Integrated morphology–control co-design of soft climbing robots via triple hybrid evolution

Abstract Soft wall-climbing robots operating on complex vertical surfaces often suffer from limited locomotion efficiency and poor environmental adaptability due to the decoupled optimization of morphology and control. To address this issue, this paper proposes an integrated morphology–control design framework based on the cooperative evolution of which combines the tree seed algorithm, harmony search, and gray wolf optimizer. Structural parameters and control variables, such as actuation timing, are jointly encoded into a hybrid real-valued vector, and a multi-objective fitness function is formulated to account for energy consumption per unit displacement, climbing stability, and adhesion success rate. The algorithm alternates among global exploration, diversity preservation, and local refinement, and candidate solutions are evaluated in a closed loop of coupled finite-element and multibody dynamic simulations with constraint penalties to ensure engineering feasibility. Simulation results demonstrate that the optimized robot achieves an energy consumption of 0.82–1.42 J cm −1 and an average climbing speed of 1.95–3.65 cm s −1 under different inclination angles within the high-fidelity co-simulation environment. Adhesion success rates of 78.5%–96.7% are obtained on surfaces with varying roughness, and a 1 m vertical climb is completed within 28.4–45.6 s, demonstrating superior energy efficiency and surface adaptability.