Behavioral cloning from human demonstrations has succeeded in programming a robot to generate fine-grained motion, but it is still challenging to learn multimodal trajectories such as with various speeds. This restricts the use of a robot dataset collected by multiusers because the different proficiency of robot operators makes the dataset have diverse distributions of speed. To tackle this issue, we develop Hierarchical Action Chunking Transformer with Vector-quantization (HACT-Vq) to efficiently learn temporal multimodality in addition to fine-grained motion. The proposed hierarchical model consists of a high-level policy to make planning for a latent subgoal and style, and a low-level policy to predict an action chunk conditioned with the latent subgoal and style. The latent subgoal and style are trained as discrete representations so that high-level policy can efficiently learn multimodal distributions of demonstrations and retrieve the mode of fast behavior. In experiments, we set up bimanual robots in both simulation and real-world environments, and collected demonstrations with various speeds. The proposed model with the quantized subgoal and style showed the highest success rates with fast imitation behavior. Our code is available at https://github.com/SamsungLabs/hierarchical-act.

This paper proposes a novel approach to address the technical challenges of stable object grasping, particularly in the context of handling tableware in a home environment. Handling tableware is particularly important, yet challenging, due to the flat nature of most tableware objects and the need to maintain a stable posture to prevent spills. To address these challenges, we present three key contributions: 1) a large-scale tableware dataset, not commonly found in the previous datasets; 2) a novel sampling method for stable grasp pose generation; and 3) a multi-modal fusion grasp network that effectively learns 6- DoF grasp pose, including flat objects. Our dataset contains over 45 million grasp poses and 1 million RGBD images captured in 800 scenes, which include randomly selected 10–18 tableware objects under 4 different lighting conditions. The grasp poses in the dataset are generated using a novel sampling method that incorporates geometric analysis to ensure stable grasping with minimal object movement. Furthermore, we design an RGBD fusion grasp network (RGBD-FGN) that can combine information from RGB and depth images considering each characteristic. Our experimental results demonstrate the superior performance of our approach over existing techniques, which is a significant contribution towards developing a multitasking home robot. Our dataset and source code can be accessed at https://github.com/SamsungLabs/RGBD-FGN.

Active camouflage exhibited by certain creatures in nature such as cephalopods has inspired the fabrication of display devices for human-adaptive camouflage technologies. In order to realize that, electrochromic devices (ECDs) have attracted significant attention owing to their low-voltage operations and fast responses. However, the effective utilization of ECDs requires multicolor patterning, durable functioning, and wearable characteristics, simultaneously, but has not been explored. Here, we demonstrate a low-power, deformable, dynamic multicolor electrochromic skin (DMECS) that mimics the multicolor patterning and the active camouflage functionalities of the skins of cephalopods. The electrochromic polymers such as poly(3-hexylthiophene-2,5-diyl), poly[2-methoxy-5-(2-ethylhexyl-oxy)-1,4-phenylenevinylene], and P4a (green color polymer) are used to create purple, orange, and green colors, respectively. An iontronic polymer pump composed of an ionic liquid (1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide) incorporated in thermoplastic polyurethane is used as a deformable and transparent solid-state electrolyte that enables low-voltage (+/- 3 V) operated DMECS with excellent cyclic coloration/bleaching stability (>35,000 s), fast response (similar to 1.75 s), and high durability under repeated 10,000 cycles of compressive force (with a bending radius of 8 mm) and tensile strains (similar to 100% up to 15,000 s). We believe that our DMECS can offer user-controlled selective coloration/bleaching of arbitrary display patterns and open new avenues for next-generation wearable optoelectronics.

A previous trigger assembly with a single embedded permanent magnet has malfunction problems such as the generation of a virtual trigger signal by an external magnetic field. In this study, we improved a trigger assembly, which can minimize the influence of external magnetic field. In order to improve the trigger assembly, two designs were considered. Through an M&S study for the optimal characteristics of Hall-effect sensor, we confirmed that the magnitude of the driving magnetic field of the Hall-effect sensor should be at least 50 gauss for the application of a trigger assembly. The second design was a magnetic bipolar simultaneous recognition system with a time-interval of 10 ms, which occurs when two embedded permanent magnets with different polarities are recognized simultaneously. As a result, the improved trigger assembly, which reflects the two design results, excluded the malfunction of small arms by the external magnetic field without magnetic shielding.

We propose a novel model-free optimal estimation and sensor placement framework for a high-DOF (degree-of-freedom) EKC (elastic kinematic chain) with only a limited number of IMU (inertial measurement unit) sensors based on POD (proper orthogonal decomposition) and MAP (maximum a posteriori) estimation. First, we (off-line) excite the system richly enough, collect the data and perform the POD to extract dominant and non-dominant modes. We then decide the minimum number of IMUs according to the dominant modes, and construct the prior distribution of the output (i.e., top-end position of EKC) based on the singular value of each POD mode. We also formulate the MAP estimation given the prior distribution and different placements of the IMUs and choose the optimal IMU placement to maximize the posterior probability. This optimal placement is then used for real-time output estimation of the EKC. Experiments are also performed to verify the theory.

We propose a novel aerial manipulation platform, an omnidirectional aerial robot, that is capable of omnidirectional wrench generation with opportunistically distributed/aligned Sectional rotors. To circumvent the tight thrust margin and weight budget of currently available rotor and battery technologies, we propose a novel design optimization framework, which maximizes the minimum-guaranteed control force/torque for any attitude while incorporating such important and useful aspects as interrotor aerointerference, anisotropic task requirement, gravity compensation, etc. We also provide a closed-form solution of infinity-norm optimal control allocation to avoid rotor saturation with the tight thrust margin. Further, we elaborate the notion of electronic speed controller induced singularity and devise a novel selective mapping algorithm to substantially subdue its destabilizing effect. Experiments are performed to validate the theory, which demonstrate such capabilities not possible with typical aerial manipulation systems, namely, separate translation and attitude control on SE(3), hybrid pose/wrench control with downward force of 60 N much larger than its own weight (2.6 kg), and peg-in-hole teleoperation with a radial tolerance of 0.5 mm.

A novel user interface design is presented for semi-autonomous teleoperation system consisting of 3-DOF manipulator, 1-DOF linear stage, and flexible beam erected vertically. An LQR(Linear Quadratic Regulator) control designed for vibration suppression of the beam is applied to the slave-robot system autonomously, and a 3-DOF manipulator performs peg-in-hole tasks by the master in a remote place through inverse kinematics-based workspace tracking control. 3D vision transmitted to the master via HMD(Head Mounted Display) device and haptic feedback calculated from contact information are main components of our interface. From the perspective of human perception, optimization for 3D camera placement is performed to maximize human work efficiency. To verify the proposed user interface performance, human subject tests are performed for eight participants by comparing our interface with a 2D webcam interface and/or with a non-haptic feedback interface. Furthermore, in order to make it possible to run the robotic system in any environment, system on-board implementation by using three strain gauges, one IMU (Inertial Measurement Unit), and encoders of actuators is progressed in this paper.

Electrical motor and hydraulic actuation widely-used in robotics are “internal actuation” with their actuators sitting at the joint between two links. This internal actuation is fundamentally limiting to construct a large-size dexterously-articulated robot, since any external force (and its own link weight) is to be accumulated to the base multiplied by the moment arm length, requiring extremely strong/sturdy base actuator/structure as the system size increases. In this paper, we propose a novel robotic system, LASDRA (large-size aerial skeleton with distributed rotor actuation), which, by utilizing distributed rotors as “external actuation”, can overcome this limitation of internal actuation and enables us to realize large-size dexterously-articulated robots. We present its design and modeling, joint locking strategy to increase its loading capability, and also a novel decentralized control scheme to allow for compliant operation with scalability against the number of links. Trajectory tracking and valve turning experiments are also performed to validate the theory.

We synthesised Sb2S3 quantum dot sensitiser-coated mesoporous TiO2 (mp-TiO2) electrodes using chemical bath deposition (CBD) for use in all-solid-state heterojunction (P3HT/PEDOT:PSS/Sb2S3/mp-TiO2) solar cells. The cell processing parameters such as Sb2S3 sensitiser reaction time and annealing temperature were optimised for the synthesis of efficient sensitised photo-anodes. It was found that the microstructure, such as interior pores and surface over-layers of devices, plays an important role in the photovoltaic properties of all-solid-state heterojunction solar cells. Devices synthesised by the optimised process exhibited a power conversion efficiency (PCE) of 4.42%.

Although the Bi2S3 QDs have an ideal bandgap for the efficient light absorbers, the Bi2S3 QDs sensitised solar cells generally indicate low photovoltaic properties owing to the active photocorrosion of Bi2S3 QDs. Herein, we report the enhancement of Bi2S3 QDs sensitised solar cells using the co-sensitisation of Bi2S3 with CdS QDs. By the homogeneous covering as well as effective blocking of photocorrosion and recombination, the CdS/Bi2S3 co-sensitised solar cells showed three-fold increase in efficiency.

We report a defect-engineered self-assembly route to a mesoporous ternary ZnCo2O4/nitrogen-doped graphene nanoarchitecture as an anode material for lithium ion batteries through a hydrothermal and thermal annealing process. A hetero-nanostructure showed flower-like ZnCo2O4 nanosheets which were well dispersed and firmly decorated on nitrogen-doped reduced graphene oxide, as atomic-scale defects such as nitrogen-doped sites and oxygen-functional groups in chemically modified graphene oxide can be more reactive nucleation sites to anchor metallic nanoparticles strongly. Strong synergy between N-doped graphene and ZnCo2O4 is observed as a high-performance anode electrode material for much higher capacity levels and more durable electrochemical stability in lithium ion batteries. The mesoporous nanoarchitecture electrode shows enhanced reversible performance in cyclic anode tests, maintaining a specific energy capacity of 998mAhg−1 after 30cycles at current density of 100mAg−1.

Amphiphilic block-graft copolymers composed of poly(styrene-b-butadiene-b-styrene) (SBS) backbone and poly(oxyethylene methacrylate) (POEM) side chains are synthesized and combined with hydrophilically preformed TiO2 (Pre-TiO2), which works as a structural binder as well as titania source. This results in the formation of crack free, 6-mu m-thick, organized mesoporous TiO2 (OM-TiO2) films via one-step doctor-blading based on self-assembly of SBS-g-POEM as well as preferential interaction of POEM chains with Pre-TiO2. SBS-g-POEM with different numbers of ethylene oxide repeating units, SBS-g-POEM(500) and SBS-g-POEM(950), are used to form OM-TiO2(500) and OM-TiO2(950), respectively. The efficiencies of dye-sensitized solar cells (DSSCs) with a quasi-solid-state polymer electrolyte reach 5.7% and 5.8% at 100 mW/cm(2) for OM-TiO2(500) and OM-TiO2(950), respectively. The surface area of OM-TiO2(950) was greater than that of OM-TiO2(500) but the light reflectance was lower in the former, which is responsible for similar efficiency. Both DSSCs exhibit much higher efficiency than one (4.8%) with randomly-organized particulate TiO2 (Ran-TiO2), which is attributed to the higher dye loading, reduced charge recombination and improved pore infiltration of OM-TiO2. When utilizing poly((1-(4-ethenylphenyl)methyl)-3-butyl-imidazolium iodide) (PEBII) and mesoporous TiO2 spheres as the solid electrolyte and the scattering layer, the efficiency increases up to 7.5%, one of the highest values for N719-based solid-state DSSCs. (C) 2015 Elsevier B.V. All rights reserved.

This paper reports a new type of transmitting mode electrochromic device that uses the high-contrast electrochromism of poly(3,4-ethylenedioxythiophene) (PEDOT) and operates at long-wavelength infrared (8–12 μm) . To maximize the transmittance contrast and transmittance contrast ratio of the device for thermal camouflage, we control the thickness of the thin PEDOT layer from 25 nm to 400 nm and develop a design of grid-type counter electrodes. The cyclability can be greatly improved by selective deposition of the PEDOT film on grid electrodes as an ion storage layer without any loss of overall transmittance. The device with optimized architectures shows a high transmittance contrast ratio of 83 % at a wavelength of 10 μm with a response rate under 1.4 s when alternating voltage is applied. Captured images of an LED lamp behind the device prove the possibility of active, film-type camouflage against thermal detection.

•Light management strategy is crucial for the performance of optoelectronic devices.•Transparent dielectric structures can efficiently control the light behavior in the device without parasitic absorption or electrical property degradation.•Recent transparent dielectric structures and their optoelectronic applications are reviewed.

Detection of an object exposed to natural environmental conditions can be performed by using an IR sensor. In this instance the radiant contrast between the object and the background is an important parameter. There are various ways of estimating the radiant contrast developed for many types of backgrounds. To estimate the radiant contrast accurately it is necessary to reflect the near field background effectively. In this study we propose a radiation contrast estimation method called the inverse distance weighted method and this is compared with the existing methods as the areal intensity method and the row pixel method. Comparisons are made for two different types of backgrounds including the municipal and the oceanic scenes. From this study we conclude that the method called the inverse distance weighted method suggested in this study most effectively reflects the near field effect of the background around the object.

We report the performance improvement of electrochromic devices for modulating the transmittance contrast of long wavelength infrared light between 1.5 and 5.0 ${\mu}m$ based on a double layer of conducting polymers. The device, fabricated with poly(3-hexylthiophene) (P3HT) and poly(3,4-ethylenedioxythiophene) (PEDOT) as the first and second layers, respectively, showed an transmittance contrast of 60% with a response rate under 5 s, which is greater than the transmittance contrast of cells based on only P3HT or PEDOT (approximately 40%).

We demonstrate that topographic artifacts are independent of optical polarizations of incident light in reflection-mode scattering-type scanning near-field optical microscopy. Experiments were performed for incident p- and s-polarizations with respect to the incident plane of a diffraction grating. The experimental results show good agreement with the simulation images. This study may be useful for the investigation of the optical properties of nanostructures affected by incident polarizations in scattering-type near-field microscopy.

Conductive and magnetic microspheres are fabricated by plating of Co–Fe alloy thin films on hollow ceramic microspheres of low density for the application to lightweight microwave absorbers. Metal plating was carried out in a two-step electroless plating process (pre-treatment of sensitizing and subsequent plating). Uniform coating of the film with about 2μm thickness was identified by SEM. High-frequency magnetic and microwave absorbing properties were determined in the rubber composites containing the metal-coated microspheres. Due to the conductive and ferromagnetic behavior of the Co–Fe thin films, high dielectric constant and magnetic loss can be obtained in the microwave frequencies. In particular, the magnetic loss increases with Fe content in the alloy films and its frequency dispersion can be explained by ferromagnetic resonance theory. Due to the electromagnetic properties, high absorption rate and thin matching thickness are predicted in the composite layers containing the metal-coated microspheres of low density (about 0.8g/cc) for the electromagnetic radiation in microwave frequencies.