Design and control of a spine-like robotic arm actuated by shape memory alloys for high radiation environments

摘要: This paper presents a radiation-resilient spine-inspired robotic arm actuated by shape-memory alloys (SMAs) to address environmental adaptability challenges in high-radiation nuclear environments such as power plants and decommissioning facilities. Forward- and inverse kinematics models are systematically established, followed by the development of a radiation-hardened control board utilizing field-programmable gate arrays (FPGAs) with multi-layer shielding to ensure operational reliability. To precisely regulate bending deformation, a predictive control algorithm integrating proportional-integral-derivative (PID) and radial basis function (RBF) neural networks calculates the optimal output voltage for heating SMA springs, achieving enhanced shape control accuracy. Furthermore, the robotic arm incorporates programmable variable-stiffness joints enabled by low-melting-point alloys, endowing it with shape-reconfiguration capabilities to adaptively perform tasks in unstructured spaces. Experimental validation confirmed the dexterous and compliant manipulation performance of the system, demonstrating its potential for critical applications, including maintenance, emergency response, and precision inspection within confined radioactive environments.