Explainable Goal-driven Agents and Robots - A Comprehensive Review

Recent applications of autonomous agents and robots have brought attention to crucial trust-related challenges associated with the current generation artificial intelligence (AI) systems. AI systems based on connectionist deep learning neural network approach lack capabilities explaining their decisions actions others, despite great successes. Without symbolic interpretation capabilities, they are ‘black boxes’, which renders choices or opaque, making it difficult trust them in safety-critical applications. The recent stance explainability has witnessed several approaches eXplainable Artificial Intelligence (XAI) ; however, most studies focused data-driven XAI applied computational sciences. Studies addressing increasingly pervasive goal-driven sparse at this point time. This paper reviews explainable intelligent robots, focusing techniques for communicating agents’ perceptual functions (e.g., senses, vision) cognitive reasoning beliefs, desires, intentions, plans, goals) humans loop. review highlights key strategies that emphasize transparency, understandability, continual explainability. Finally, presents requirements suggests a road map possible realization effective robots.