Previous research on finite-time control focuses on forcing a system state (vector) to converge within a certain time moment, regardless of how each state element converges. In the present work, we introduce a control problem with unique finite/fixed-time stability considerations, namely time-synchronized stability (TSS), where at the same time , all the system state elements converge to the origin, and fixed-TSS, where the upper bound of the synchronized settling time is invariant with any initial state. Accordingly, sufficient conditions for (fixed-) TSS are presented. On the basis of these formulations of the time-synchronized convergence property, the classical sign function, and also a norm-normalized sign function , are first revisited. Then in terms of this notion of TSS, we investigate their differences with applications in control system design for first-order systems (to illustrate the key concepts and outcomes), paying special attention to their convergence performance. It is found that while both these sign functions contribute to system stability, nevertheless an important result can be drawn that norm-normalized sign functions help a system to additionally achieve TSS. Furthermore, we propose a fixed-time-synchronized sliding-mode controller for second-order systems; and we also consider the important related matters of singularity avoidance there. Finally, numerical simulations are conducted to present the (fixed-) time-synchronized features attained; and further explorations of the merits of the proposed (fixed-) TSS are described.

Semantic segmentation is one of the significant tasks in understanding aerial images with high spatial resolution. Recently, Graph Neural Network (GNN) and attention mechanism have achieved excellent performance in semantic segmentation tasks in general images and been applied to aerial images. In this paper, we propose a novel Superpixel-based Attention Graph Neural Network (SAGNN) for semantic segmentation of high spatial resolution aerial images. A K-Nearest Neighbor (KNN) graph is constructed from our network for each image, where each node corresponds to a superpixel in the image and is associated with a hidden representation vector. On this basis, the initialization of the hidden representation vector is the appearance feature extracted by a unary Convolutional Neural Network (CNN) from the image. Moreover, relying on the attention mechanism and recursive functions, each node can update its hidden representation according to the current state and the incoming information from its neighbors. The final representation of each node is used to predict the semantic class of each superpixel. The attention mechanism enables graph nodes to differentially aggregate neighbor information, which can extract higher-quality features. Furthermore, the superpixels not only save computational resources, but also maintain object boundary to achieve more accurate predictions. The accuracy of our model on the Potsdam and Vaihingen public datasets exceeds all benchmark approaches, reaching 90.23% and 89.32%, respectively.

Continuum manipulators have infinite degrees of freedom and high flexibility, making it challenging for accurate modeling and control. Some common modeling methods include mechanical modeling strategy, neural network strategy, constant curvature assumption, etc. However, the inverse kinematics of the mechanical modeling strategy is difficult to obtain while a strategy using neural networks may not converge in some applications. For algorithm implementation, the constant curvature assumption is used as the basis to design the controller. When the driving wire is tight, the linear controller under constant curvature assumption works well in manipulator position control. However, this assumption of linearity between the deformation angle and the driving input value breaks upon repeated use of the driving wires which get inevitably lengthened. This degrades the accuracy of the controller. In this work, the Koopman theory is used to identify the nonlinear model of the continuum manipulator. Under the linearized model, the control input is obtained through model predictive control (MPC). As the lifted function can affect the effectiveness of the Koopman operator-based MPC (K-MPC), a novel design method of the lifted function through the Legendre polynomial is proposed. To attain higher control efficiency and computational accuracy, we use a selective control scheme according to the state of the driving wires. When the driving wire is tight, the linear controller is employed; otherwise, the K-MPC is adopted. Finally, a set of static and dynamic experiments have been conducted using an experimental prototype. The results demonstrate high effectiveness and good performance of the selective control scheme.

Robots are increasingly expected to manipulate objects, of which properties have high perceptual uncertainty from any single sensory modality. This directly impacts successful object manipulation. Object packing is one of the challenging tasks in robot manipulation. In this work, a new visuo-tactile feedback-based manipulation planning framework for object packing is proposed, which makes use of the on-the-fly multisensory feedback and an attention-guided deep affordance model as perceptual states as well as a deep reinforcement learning (DRL) pipeline. Significantly, multiple sensory modalities, vision and touch [tactile and force/torque (F/T)], are employed in predicting and indicating the manipulable regions of multiple affordances (i.e., graspability and pushability) for objects with similar appearances but different intrinsic properties (e.g., mass distribution). To improve the manipulation efficiency, the DRL algorithm is trained to select the optimal actions for successful object manipulation. The proposed method is evaluated on both an open dataset and our collected dataset and demonstrated in the use case of the object packing task. The results show that the proposed method outperforms the existing methods and achieves better accuracy with much higher efficiency.

This article studies the planning problem for a robot residing in partially unknown environments under signal temporal logic (STL) specifications, where most of the existing planning methods using STL rely on a fully known environment. In many practical scenarios, however, robots do not have prior information of all the obstacles. In this article, a novel two-phase planning method, i.e., offline exploration followed by online planning, is proposed to efficiently synthesize paths that satisfy STL tasks. In the offline exploration phase, a rapidly exploring random tree* (RRT*) is grown from task regions under the guidance of timed transducers, which guarantees that the resultant paths satisfy the task specifications. In the online phase, the path with minimum cost in RRT* is determined when an initial configuration is assigned. This path is then set as the reference of the time elastic band algorithm, which modifies the path until it has no collisions with obstacles. It is shown that the online computational burden is reduced, and collisions with unknown obstacles are avoided by using the proposed planning framework. The effectiveness and superiority of the proposed method are demonstrated in simulations and real-world experiments.

Active object detection (AOD) offers significant advantage in expanding the perceptual capacity of a robotics system. AOD is formulated as a sequential action decision process to determine optimal viewpoints to identify objects of interest in a visual scene. While reinforcement learning (RL) has been successfully used to solve many AOD problems, conventional RL methods suffer from (i) sample inefficiency, and (ii) unstable outcome due to inter-dependencies of action type (direction of view change) and action range (step size of view change). To address these issues, we propose a novel self-supervised RL method, which employs self-supervised representations of viewpoints to initialize the policy network, and a self-supervised loss on action range to enhance the network parameter optimization. The output and target pairs of self-supervised learning loss are automatically generated from the policy network online prediction and a range shrinkage algorithm (RSA), respectively. The proposed method is evaluated and benchmarked on two public datasets (T-LESS and AVD) using on-policy and off-policy RL algorithms. The results show that our method enhances detection accuracy and achieves faster convergence on both datasets. By evaluating on a more complex environment with a larger state space (where viewpoints are more densely sampled), our method achieves more robust and stable performance. Our experiment on real robot application scenario to disambiguate similar objects in a cluttered scene has also demonstrated the effectiveness of the proposed method.