Design of compliant adaptive constant-force soft robotic fingers based on a sensor-free force estimation method

Soft robotic grippers show great promise for handling fragile and geometrically complex objects, but most existing designs still rely on complex control strategies or advanced tactile sensing for force regulation. Compliant constant-force mechanisms offer a simpler, more attractive alternative, enabling force regulation through mechanical intelligence without the need for sensors or active control. In this paper, we propose a novel type of compliant adaptive constant-force soft robotic finger, which combines a reconfigurable fin-ray-inspired structure with typical bistable beams to achieve both adaptive grasping and constant-force output. We also develop a sensor-free force estimation method to analyze deformed shapes and reaction forces during grasping. A particle swarm optimization algorithm is employed to determine the geometric parameters that maximize the constant-force range. Finite element simulations and experimental tests validate the proposed model and demonstrate the finger's performance. Furthermore, extensive experiments with daily items of varying shapes, sizes, and stiffness confirm the feasibility and robustness of the design. The results highlight the potential of this finger for adaptive, efficient, and safe manipulation of fragile or deformable objects in everyday scenarios.