Adaptive Force Control Strategy for Flexible Grasping of Multiple Power Metering Devices Oriented to Power System Operation

As intelligent power grid construction accelerates, power metering warehouses increasingly require fully automatic handling of diverse and fragile devices, ranging from smart meters to current transformers. Conventional rigid clamping schemes apply fixed empirical thresholds and cannot adapt to objects with different sizes, masses, and surface properties, so positioning errors and unexpected collisions often cause component damage. This paper proposes a flexible grasping scheme with adaptive force regulation for power metering equipment. A multifunctional end-effector controller generates gripping force profiles and impedance parameters according to the estimated properties of each target object. An online optimization module based on deep reinforcement learning fuses feedback from a six-axis force sensor and adjusts the grasping trajectory and controller parameters during contact. A closed-loop framework coordinates visual perception before contact with force-controlled execution after contact. Experiments on a UR5e platform with four categories of metering equipment show that the proposed method achieves a grasp success rate of 91.8% and a damage rate of 2.8%, outperforming five comparison methods.