The fault diagnosis of rolling bearings is vital for the safe and reliable operation of mechanical equipment. However, the imbalanced data collected from the real engineering scenario bring great challenges to the deep learning-based diagnosis methods. For this purpose, this article proposes a methodology called full attention Wasserstein generative adversarial network (WGAN) with gradient normalization (FAWGAN-GN) for data augmentation and uses a shallow 1-D convolutional neural network (CNN) to perform fault diagnosis. First, a gradient normalization (GN) is introduced into the discriminator as a model-wise constraint to make it more flexible in setting the structure of the network, which leads to a more stable and faster training process. Second, the full attention (FA) mechanism is utilized to let the generator pay more attention to learning the discriminative features of the original data and generate high-quality samples. Third, to more thoroughly and deeply evaluate the data generation performance of generative adversarial networks (GANs), a more comprehensive multiple indicator-based evaluation framework is developed to avoid the one-sidedness and superficiality of using one or two simple indicators. Based on two widely applied fault diagnosis datasets and a real rolling bearing fault diagnosis testbed, extensive comparative fault diagnosis experiments are conducted to validate the effectiveness of the proposed method. Experimental results reveal that the proposed FAWGAN-GN can effectively solve the sample imbalance problem and outperforms the state-of-the-art imbalanced fault diagnosis methods.

In recent years, intelligent fault diagnosis methods based on deep learning have developed rapidly. However, most of the existing work performs well under the assumption that training and testing samples are collected from the same distribution, and the performance drops sharply when the data distribution changes. For rolling bearings, the data distribution will change when the load and speed change. In this article, to improve fault diagnosis accuracy and anti-noise ability under different working loads, a transfer learning method based on multi-scale capsule attention network and joint distributed optimal transport (MSCAN-JDOT) is proposed for bearing fault diagnosis under different loads. Because multi-scale capsule attention networks can improve feature expression ability and anti-noise performance, the fault data can be better expressed. Using the domain adaptation ability of joint distribution optimal transport, the feature distribution of fault data under different loads is aligned, and domain-invariant features are learned. Through experiments that investigate bearings fault diagnosis under different loads, the effectiveness of MSCAN-JDOT is verified; the fault diagnosis accuracy is higher than that of other methods. In addition, fault diagnosis experiment is carried out in different noise environments to demonstrate MSCAN-JDOT, which achieves a better anti-noise ability than other transfer learning methods.

Hindsight experience replay (HER) has greatly increased the possibility of using deep reinforcement learning (DRL) for robotic manipulation with sparse rewards. However, there are still concerns about low learning efficiency and poor performance due to its insufficient exploration ability and bias against the initial goal introduced by HER. In this article, to solve this problem, a multigoal robotic manipulation DRL method based on adaptive intrinsic curiosity and universal value network fitting (AIC-UVNF) is proposed to further improve the exploration ability and learning performance. Specifically, this method utilizes an improved curiosity mechanism to construct a joint intrinsic reward and adaptively adjust the proportion, which can enhance exploration ability and avoid excessive pursuit of novel states. In addition, a universal value network fitting approach is proposed to incorporate the initial goal into the value function fitting process, which employs the value of the initial goal to eliminate the bias of HER in the algorithm update. Combined with the off-policy soft actor-critic method, AIC-UVNF is verified on multigoal robotic manipulation tasks. The results show that the proposed method achieves better convergence efficiency and learning performance.

Multi-robot systems have demonstrated significant potential in accomplishing complex tasks, such as cooperative pursuit, search-and-rescue operations. The emergence of heterogeneous robots with diverse capabilities and characteristics shows superior adaptability compared with homogeneous teams. However, in practical applications, global information is typically inaccessible, and composite teams must contend with partial observability and coordination difficulties. To address the issue in heterogeneous multi-robot systems, we propose a novel Intention-Guided reinforcement learning approach with Dirichlet Energy constraint (IGDE). Specifically, an intention-guided module is designed to derive long-horizon strategies based solely on local observations, enabling foresighted decision-making. In addition, a Dirichlet energy constraint is incorporated into the communication process to enhance the diversity of environmental cognition among different classes of robots. Heterogeneous robots perform class-aware actions driven by distinct cognitive representations, thereby enhancing cooperative efficiency. Notably, our approach alleviates the need of prior knowledge and heterogeneity modeling. Extensive comparative experiments and ablation studies verify the effectiveness of the proposed framework. Additionally, real-world deployment is conducted to demonstrate the practicality.