Maintaining object grasp stability represents a pivotal challenge within the domain of robotic manipulation and upper-limb prosthetics. Perturbations originating from external sources frequently disrupt the stability of grasps, resulting in slippage occurrences. Also, if the grasping forces are not optimal while controlling the slip, it may result in the deformation of the objects. This study investigates the robustification of a reinforcement learning (RL) policy for implementing intelligent bionic reflex control, i.e., slip and deformation prevention of the grasped objects. RL-derived policies are vulnerable to failures in environments characterized by dynamic variability. To mitigate this vulnerability, we propose a methodology involving the incorporation of an adaptive sliding mode controller into a pre-trained RL policy. By exploiting the inherent invariance property of the sliding mode algorithm in the presence of uncertainties, our approach strengthens the robustness of the RL policies against diverse and dynamic variations. Numerical simulations substantiate the efficacy of our approach in robustifying RL policies trained within simulated environments.

Soft robotic devices are designed for applications such as exploration, manipulation, search and rescue, medical surgery, rehabilitation, and assistance. Due to their complex kinematics, various and often hard-to-define degrees of freedom, and nonlinear properties of their material, designing and operating these devices can be quite challenging. Using tools such as optimization methods can improve the efficiency of these devices and help roboticists manufacture the robots they need. In this work, we present an extensive and systematic literature search on the optimization methods used for the mechanical design of soft robots, particularly focusing on literature exploiting evolutionary computation (EC). We completed the search in the IEEE, ACM, Springer, SAGE, Elsevier, MDPI, Scholar, and Scopus databases between 2009 and 2024 using the keywords “soft robot,” “design,” and “optimization.” We categorized our findings in terms of the type of soft robot (i.e., bio-inspired, cable-driven, continuum, fluid-driven, gripper, manipulator, modular), its application (exploration, manipulation, surgery), the optimization metrics (topology, force, locomotion, kinematics, sensors, and energy), and the optimization method (categorized as EC or non-EC methods). After providing a road map of our findings in the state of the art, we offer our observations concerning the implementation of the optimization methods and their advantages. We then conclude our paper with suggestions for future research.

This work proposes a novel grasp detection method, the Efficient Grasp Aware Network (EGA-Net), for robotic visual grasp detection. Our method obtains semantic information for grasping through feature extraction. It efficiently obtains feature channel weights related to grasping tasks through the constructed ECA-ResNet module, which can smooth the network’s learning. Meanwhile, we use concatenation to obtain low-level features with rich spatial information. Our method inputs an RGB-D image and outputs the grasp poses and their quality score. The EGA-Net is trained and tested on the Cornell and Jacquard datasets, and we achieve 98.9% and 95.8% accuracy, respectively. The proposed method only takes 24 ms for real-time performance to process an RGB-D image. Moreover, our method achieved better results in the comparison experiment. In the real-world grasp experiments, we use a 6-degree of freedom (DOF) UR-5 robotic arm to demonstrate its robust grasping of unseen objects in various scenes. We also demonstrate that our model can successfully grasp different types of objects without any processing in advance. The experiment results validate our model’s exceptional robustness and generalization.

With the rise of deep reinforcement learning (RL) methods, many complex robotic manipulation tasks are being solved. However, harnessing the full power of deep learning requires large datasets. Online RL does not suit itself readily into this paradigm due to costly and time-consuming agent-environment interaction. Therefore, many offline RL algorithms have recently been proposed to learn robotic tasks. But mainly, all such methods focus on a single-task or multitask learning, which requires retraining whenever we need to learn a new task. Continuously learning tasks without forgetting previous knowledge combined with the power of offline deep RL would allow us to scale the number of tasks by adding them one after another. This paper investigates the effectiveness of regularisation-based methods like synaptic intelligence for sequentially learning image-based robotic manipulation tasks in an offline-RL setup. We evaluate the performance of this combined framework against common challenges of sequential learning: catastrophic forgetting and forward knowledge transfer. We performed experiments with different task combinations to analyse the effect of task ordering. We also investigated the effect of the number of object configurations and the density of robot trajectories. We found that learning tasks sequentially helps in the retention of knowledge from previous tasks, thereby reducing the time required to learn a new task. Regularisation-based approaches for continuous learning, like the synaptic intelligence method, help mitigate catastrophic forgetting but have shown only limited transfer of knowledge from previous tasks.

The demand for flexible grasping of various objects by robotic hands in the industry is rapidly growing. To address this, we propose a novel variable stiffness gripper (VSG). The VSG design is based on a parallel-guided beam structure inserted by a slider from one end, allowing stiffness variation by changing the length of the parallel beams participating in the system. This design enables continuous adjustment between high compliance and high stiffness of the gripper fingers, providing robustness through its mechanical structure. The linear analytical model of the deflection and stiffness of the parallel beam is derived, which is suitable for small and medium deflections. The contribution of each parameter of the parallel beam to the stiffness is analyzed and discussed. Also, a prototype of the VSG is developed, achieving a stiffness ratio of 70.9, which is highly competitive. Moreover, a vision-based force sensing method utilizing ArUco markers is proposed as a replacement for traditional force sensors. By this method, the VSG is capable of closed-loop control during the grasping process, ensuring efficiency and safety under a well-defined grasping strategy framework. Experimental tests are conducted to emphasize the importance and safety of stiffness variation. In addition, it shows the high performance of the VSG in adaptive grasping for asymmetric scenarios and its ability to flexible grasping for objects with various hardness and fragility. These findings provide new insights for future developments in the field of variable stiffness grippers.

In recent years, deep learning-based robotic grasping methods have surpassed analytical methods in grasping performance. Despite the results obtained, most of these methods use only planar grasps due to the high computational cost found in 6D grasps. However, planar grasps have spatial limitations that prevent their applicability in complex environments, such as grasping manufactured objects inside 3D printers. Furthermore, some robotic grasping techniques only generate one feasible grasp per object. However, it is necessary to obtain multiple possible grasps per object because not every grasp generated is kinematically feasible for the robot manipulator or does not collide with other close obstacles. Therefore, a new grasping pipeline is proposed to yield 6D grasps and select a specific object in the environment, preventing collisions with obstacles nearby. The grasping trials are performed in an additive manufacturing unit that has a considerable level of complexity due to the high chance of collision. The experimental results prove that it is possible to achieve a considerable success rate in grasping additive manufactured objects. The UR5 robot arm, Intel Realsense D435 camera, and Robotiq 2F-140 gripper are used to validate the proposed method in real experiments.

In order to complete many complex operations and attain more general-purpose utility, robotic grasp is a necessary skill to master. As the most common essential action of robots in factory and daily life environments, robotic autonomous grasping has a wide range of application prospects and has received much attention from researchers in the past decade. However, the accurate grasp of arbitrary objects in unstructured environments is still a research challenge that has not yet been completely overcome. A complete robotic grasp system usually involves three aspects: grasp detection, grasp planning, and control subsystem. As the first step, identifying the location of the object and generating the grasp pose is the premise of successful grasp, which is conducive to planning the subsequent grasp path and the realization of the entire grasp action. Therefore, this paper conducts a literature review focusing on grasp detection technology and concludes two significant aspects: the analytic and data-driven methods. According to the previous grasp experience of the target object, this paper divides the data-driven methods into the grasp of known and unknown objects. Then it describes in detail the typical grasp detection methods and related characteristics of each classification in the grasp of unknown objects. Finally, current research status and potential research directions in this field are discussed to provide some reference for related research.

As a solution to soft material manipulation, dual-arm robots that are capable of multi-point grasping have become the latest trend due to their higher capability and efficiency. To explore further development in soft material manipulation and go beyond the hardware limit of dual arms, in this paper the robot platform “Quad-SCARA” that consists of four Selective Compliance Assembly Robot Arm (SCARA) is developed and introduced. With the base of hardware platform, a novel collision avoidance system designed for the motion planning of Quad-SCARA in multi-point grasping state is proposed. This system is inspired by the idea of “Buffered Voronoi Cell (BVC),” an algorithm originally proposed for multi-agent collision avoidance. After appropriate adaptation and new proposition in implementation of BVC, the experiments of folding and spreading a handkerchief are performed, and the results in simulation and robot experiments are presented and discussed for the validation and evaluation of entire system.

Handling and manipulating flexible porous objects is one of the main challenges in robotics for household and industrial tasks. Improving the design of grippers for flexible objects of manipulation is an important stage in the development of this topic. This article proposes a method of modeling a gripper for porous objects using the finite element method. It identifies the main parameters of the model that will affect the grasping force and the permeability of porous objects. The power characteristics of the obtained gripper model for different supply pressures, with varying porosity of the manipulated objects, are determined. The obtained characteristics are then used to find the correspondence of channel length for three textile materials with different permeable properties. An experimental study of the lifting force is conducted, and a comparison is made with the obtained modeling data for the presented samples. Additionally, using the obtained simulation data, an analysis of the pressure distribution on the surface of the porous object of manipulation is performed. As a result, it is found that the gripping device must use a design with elements to stabilize the distribution of pressure in its chamber, which will increase the stability of the gripping process.

Recently, Indrek Reiland proposed a new version of the act-type theory of propositions (ATT) in which predication is still committal. However, the Frege-Geach problem can be addressed without resorting to Peter Hanks's cancellation manoeuvre. In this article, I argue that if we take predication as a committal act, we will then have to tackle another problem: non-committal representational acts. I argue that Reiland still needs a notion of cancellation to deal with the latter problem. On this account, he cannot avoid the major flaw he attributes to Hanks's version.

Every year the shortage of biosamples is increasing, while the requirements for their quality are constantly tightening, which requires the introduction of new technological solutions. To solve these problems, a robotic system for aliquoting biological liquid was developed. The aliquotation process is described. The station includes a serial robot on which a gripper based on a globoid worm is installed. The gripping device is parameterized and takes into account the gripping force for different finger deflections. 3D models were developed using Computer Aided Design (CAD) system tools, after which working layouts were created using 3D printing. The design process and test results are discussed to show the efficiency of the built prototype with lab tests.

Knowledge is widely regarded as being incompatible with epistemic luck, but according to several philosophers, the same does not hold for understanding. This paper examines to what extent understanding is vulnerable to epistemic luck. After discussing the weaknesses of some of the cases that have been offered to support the conclusion that understanding tolerates environmental epistemic luck, I turn to a more recent one offered in favour of the opposite conclusion. I argue that this case does not manage to establish that understanding is vulnerable to environmental luck; this even if the fact that understanding comes in degrees is taken into account. Finally, I examine the vulnerability of understanding to intervening luck – the type of luck present in classical Gettier cases – and conclude that when such luck is present, one's understanding is necessarily sub-optimal; a conclusion that does not hold, according to what I argue, when it comes to environmental luck.

This paper presents a study on the effect of design parameters of an underactuated hand on its grasp performance. Three kinds of grasp performance characteristics are considered: grasp range, grasp strength, and immobilizing grasp range. Grasp strength is defined as the stiffness of the grasp. Immobilizing grasps are those in which the object cannot be moved for a force up to a certain threshold. In general, underactuated hands cannot produce immobilizing grasps. However, we show that immobilizing grasps can be created by including joint limits in the hand design. We consider the effect of two design parameters on the grasp performance: torque ratio and finger-base distance. Results show that an optimal finger-base distance and torque ratio exists that maximizes the grasp range and grasp strength. Also, the immobilizing grasp range is increased by decreasing the finger-base distance and increasing the torque ratio and joint limits.

The paper proposes a novel method to detect graspable handles for picking objects from a confined and cluttered space, such as the bins of a rack in a retail warehouse. The proposed method combines color and depth curvature information to create a Gaussian mixture model that can segment the target object from its background and imposes the geometrical constraints of a two-finger gripper to localize the graspable regions. This helps in overcoming the limitations of a poorly trained deep network object detector and provides a simple and efficient method for grasp pose detection that does not require a priori knowledge about object geometry and can be implemented online with near real-time performance. The efficacy of the proposed approach is demonstrated through simulation as well as real-world experiment.

The procyonid †Cyonasua is endemic to South America and recorded from the Late Miocene to the Early Pleistocene. This paper studies the forelimb of †Cyonasua sp. (late Pliocene of Miramar, Argentina), using an ecomorphological approach to infer morphological adaptations linked to substrate preference and locomotory mode, as well as to grasping and digging ability. Twenty linear measurements of forelimb and pectoral girdle were taken from †Cyonasua sp. and a sample of 87 specimens of extant carnivoran families (Procyonidae, Mustelidae, Ursidae, Viverridae, Canidae and Felidae). Raw values were transformed to minimise the effect of size. Morphological variation was explored by principal component analysis (PCA); substrate preference and locomotory mode were further analysed by multivariate analysis of variance (MAV) and discriminant analysis (DA); grasping and digging ability were analysed by DA. In the PCA morphospace, †Cyonasua sp. occupied a unique position, close to extant procyonids. DA classified it as non-specialised digger with poor grasping ability. The results lead to the interpretation of †Cyonasua sp. as having a moderately stabilised elbow joint with poor pronation–supination, although some climbing skills cannot be ruled out. Thus, †Cyonasua sp. could have had generalised habits, in agreement with reconstructed palaeoenvironmental conditions.

This paper presents recent results and applications of our planar kinematic synthesis of serial and parallel linkages to guide a rigid body, such that it does not violate normal direction and curvature constraints imposed by contact with objects in the environment. The paper briefly reviews the recently developed theory on transforming contact direction and curvature constraints into conditions on velocity and acceleration of certain points in the moving body to obtain synthesis equations which can, subsequently be solved to find the dimensions of a mechanical linkage. The main contribution of the paper is in demonstrating the applicability of the proposed theory to the kinematic synthesis of both open and closed-loop kinematic linkages. We provide preliminary results on the synthesis of kinematic chains based on novel task specifications that incorporate curvature constraints with a variety of applications, such as passive suspensions for small rovers, assistive technologies, as well as grasping.

It has become a common practice to use simulation to generate large databases of good grasps for grasp planning in robotics research. However, the existence of a generic simulation context that enables the generation of high quality grasps that can be used in several different contexts such as bin-picking or picking objects from a table, has to our knowledge not yet been discussed in the literature.

In this paper, we investigate how well the quality of grasps simulated in a commonly used “generic” context transfers to a specific context, both, in simulation and in the real world.

We generate a large database of grasp hypotheses for several objects and grippers, which we then evaluate in different dynamic simulation contexts e.g., free floating (no gravity, no obstacles), standing on a table and lying on a table.

We present a comparison on the intersection of the grasp outcome space across the different contexts and quantitatively show that to generate reliable grasp databases, it is important to use context specific simulation.

We furthermore evaluate how well a state of the art grasp database transfers from two simulated contexts to a real world context of picking an object from a table and discuss how to evaluate transferability into non-deterministic real world contexts.