AI, Machine Learning, and Deep Learning in Advanced Robotics: A Review of Current Innovations and Future Prospects

Abstract

This narrative review addresses the applications of artificial intelligence (AI), machine learning (ML), and deep learning (DL) in robotics. It explains the differences between these paradigms, generalises their use in the area, and determines new innovations and challenges that define the future trends. The literature in the last 20 years of peer-reviewed articles was synthesised through a narrative dissection. The sources were chosen from large databases concerning their relevance to technical underpinnings, field of application, sustainability, and human-robot interaction, as well as cross-disciplinary issues, such as ethics and governance. It was stressed on conceptual integration and comparative analysis at the expense of systematic protocol. The review finds that symbolic AI, ML, and DL provide complementary strengths: transparency, adaptability, and perception capabilities. They find use in surgical and rehabilitative systems of the medical field or in concurrent assembly and predictive maintenance of industrial systems, and have other applications in sustainable agriculture, waste disposal and energy efficiency. Affective computing and augmented reality contribute to the growth of human-robot interaction. Transformers, edge AI, digital twins, bio-inspired robotics, and others are all innovations that are changing the landscape, but explainability, power consumption, and control remain problematic. Robotics is evolving to integrated, hybrid AI solutions. Some of the priorities will be energy-efficient architectures, explainable models, and ethical frameworks of safe, sustainable deployment.