Enhancing the Versatility and Performance of Soft Robotic Grippers, Hands, and Crawling Robots Through Three-Dimensional-Printed Multifunctional Buckling Joints.

Soft robotic grippers and hands offer adaptability, lightweight construction, and enhanced safety in human-robot interactions. In this study, we introduce vacuum-actuated soft robotic finger joints to overcome their limitations in stiffness, response, and load-carrying capability. Our design-optimized through parametric design and three-dimensional (3D) printing-achieves high stiffness using vacuum pressure and a buckling mechanism for large bending angles (>90°) and rapid response times (0.24 s). We develop a theoretical model and nonlinear finite-element simulations to validate the experimental results and provide valuable insights into the underlying mechanics and visualization of the deformation and stress field. We showcase versatile applications of the buckling joints: a three-finger gripper with a large lifting ratio (∼96), a five-finger robotic hand capable of replicating human gestures and adeptly grasping objects of various characteristics in static and dynamic scenarios, and a planar-crawling robot carrying loads 30 times its weight at 0.89 body length per second (BL/s). In addition, a jellyfish-inspired robot crawls in circular pipes at 0.47 BL/s. By enhancing soft robotic grippers' functionality and performance, our study expands their applications and paves the way for innovation through 3D-printed multifunctional buckling joints.

Working Memory Ability Evaluation Based on Fuzzy Support Vector Regression.

One's working memory process is a fundamental cognitive activity which often serves as an indicator of brain disease and cognitive impairment. In this research, the approach to evaluate working memory ability by means of electroencephalography (EEG) analysis was proposed. The result shows that the EEG signals of subjects share some characteristics when performing working memory tasks. Through correlation analysis, a working memory model describes the changes in EEG signals within alpha, beta and gamma waves, which shows an inverse tendency compared to Zen meditation. The working memory ability of subjects can be predicted using multi-linear support vector regression (SVR) with fuzzy C-mean (FCM) clustering and knowledge-based fuzzy support vector regression (FSVR), which reaches the mean square error of 0.6 in our collected data. The latter, designed based on the working memory model, achieves the best performance. The research provides the insight of the working memory process from the EEG aspect to become an example of cognitive function analysis and prediction.

Development of a Visual Perception System on a Dual-Arm Mobile Robot for Human-Robot Interaction.

This paper presents the development of a visual-perception system on a dual-arm mobile robot for human-robot interaction. This visual system integrates three subsystems. Hand gesture recognition is utilized to trigger human-robot interaction. Engagement and intention of the participants are detected and quantified through a cognitive system. Visual servoing uses YOLO to identify the object to be tracked and hybrid, model-based tracking to follow the object's geometry. The proposed visual-perception system is implemented in the developed dual-arm mobile robot, and experiments are conducted to validate the proposed method's effects on human-robot interaction applications.

Robot Intelligent Grasp of Unknown Objects Based on Multi-Sensor Information.

Robots frequently need to work in human environments and handle many different types of objects. There are two problems that make this challenging for robots: human environments are typically cluttered, and the multi-finger robot hand needs to grasp and to lift objects without knowing their mass and damping properties. Therefore, this study combined vision and robot hand real-time grasp control action to achieve reliable and accurate object grasping in a cluttered scene. An efficient online algorithm for collision-free grasping pose generation according to a bounding box is proposed, and the grasp pose will be further checked for grasp quality. Finally, by fusing all available sensor data appropriately, an intelligent real-time grasp system was achieved that is reliable enough to handle various objects with unknown weights, friction, and stiffness. The robots used in this paper are the NTU 21-DOF five-finger robot hand and the NTU 6-DOF robot arm, which are both constructed by our Lab.

Learn the Lagrangian: A Vector-Valued RKHS Approach to Identifying Lagrangian Systems.

We study the modeling of Lagrangian systems with multiple degrees of freedom. Based on system dynamics, canonical parametric models require ad hoc derivations and sometimes simplification for a computable solution; on the other hand, due to the lack of prior knowledge in the system's structure, modern nonparametric models in machine learning face the curse of dimensionality, especially in learning large systems. In this paper, we bridge this gap by unifying the theories of Lagrangian systems and vector-valued reproducing kernel Hilbert space. We reformulate Lagrangian systems with kernels that embed the governing Euler-Lagrange equation-the Lagrangian kernels-and show that these kernels span a subspace capturing the Lagrangian's projection as inverse dynamics. By such property, our model uses only inputs and outputs as in machine learning and inherits the structured form as in system dynamics, thereby removing the need for the mundane derivations for new systems as well as the generalization problem in learning from scratches. In effect, it learns the system's Lagrangian, a simpler task than directly learning the dynamics. To demonstrate, we applied the proposed kernel to identify the robot inverse dynamics in simulations and experiments. Our results present a competitive novel approach to identifying Lagrangian systems, despite using only inputs and outputs.

Learning the Inverse Dynamics of Robotic Manipulators in Structured Reproducing Kernel Hilbert Space.

We investigate the modeling of inverse dynamics without prior kinematic information for holonomic rigid-body robots. Despite success in compensating robot dynamics and friction, general inverse dynamics models are nontrivial. Rigid-body models are restrictive or inefficient; learning-based models are generalizable yet require large training data. The structured kernels address the dilemma by embedding the robot dynamics in reproducing kernel Hilbert space. The proposed kernels autonomously converge to rigid-body models but require fewer samples; with a semi-parametric framework that incorporates additional parametric basis for friction, the structured kernels can efficiently model general rigid-body robots. We tested the proposed scheme in simulations and experiments; the models that consider the structure of function space are more accurate.

A novel offset cancellation based on parasitic-insensitive switched-capacitor sensing circuit for the out-of-plane single-Gimbaled decoupled CMOS-MEMS gyroscope.

This paper presents a novel parasitic-insensitive switched-capacitor (PISC) sensing circuit design in order to obtain high sensitivity and ultra linearity and reduce the parasitic effect for the out-of-plane single-gimbaled decoupled CMOS-MEMS gyroscope (SGDG). According to the simulation results, the proposed PISC circuit has better sensitivity and high linearity in a wide dynamic range. Experimental results also show a better performance. In addition, the PISC circuit can use signal processing to cancel the offset and noise. Thus, this circuit is very suitable for gyroscope measurement.

Chip implementation with a combined wireless temperature sensor and reference devices based on the DZTC principle.

This paper presents a novel CMOS wireless temperature sensor design in order to improve the sensitivity and linearity of our previous work on such devices. Based on the principle of CMOS double zero temperature coefficient (DZTC) points, a combined device is first created at the chip level with two voltage references, one current reference, and one temperature sensor. It was successfully fabricated using the 0.35 µm CMOS process. According to the chip results in a wide temperature range from -20 °C to 120 °C, two voltage references can provide temperature-stable outputs of 823 mV and 1,265 mV with maximum deviations of 0.2 mV and 8.9 mV, respectively. The result for the current reference gives a measurement of 23.5 µA, with a maximum deviation of 1.2 µA. The measurements also show that the wireless temperature sensor has good sensitivity of 9.55 mV/°C and high linearity of 97%. The proposed temperature sensor has 4.15-times better sensitivity than the previous design. Moreover, to facilitate temperature data collection, standard wireless data transmission is chosen; therefore, an 8-bit successive-approximation-register (SAR) analog-to-digital converter (ADC) and a 433 MHz wireless transmitter are also integrated in this chip. Sensing data from different places can be collected remotely avoiding the need for complex wire lines.