Soft robots exhibit promising dexterity and adaptability for manipulation because of their high compliance. However, the existing soft robots with invariant stiffness hardly interact with cluttered environments with varying curvatures. In this study, inspired by the maneuverability of an elephant trunk, we proposed a pneumatic soft robot comprised of soft pneumatic actuators and interference plates. By implementing interference plates in diverse patterns, the localized stiffness of our robot can be pre-programmed. To quantify the availability of localized stiffness regulation, we constructed four typical assembly fashions strengthened by the interference plates, and predicted robot profiles by simulation that was validated via experiments. We demonstrated the application of our soft robot with localized stiffness regulation in a variety of scenarios demanding specific configurations. Our design paradigm not only uncovers the potential methodology for regulating robotic curvature precisely, but may inspire next-generation soft robots integrating specified structural and material characteristics learned from natural intelligence.

Odor source localization (OSL) robots play a vital role in detecting hazardous gas leaks, showing significant practical value and potential applications. In this study, we designed a 3D bio-inspired OSL algorithm that employs a pair of bilateral sensors and uses relatively few computational resources, requiring only its previously sampled concentration data and wind information. Initially, the robot’s Spatial Swiveling Sensor Model is constructed. Following this, an animal klinotaxis behavior inspired algorithm, which incorporates the fusion of bilateral odor information and wind information (B-AKBI algorithm), is employed during the plume tracking phase. In the plume finding phase, a cross-wind plane Lévy walk strategy is utilized. An adaptive strategy is developed, allowing the robot to modify step length based on detected concentration, preventing missed detection and maintaining efficiency. We investigate the impact of incorporating bilateral concentration information into the algorithm, and the results indicate a 10.8% to 14.1% increase in success rate. The robustness related to initial orientations is also studied, and the results indicate that for the six initial postures, the algorithm exhibits strong robustness with success rates ranging from 79.0% to 81%. The algorithm’s performance is assessed in 3 scenarios and compared with a surge-spiral algorithm when initially positioned behind the odor source. The results show the proposed algorithm achieves a higher success rate while the average search distance during the plume tracking phase is 41.6% and 30.67% of that of the comparison methods, indicating the efficiency of the B-AKBI algorithm.

Drawing inspiration from the intricate soft structures found in nature, soft actuators possess the ability to incrementally execute complex tasks and adapt to dynamic and interactive environments. In particular, the integration of sensor data feedback allows actuators to respond to environmental stimuli with heightened intelligence. However, conventional rigid sensors are constrained by their inherent lack of flexibility. The current manufacturing processes for flexible sensors are complex and fail to align with the inherent simplicity of soft actuators. In this study, to facilitate the straightforward and consistent sensing of soft pneumatic actuators, carbon–polydimethylsiloxane (CPDMS) materials were employed, utilizing 3D printing and laser-cutting techniques to fabricate a flexible sensor with ease. The preparation of standard tensile specimens verified that the sensor exhibits a fatigue life extending to several hundred cycles and determined its gauge factor to be −3.2. Experimental results indicate that the sensor is suitable for application in soft pneumatic actuators. Additionally, a printed circuit board (PCB) was fabricated and the piecewise constant curvature (PCC) kinematic method was utilized to enable real-time pose estimation of the soft pneumatic actuator. Compared with computer vision methods, the pose estimation error obtained by the sensing method is as low as 4.26%. This work demonstrates that this easily fabricated sensor can deliver real-time pose feedback for flexible pneumatic actuators, thereby expanding the potential application scenarios for soft pneumatic actuators (SPAs).