Motion-force duality in control of anthropomorphic dual redundant robotic arms: modeling, performance, and optimization for dexterous manipulation

The progression of industrial revolutions has redefined robots, moving them from isolated motions in structured settings to adaptive force-aware interactions in human environments. This evolution reveals a central pivot, both philosophical and technical: the shift from treating motion and force as separate domains to understanding their duality as the core of dexterity. This review offers a comprehensive synthesis of this theme for dual-arm robotic systems. We articulate a perspective that views motion-force duality not merely as a mathematical symmetry but as an underlying principle integrating modeling, analysis, and control. The article establishes a unified modeling foundation from anthropomorphic design and screw theory, then examines performance metrics that quantify the trade-offs between agility and strength embedded in this duality. We synthesize how planning and control architectures, ranging from redundancy resolution and hybrid control to learning-based policies, negotiate these trade-offs to execute complex tasks. This work serves as an integrative guide for newcomers and a structured framework for researchers, focusing on the challenge of blending motion and force. We conclude by identifying open frontiers, including real-time compromise optimization and safe adaptive learning, charting a course toward fluid collaborative bimanual manipulation.