Soft robots leverage compliant materials to generate motion through controlled elastic deformation, making them ideal for delicate tasks such as underwater exploration and biomimetic marine systems. Although hydraulic/pneumatic actuation remains pivotal for such systems, the lack of systematic design frameworks has hindered the development of robots capable of complex 3D motion, such as fish-like swimming. This work introduces a topology optimization method to automate the design of a hydraulic soft fish tail, explicitly addressing the design-dependent coupling between fluidic actuation and structural deformation. We use a Darcy law-based model augmented with a drainage term to simulate spatially varying hydraulic pressure loads, translating these into consistent nodal forces via finite element analysis. The employed robust multi-criteria optimization formulation balances deformation efficiency, fluid-structure interaction, geometric manufacturability, and required stiffness for optimizing a bioinspired soft fish tail for 3D swimming kinematics. The optimized tail topology is incorporated into a pneumatic network actuator and computationally validated under various hydraulic loads, achieving tunable undulatory amplitudes and multiaxis bending for depth adjustment. The optimized 2D tail outperforms its rectangular counterpart. By cascading optimized tail segments, we demonstrate programmable swimming patterns in soft robotic fish tails at different hydraulic loads. This work advances the systematic codesign of hydraulic actuators and soft structures, offering a pathway to automate underwater robots with optimized design and vertebrate-like agility in confined aquatic environments. Our implementations and simulations are publicly available at 'https://github.com/PrabhatIn/TO-SoFiT'.

翻译：软体机器人利用柔性材料通过受控弹性变形产生运动，使其成为水下勘探和仿生海洋系统等精细任务的理想选择。尽管液压/气动驱动仍是此类系统的关键，但缺乏系统性设计框架阻碍了能够实现复杂三维运动（如鱼类游动）的机器人开发。本文提出了一种拓扑优化方法，用于自动化设计液压软鱼尾，明确处理流体驱动与结构变形之间的设计依赖耦合关系。我们采用基于达西定律并引入排水项的模型来模拟空间变化的液压载荷，通过有限元分析将其转换为一致的节点力。所采用的稳健多准则优化公式平衡了变形效率、流固耦合、几何可制造性以及仿生软鱼尾三维游动运动学优化所需的刚度。优化后的尾部拓扑结构被集成到气动网络执行器中，并在多种液压载荷下进行了计算验证，实现了可调波动幅度和用于深度调节的多轴弯曲。优化后的二维尾部性能优于其矩形对应物。通过级联优化的尾部段，我们展示了在不同液压载荷下软体机器鱼尾的可编程游动模式。本研究推进了液压执行器与软体结构的系统共设计，为在受限水生环境中实现具有优化设计和脊椎动物般敏捷性的自动化水下机器人开辟了路径。我们的实现代码与仿真结果已公开于'https://github.com/PrabhatIn/TO-SoFiT'。