-critical and socially interactive domains.



Despite significant progress, challenges remain in building perception pipelines that are robust, adaptive, and explainable across varied application domains. Emerging work demonstrates the potential of neuromorphic hardware, physics-informed sensing, and learning-based calibration to revolutionize the sensing and reasoning capabilities of robots. However, gaps persist in unifying low- and high-level perception with cognitive reasoning, ensuring trustworthy operation in unstructured and adversarial conditions, and developing systems that generalize across platforms and environments. There is a pressing need for new algorithms, sensor technologies, and cognitive architectures that enhance interpretability, enable reliable sensor fusion, and seamlessly connect perception with planning, control, and human interaction.



This Research Topic aims to bridge advances from robotics, artificial intelligence, sensor engineering, and neuroscience to foster the next generation of perceptual intelligence in autonomous systems. The primary objectives are to (1) characterize novel sensing modalities and their calibration for robust operation, (2) advance algorithms enabling explainable and transparent perception and decision-making, and (3) showcase application-driven case studies from diverse domains—ranging from aerial and underwater robotics to humanoid and wearable systems. Contributions that pursue symbolic, data-driven, or hybrid approaches for enhancing transparency and reliability are particularly encouraged.



The Research Topic focuses on the entire perception pipeline in robotics, from low-level sensing to high-level cognitive integration, emphasizing transparency, robustness, and explainability. Contributions should target innovative methods, technologies, and integrative frameworks, with particular interest in cross-domain insights and emerging challenges. We welcome articles addressing, but not limited to, the following themes:



Interpretable sensor and calibration technologies, including symbolic and physics-informed models

Sensor fusion and learning for perception in uncertain or dynamic environments

Neuromorphic and low-power sensory hardware for embodied intelligence

Machine learning and cognitive architectures for explainable perception and decision-making

Integration of perception with planning, control, and human-robot interaction

Applications in aerial, maritime, space, humanoid, wearable, and swarm robotics

Trust, transparency, and explainability in mission-critical and safety-sensitive contexts