This article investigates the gate recurrent unit neural network (GRUNN)-based learning control framework for the online tracking of manipulators with guaranteed prescribed performance and the absence of velocity measurements. In contrast with methods requiring dynamics modeling, the proposed scheme deals with completely unknown manipulator dynamics, input saturation, and disturbances by virtue of a modified GRUNN. First, the guaranteed output bounds are used to achieve prescribed performance based on the barrier Lyapunov function (BLF). Then, a high-gain observer (HGO) is designed to observe the velocities and construct output feedback control. Finally, a Lyapunov-based GRUNN learning controller is derived by utilizing the backstepping method, in which the GRUNN approximates real-time lumped uncertainties. The advantages are that the proposed solution not only guarantees the prescribed performance with the least sensors but also can handle model-free situations. Indispensably, the practical uniform ultimate boundedness of the closed-loop system is proved. Both static and dynamic tracking tests are performed. Experiment results indicate that the tracking precision is improved, which convincingly demonstrates the significant online estimation performance of the novel GRUNN and the superiority of the proposed control framework.
20 September 2024 in IEEE Transactions on Medical Robotics and Bionics
With the increasing adoption of robotic surgery systems, the need for automated surgical tasks has become more pressing. Recent learning-based approaches provide solutions to surgical automation but typically rely on low-dimensional observations. To further imitate the actions of surgeons in an end-to-end paradigm, this paper introduces a novel visual-based approach to automating surgical tasks using generative imitation learning for robotic systems. We develop a hybrid model integrating state space models transformer, and conditional variational autoencoders (CVAE) to enhance performance and generalization called ACMT. The proposed model, leveraging the Mamba block and multi-head cross-attention mechanisms for sequential modeling, achieves a 75-100% success rate with just 100 demonstrations for most of the tasks. This work significantly advances data-driven automation in surgical robotics, aiming to alleviate the burden on surgeons and improve surgical outcomes.
4 March 2011 in International Journal of Human–Computer Interaction
Whereas the concept of usability is predominantly defined analytically, people relate to systems through personal usability constructs. Based on 48 repertory-grid interviews, this study investigates how such personal constructs are affected by two factors crucial to the international development and uptake of systems: nationality (Chinese, Danish, or Indian) and stakeholder group (developer or user). We find no significant overall difference across nationalities, but further analyses suggest that conventional usability aspects such as ease of use and simplicity are prominent for Chinese and Danish but not Indian participants and that a distinction between work and leisure-related communication is central to Chinese and Indian but not Danish participants. For stakeholder groups, we find a significant overall difference between developers and users. Unlike developers, users associate ease of use with leisure and, conversely, difficulty in use with work-relatedness. Further, users perceive usefulness as related to frustration and separate from ease of use, whereas developers construe usefulness, fun, and ease of use as related. In construing usability, participants make use of several constructs that are not part of prevailing usability definitions, including usefulness, fun, and security.
This paper proposes a learning control framework for the robotic manipulator’s dynamic tracking task demanding fixed-time convergence and constrained output. In contrast with model-dependent methods, the proposed solution deals with unknown manipulator dynamics and external disturbances by virtue of a recurrent neural network (RNN)-based online approximator. First, a time-varying tangent-type barrier Lyapunov function (BLF) is introduced to construct a fixed-time virtual controller. Then, the RNN approximator is embedded in the closed-loop system to compensate for the lumped unknown term in the feedforward loop. Finally, we devise a novel fixed-time, output-constrained neural learning controller by integrating the BLF and RNN approximator into the main framework of the dynamic surface control (DSC). The proposed scheme not only guarantees the tracking errors converge to the small neighborhoods about the origin in a fixed time, but also preserves the actual trajectories always within the prescribed ranges and thus improves the tracking accuracy. Experiment results illustrate the excellent tracking performance and verify the effectiveness of the online RNN estimate for unknown dynamics and external disturbances.
30 December 2024 in IEEE/ASME Transactions on Mechatronics
This article explores the challenges of input saturation and output constraints in the context of a dynamic tracking task for robotic manipulators, considering model uncertainties and the absence of velocity measurements. First, a time-varying global barrier Lyapunov function (GBLF) is designed without restrictions on its application to achieve finite output constraints. Subsequently, an adaptive boundary mode and a new auxiliary system are proposed as part of a composite desaturation scheme to address input saturation. The adaptive boundary mode adjusts GBLF-related control signals while the auxiliary system generates compensation signals for each order subsystem for desaturation. Then, an extended state observer and a radial basis function neural network are employed to estimate velocities and lumped uncertainties, respectively. Finally, a novel controller integrating the aforementioned techniques is derived, achieving satisfactory convergence for each processor of the controller. The advantages of this controller lie in its effective resolution of existing BLFs' applicability issues and noticeable improvement in desaturation rate. Experiments validate the effectiveness and superiority of the proposed method.