Search

In certain precision work scenarios, underwater robots require the ability to adhere to surfaces in order to perform tasks effectively. An efficient and stable suction device plays a pivotal role in the functionality of such underwater robots. The vortex suction cup, distinguished by its uncomplicated design, high suction efficiency, and capability for non-contact adhesion, holds significant promise for integration into underwater robotic systems. This paper presents a novel design for a vortex suction cup and investigates its suction force and torque when encountering surfaces with varying curvature radii using Computational Fluid Dynamics (CFD) simulations and experimental testing. These findings offer valuable insights for the development of robots capable of adapting to underwater structures of different dimensions. Results from both experiments and simulations indicate that reducing the curvature radius of the adhered surface results in a decrease in suction force and an increase in torque exerted on the suction cup. As the adhered surface transitions from flat to a curvature radius of 150 mm, the adhesion force of our proposed vortex suction cup decreases by approximately 10%, while the torque increases by approximately 20% to 30%. Consequently, the adhesion efficiency of the suction cup decreases by about 25% to 30%.

Underwater bionic-legged robots encounter significant challenges in attitude, velocity, and positional control due to lift and drag in water current environments, making it difficult to balance operational efficiency with motion stability. This study delves into the hydrodynamic properties of a bionic crab robot’s shell, drawing inspiration from the sea crab’s motion postures. It further refines the robot’s underwater locomotion strategy based on these insights. Initially, the research involved collecting attitude data from crabs during underwater movement through biological observation. Subsequently, hydrodynamic simulations and experimental validations of the bionic shell were conducted, examining the impact of attitude parameters on hydrodynamic performance. The findings reveal that the transverse angle predominantly influences lift and drag. Experiments in a test pool with a crab-like robot, altering transverse angles, demonstrated that increased transverse angles enhance the robot’s underwater walking efficiency, stability, and overall performance.

Bionic-legged robots draw inspiration from animal locomotion methods and structures, demonstrating the potential to traverse irregular and unstructured environments. The ability of Portunus trituberculatus (Portunus) to run flexibly and quickly in amphibious environments inspires the design of systems and locomotion methods for amphibious robots. This research describes an amphibious crab-like robot based on Portunus and designs a parallel leg mechanism for the robot based on biological observations. The research creates the group and sequential gait commonly used in multiped robots combined with the form of the robot’s leg mechanism arrangement. This research designed the parallel leg mechanism and modeled its dynamics. Utilizing the outcomes of the dynamics modeling, we calculate the force and torque exerted on each joint of the leg mechanism during group gait and sequential gait when the robot is moving with a load. This analysis aims to assess the performance of the robot’s motion. Finally, a series of performance evaluation experiments are conducted on land and underwater, which show that the amphibious crab-like robot has good walking performance. The crab-like robot can perform forward, backward, left, and right walking well using group and sequential gaits. Simultaneously, the crab-like robot showcases faster movement in group gaits and a more substantial load capacity in sequential gaits.

IntroductionThe value approximation bias is known to lead to suboptimal policies or catastrophic overestimation bias accumulation that prevent the agent from making the right decisions between exploration and exploitation. Algorithms have been proposed to mitigate the above contradiction. However, we still lack an understanding of how the value bias impact performance and a method for efficient exploration while keeping stable updates. This study aims to clarify the effect of the value bias and improve the reinforcement learning algorithms to enhance sample efficiency.MethodsThis study designs a simple episodic tabular MDP to research value underestimation and overestimation in actor-critic methods. This study proposes a unified framework called Realistic Actor-Critic (RAC), which employs Universal Value Function Approximators (UVFA) to simultaneously learn policies with different value confidence-bound with the same neural network, each with a different under overestimation trade-off.ResultsThis study highlights that agents could over-explore low-value states due to inflexible under-overestimation trade-off in the fixed hyperparameters setting, which is a particular form of the exploration-exploitation dilemma. And RAC performs directed exploration without over-exploration using the upper bounds while still avoiding overestimation using the lower bounds. Through carefully designed experiments, this study empirically verifies that RAC achieves 10x sample efficiency and 25% performance improvement compared to Soft Actor-Critic in the most challenging Humanoid environment. All the source codes are available at https://github.com/ihuhuhu/RAC.DiscussionThis research not only provides valuable insights for research on the exploration-exploitation trade-off by studying the frequency of policies access to low-value states under different value confidence-bounds guidance, but also proposes a new unified framework that can be combined with current actor-critic methods to improve sample efficiency in the continuous control domain.