Mechanical reliability in chiplet packaging is heavily influenced by the layout-dependent anisotropic properties of redistribution layers (RDLs). Traditional evaluation of these mechanical responses requires detailed GDSII layouts and computationally expensive finite element method (FEM) simulations, which become impractical during early-stage design exploration when floorplans and detailed routing are not yet available. To address this limitation, we propose a scalable framework for the rapid, layout-aware estimation of effective orthotropic elastic properties. The approach decouples intensive material characterization from the real-time inference flow through a three-phase pipeline: (1) an offline data generation stage that employs a parallelized FEM solver and homogenization theory to compute ground-truth elastic stiffness matrices (D); (2) a density-constrained A* global routing engine that generates representative copper density maps from given chiplet placements; and (3) a multi-layer perceptron (MLP) surrogate model that maps localized 3 × 3 Moore neighborhood copper volume fractions to the nine independent components of the orthotropic stiffness matrix (D1111,…, D2323). The proposed framework enables near-instantaneous ML-based inference of spatially varying elastic properties across the entire RDL footprint, delivering substantial speed-ups compared to full FEM simulations. Validation on an independent 100-µm-resolution test set demonstrates high predictive accuracy, with R2 values of 0.9519 for D1111 and 0.9299 for D2222, and mean absolute percentage errors (MAPE) as low as 4.60%. This framework establishes a foundational tool for layout-driven mechanical optimization, enabling designers to proactively mitigate thermo-mechanical stress risks in next-generation heterogeneous integration systems during the pre-routing design phase.

This paper introduces a quantum-enhanced finite element method (FEM) designed for noisy intermediate-scale quantum (NISQ) devices, leveraging variational quantum algorithms (VQAs) to solve engineering partial differential equations. We demonstrate the framework by solving the Euler-Bernoulli beam and the NAFEMS T4 heat transfer problems, which involve Dirichlet, Neumann, and Robin boundary conditions. A key innovation is a 'set-to-zero' strategy that incorporates boundary conditions through a correction matrix, Kbc, allowing for flexible imposition at any node without domain decomposition. The global stiffness matrix is decomposed into a constant number of Pauli terms, O(1), using the method by Sato et al while boundary terms are handled with a sublinearly scaling partial Pauli measurement technique. The algorithm achieves logarithmic qubit scaling ( n=log2N qubits for N degrees of freedom(DOF)) and employs shallow, hardware-efficient circuits with empirically trainable depth for small-scale systems. Validation on the Qiskit statevector simulator shows high accuracy. For the Euler-Bernoulli beam problem with 4 to 64 DOF, the algorithm achieves relative errors of 0.5%-1.5% and fidelities of 0.998-0.999. For the NAFEMS T4 heat transfer benchmark, a 5.4% relative error is observed. The VQA converges robustly within 77-350 iterations, though barren plateaus are a known challenge for scaling to larger systems. This work establishes a scalable framework for quantum FEM, offering a significant memory advantage over classical methods and advancing the potential for quantum-enhanced engineering simulations.

This paper presents a novel approach for optimizing the design of power distribution networks (PDNs) in heterogeneous multi-chiplet systems by leveraging a graph neural network (GNN). The proposed method predicts self-impedances at observation ports and optimizes decoupling capacitor placement to minimize PDN impedance across multiple voltage domains. Addressing the inherent complexity of PDN design in multi-chiplet architectures, a GNN-based surrogate model is employed to efficiently explore the high-dimensional design space, streamlining capacitor selection and placement. The optimization framework integrates PDN design objectives and constraints into a feedback-driven deep reinforcement learning process, enabling impedance reduction while minimizing the total number of capacitors. This approach ensures adherence to key design rules while achieving optimal PDN performance within a targeted bandwidth. By combining GNN-based modeling with reinforcement learning, this work represents a significant advancement in PDN design methodology, offering a faster and more cost-effective solution for the heterogeneous integration of multi-chiplet systems.

The logarithmic spiral is observed as a common pattern in several living beings across kingdoms and species. Some examples include fern shoots, prehensile tails, and soft appendages like octopus arms and elephant trunks. In the latter cases, spiraling is also used for grasping. Motivated by how this strategy simplifies behavior into kinematic primitives and combines them to develop smart grasping movements, this work focuses on the elephant trunk, which is more deeply investigated in the literature. We present a soft arm combined with a rigid robotic arm to replicate elephant grasping capabilities. In our system, the rigid arm ensures positioning and orientation, mimicking the role of the elephant’s head, while the soft manipulator reproduces trunk motion primitives of bending and twisting under proper actuation patterns. The synergy between rigid and soft components replicates 9 distinct elephant grasping strategies, enabling adaptation to various object shapes and sizes while reducing control complexity.

Soft grippers, that can mimic human fingers grasping objects, have emerged as a game-changer in the automated food service industry. Their flexibility and compliance enable them to handle various food items, regardless of size, shape, or stiffness, outperforming their rigid counterparts while maintaining cost-efficiency and greater adaptability. However, the working scenarios for soft grippers are also more complex and unpredictable, leading to a challenge in pre-programming the gesture and trajectory of the system. Current research on soft grippers primarily focuses on their compliance characteristics, which is effective for objects with characteristic features but faces challenges with cut food items. For such cut objects, both vertical friction and compliance play crucial roles in grasping, highlighting the need for a reconfigurable soft gripper system (RSGS) with multiple degrees of freedom actuators. To address these challenges, we introduce an intelligent automated gesture planning strategy for RSGSs with multiple degrees of freedom. Our proposed framework comprises five modules: Feature Engineering, which parameterizes and samples arbitrary polygon cross-sections of food items; Simulation, which automates the creation of numerical models in the design space, run and post-processing of simulation; GraspingFormer, which estimates reaction forces during grasping; AgentVAE, which uses a generative variational autoencoder to sample feasible grasping gestures in latent space; Planner, which identifies the optimized gestures by solving an inverse problem. This strategy can facilitate gesture planning when grasping a target object with a RSGS, to enhance the picking-up success rate. The proposed framework can potentially benefit food-handling-like tasks and expand the use of soft robots in real-world applications.

In heterogeneous integration and advanced packaging, the mechanical reliability of the Redistribution Layer (RDL) is crucial due to its role in electrical routing and interconnect density. Traditional mechanical simulation typically assumes homogeneous and isotropic material behavior, which can lead to inaccurate stress prediction, especially under complex loading and temperature conditions. Given the heterogeneous composition and fine-scale geometry of RDL, capturing anisotropic mechanical response is vital for design accuracy and reliability. However, full-scale mechanical modeling with detailed resolution is computationally intensive. This work proposes a novel modeling scheme for fast, real-time evaluation of effective anisotropic mechanical properties of RDL structures, enabling accurate, scalable, and simulationdriven mechanical design.

This paper presents a novel approach for rapid evaluation of effective thermal property distribution in Redistribution Layers (RDLs) used in advanced packaging. The methodology combines GDSii file processing, finite element analysis (FEM), and machine learning to overcome the limitations of existing methods. The process involves dividing the RDL into small patches, assembling them into a structured mesh, conducting FEM analysis using an optimized in-house code for data generation, and training a U-Net-based machine learning model to predict effective thermal property distributions. This approach significantly reduces computational time compared to traditional methods, enabling simultaneous evaluation of the entire RDL material property distribution. The method achieves a balance between accuracy and efficiency, reducing evaluation time from days to seconds while maintaining reasonable prediction accuracy. This innovation facilitates thermal characterization and design optimization in chiplet-based packaging, potentially accelerating the development of next-generation electronic devices through improved co-design practices.

While advanced packaging technologies such as 2.5D chiplet packaging offer advantages such as improved manufacturing yield, lower cost and higher transistor densities, they come with the cost of critical thermal management challenges. Thermal-aware design optimization often requires thousands of simulation runs, however, existing thermal modelling techniques face fundamental trade-offs between accuracy and computational cost that limit their utility. Conventional finite element simulations provide high-fidelity solutions but are computationally expensive. Recently, machine learning approaches have been proposed to replace conventional 3D steady-state thermal solvers. Once trained, machine learning models can provide reasonable predictions at millisecond timescales, showing great potential in accelerating design space exploration for thermal-aware chiplet placements. However, the long training time, complex architecture, and accuracy of neural networks have been the key limitations. We propose a novel technique that synergistically combines Proper Orthogonal Decomposition (POD) and Artificial Neural Networks (ANNs) to overcome these limitations. By leveraging POD to embedding the essential information from high-dimensional thermal data onto a low-dimensional subspace, an ANN can be rapidly trained to learn the low-rank representation. This compressed predictive model can then regenerate full high-resolution thermal fields orders-of-magnitude faster than rerunning simulations.

Dielectric elastomer actuators (DEAs), a type of “artificial muscles”, can generate significant deformations and offer speedy responses when exposed to voltage. Owing to their high electromechanical conversion efficiency and great flexibility, they have been extensively used in soft robot applications, such as soft grippers, walking robots, crawling robots, climbing robots, swimming robots, etc. Although previous research has explored the use of DEAs in soft robot locomotion, achieving optimal behavior is challenging due to the complexity of the constituent materials and the highly nonlinear nature of the problem. In this study, a simulation-based design optimization approach is proposed to address this challenge. The proposed approach involves developing a computational modeling framework that evaluates the electromechanical behavior of the DEA. A graph neural network (GNN) is employed as an encoder to extract the latent representation of the geometry in a low dimensional space, which is further used to construct a surrogate model for fast prediction of target responses. To achieve an optimal actuation capability under design constraints, a multi-objective optimization function is formulated to balance the actuation distance and the actuator size, where the Pareto front demonstrates the trade-off between the actuation distance and design constraints. Finally, three optimized designs are fabricated and tested, demonstrating a performance improvement of over 140% compared to an intuitive design. This framework can greatly benefit the design of DEA-based soft robotics.

Recent advancements in soft robotics have seen the rapid development of soft grippers for industrial pick-and-place applications. They are, however, ill-suited to bear heavy loads due to their compliant nature. Paradoxically, researchers have sought to increase the stiffness of soft grippers to improve load-bearing capabilities. Unfortunately, contemporary soft actuators with variable stiffness are fabricated using manual processes and their performance is subject to an individual's mastery. They are therefore not reliable for long-term industrial use. In this article, we present our work on a 3-D-printed metal-endoskeleton-reinforced actuator (MERA) for industrial pick-and-place applications. We also highlight the fabrication processes needed to recreate it repetitively. Using stainless steel splints (SSS), we demonstrate that MERA is able to modulate its stiffness at selective junctures for stable and effective grasping. We also describe our design rationale with a qualitative mathematical model and validate its performance quantitatively using a finite element model, which is further investigated in the following fatigue test. In our experiments, the MERA equipped with SSS is able to output a peak tip force of 8 N, which is a 291% increase compared to the one without metallic reinforcement. In addition, an increase of 76.5% in gripping load and a maximum holding force per actuator of 13.8 N are realized through the stiffness tuning of a MERA-Gripper. Despite significantly improving load-bearing capabilities, the actuator manages to retain an overall low profile with a weight of 82 g. Finally, we adapted the MERA into a reconfigurable gripper and tested its grasping capabilities on objects of various shapes, sizes, and weights.

We proposed a general quantum-computing-based algorithm that harnesses the exponential power of noisy intermediate-scale quantum (NISQ) devices in solving partial differential equations (PDE). This variational quantum eigensolver (VQE)-inspired approach transcends previous idealized model demonstrations constrained by strict and simplistic boundary conditions. It enables the imposition of arbitrary boundary conditions, significantly expanding its potential and adaptability for real-world applications, achieving this"from ad-hoc to systematic"concept. We have implemented this method using the fourth-order PDE (the Euler-Bernoulli beam) as example and showcased its effectiveness with four different boundary conditions. This framework enables expectation evaluations independent of problem size, harnessing the exponentially growing state space inherent in quantum computing, resulting in exceptional scalability. This method paves the way for applying quantum computing to practical engineering applications.

Soft robots have received much attention due to their impressive capabilities including high flexibility and inherent safety features for humans or unstructured environments compared with hard-bodied robots. Soft actuators are the crucial components of soft robotic systems. Soft robots require dexterous soft actuators to provide the desired deformation for different soft robotic applications. Most of the existing soft actuators have only one or two deformation modes. In this article, a new soft pneumatic actuator (SPA) is proposed taking inspiration from Kirigami. Kirigami-inspired cuts are applied to the actuator design, which enables the SPA to be equipped with multiple deformation modes. The proposed Kirigami-inspired soft pneumatic actuator (KiriSPA) is capable of producing bending motion, stretching motion, contraction motion, combined motion of bending and stretching, and combined motion of bending and contraction. The KiriSPA can be directly manufactured using 3D printers based on the fused deposition modeling technology. Finite element method is used to analyze and predict the deformation modes of the KiriSPA. We also investigated the step response, creep, hysteresis, actuation speed, stroke, workspace, stiffness, power density, and blocked force of the KiriSPA. Moreover, we demonstrated that KiriSPAs can be combined to expand the capabilities of various soft robotic systems including the soft robotic gripper for delicate object manipulation, the soft planar robotic manipulator for picking objects in the confined environment, the quadrupedal soft crawling robot, and the soft robot with the flipping locomotion.

Inhomogeneous swelling of polymer films in liquid environments may find applications in soft actuators and sensors. Among them, fluoroelastomer based films bend up spontaneously once they are placed on an acetone-soaked filter paper. The stretchability and dielectric properties of a fluoroelastomer is attractive in the fields of soft actuators and sensors, making in-depth studies on and understanding of fluoroelastomer bending behaviors important. Here, we report an abnormal size-dependent bending phenomenon of rectangular fluoroelastomer films, which transform the bending direction from the long-side bending to the short-side bending as their length or width increases or the thickness decreases. By using finite element analysis and an analytical expression obtained using a bilayer model, we reveal the key role of gravity in determining the size-dependent bending behavior. In the bilayer model, an energy quantity is obtained to characterize the role of each material and geometrical parameters in determining the size-dependent bending behavior. We further construct phase diagrams to correlate the bending modes and the film sizes based on the finite element results, which are in good agreement with experimental results. These findings can be useful for the design of future swelling-based polymer actuators and sensors.

We proposed a general quantum-computing-based algorithm that harnesses the exponential power of noisy intermediate-scale quantum (NISQ) devices in solving partial differential equations (PDE). This variational quantum eigensolver (VQE)-inspired approach transcends previous idealized model demonstrations constrained by strict and simplistic boundary conditions. It enables the imposition of arbitrary boundary conditions, significantly expanding its potential and adaptability for real-world applications, achieving this "from ad-hoc to systematic" concept. We have implemented this method using the fourth-order PDE (the Euler-Bernoulli beam) as example and showcased its effectiveness with four different boundary conditions. This framework enables expectation evaluations independent of problem size, harnessing the exponentially growing state space inherent in quantum computing, resulting in exceptional scalability. This method paves the way for applying quantum computing to practical engineering applications.

Conventional pressure sensors rely on solid sensing elements. Instead, inspired by the air entrapment phenomenon on the surfaces of submerged lotus leaves, we designed a pressure sensor that uses the solid–liquid–liquid–gas multiphasic interfaces and the trapped elastic air layer to modulate capacitance changes with pressure at the interfaces. By creating an ultraslippery interface and structuring the electrodes at the nanoscale and microscale, we achieve near-friction-free contact line motion and thus near-ideal pressure-sensing performance. Using a closed-cell pillar array structure in synergy with the ultraslippery electrode surface, our sensor achieved outstanding linearity ( R 2 = 0.99944 ± 0.00015; nonlinearity, 1.49 ± 0.17%) while simultaneously possessing ultralow hysteresis (1.34 ± 0.20%) and very high sensitivity (79.1 ± 4.3 pF kPa −1 ). The sensor can operate under turbulent flow, in in vivo biological environments and during laparoscopic procedures. We anticipate that such a strategy will enable ultrasensitive and ultraprecise pressure monitoring in complex fluid environments with performance beyond the reach of the current state-of-the-art.

Rapid detection of ammonium nitrogen(NH4+) in soil is quit important for precise agriculture as it can significantly increase the utilization of soil fertility and reduce the environmental pollution. Compared with the traditional detection methods for NH4+ which are time-consuming and expensive, here we developed a rapid and sensitive electrochemical sensor by platinum deposited screen printed electrodes(Pt-SPEs) as sensing components. Owing to the specific catalysis to NH4+ , using the carbonate to precipitate the metal ions, coupling with the ammonium extraction syringe, NH4+ in soil can be detected in 5 min for the first time, with a detection range of 0.1-5 mmol/L(R2=0.997)and a detection limit of 13 mu mol/L(S/N=3). This sensors can be applied for on-site detection of NH4+ in soil due to its rapidity, simplicity, high selectivity and stability without large equipment, providing an important data support for precise agriculture.

In this work, we introduce a novel technique that optimizes the design and properties of support structures for Powder Bed Fusion–Laser-Based Metal (PBF-LB/M) processes. The goal is to significantly ease support removal during post-processing without compromising print quality. We propose an additive manufacturing (AM) workflow that streamlines the end-to-end AM process through the creation and integration of innovative support systems, modeling software, and efficient support removal techniques. This approach enhances productivity and minimizes trial and error between design and 3D printing. Our key concept revolves around developing a “soft” support system, which can be easily removed using sandblasting. Compared to manual or machining removal, this method substantially reduces cost and time, particularly in areas with limited or no access. For locations with steep overhangs and large cross-sectional areas, the traditional strong support system can be implemented as a “hard” support. The proposed workflow allows for the development of a hybrid system combining hard and soft supports, effectively manufacturing parts with minimal design constraints. The workflow begins from a standard operating procedure (SOP) for evaluating the appropriate printing parameters to achieve a support that can be soft enough to be sandblasted for removal, yet sufficient strong to support the overhang part structure and prevent distortion. A simulation platform is developed to assess the ideal placement and combination of the hybrid support design, achieving a balance between post-processing cost and effort, and print quality. The development of this solution truly reflects the advantage of metal 3D printing technology, i.e., design freedom.

This paper contributes to a new design of the three-dimensional printable robotic ball joints capable of creating the controllable stiffness linkage between two robot links through pneumatic actuation. The variable stiffness ball joint consists of a soft pneumatic elastomer actuator, a support platform, an inner ball and a socket. The ball joint structure, including the inner ball and the socket, is three-dimensionally printed using polyamide−12 (PA12) by selective laser sintering (SLS) technology as an integral mechanism without the requirement of assembly. The SLS technology can make the ball joint have the advantages of low weight, simple structure, easy to miniaturize and good MRI compatibility. The support platform is designed as a friction-based braking component to increase the stiffness of the ball joint while withstanding the external loads. The soft pneumatic elastomer actuator is responsible for providing the pushing force for the support platform, thereby modulating the frictional force between the inner ball, the socket and the support platform. The most remarkable feature of the proposed variable stiffness design is that the ball joint has ‘zero’ stiffness when no pressurized air is supplied. In the natural state, the inner ball can be freely rotated and twist inside the socket. The proposed ball joint can be quickly stiffened to lock the current position and orientation of the inner ball relative to the socket when the pressurized air is supplied to the soft pneumatic elastomer actuator. The relationship between the stiffness of the ball joint and the input air pressure is investigated in both rotating and twisting directions. The finite element analysis is conducted to optimize the design of the support platform. The stiffness tests are conducted, demonstrating that a significant stiffness enhancement, up to approximately 508.11 N·mm reaction torque in the rotational direction and 571.93 N·mm reaction torque in the twisting direction at the pressure of 400 kPa, can be obtained. Multiple ball joints can be easily assembled to form a variable stiffness structure, in which each ball joint has a relative position and an independent stiffness. Additionally, the degrees of freedom (DOF) of the ball joint can be readily restricted to build the single-DOF or two-DOFs variable stiffness joints for different robotic applications.

This article presents a versatile soft robotic gripper system whereby its fingers can be reconfigured into different poses such as scoop, pinch, and claw. This allows the gripper to efficiently and safely handle food samples of different shapes, sizes and stiffness such as uncooked tofu and broccoli floret. The 3D-printed fingers were tested to last up to 25 000 cycles without significant changes in the curvature profile and force output profile. A benchmark experiment was conducted to evaluate the performance of the gripper and state-of-the-art gripping solutions. Capability of versatile soft gripper was optimized by integrating vision and tactile sensing facilities. An object recognition system was developed to identify food samples such as potato, broccoli, and sausage. Position and orientation of food samples were identified and pick-and-place pathway was optimized to achieve the best gripping performance. Flexible tactile sensors were integrated into soft fingers and closed-loop force feedback control system was developed. This allowed the gripper to automatically explore and select the most stable grip pose for different food samples. Integration of vision and force feedback system ensure that objects detected by the system would be firmly gripped. The reconfigurable soft robotic gripper system has been demonstrated to perform high-speed pick-and-place tasks (∼3 s per item) with object recognition system, making it a potential solution to food and grocery supply chain needs.