Search

A critical goal in robotics and autonomy is to teach robots to adapt to real-world collaborative tasks, particularly in automatic assembly. The ability of a robot to understand the original intent of an incomplete assembly and complete missing features without human instruction is valuable but challenging. This paper introduces 3D combinatorial assembly completion, which is demonstrated using combinatorial unit primitives (i.e., Lego bricks). Combinatorial assembly is challenging due to the possible assembly combinations and complex physical constraints (e.g., no brick collisions, structure stability, inventory constraints, etc.). To address these challenges, we propose a two-part deep reinforcement learning (DRL) framework that tackles teaching the robot to understand the objective of an incomplete assembly and learning a construction policy to complete the assembly. The robot queries a stable object library to facilitate assembly inference and guide learning. In addition to the robot policy, an action mask is developed to rule out invalid actions that violate physical constraints for object-oriented construction. We demonstrate the proposed framework's feasibility and robustness in a variety of assembly scenarios in which the robot satisfies real-life assembly with respect to both solution and runtime quality. Furthermore, results demonstrate that the proposed framework effectively infers and assembles incomplete structures for unseen and unique object types., Comment: Submitted to 2025 American Control Conference (ACC)

Combinatorial assembly uses standardized unit primitives to build objects that satisfy user specifications. This paper studies assembly sequence planning (ASP) for physical combinatorial assembly. Given the shape of the desired object, the goal is to find a sequence of actions for placing unit primitives to build the target object. In particular, we aim to ensure the planned assembly sequence is physically executable. However, ASP for combinatorial assembly is particularly challenging due to its combinatorial nature. To address the challenge, we employ deep reinforcement learning to learn a construction policy for placing unit primitives sequentially to build the desired object. Specifically, we design an online physics-aware action mask that filters out invalid actions, which effectively guides policy learning and ensures violation-free deployment. In the end, we apply the proposed method to Lego assembly with more than 250 3D structures. The experiment results demonstrate that the proposed method plans physically valid assembly sequences to build all structures, achieving a $100\%$ success rate, whereas the best comparable baseline fails more than $40$ structures. Our implementation is available at \url{https://github.com/intelligent-control-lab/PhysicsAwareCombinatorialASP}.

Electronic properties of crystals can be manipulated using spatially periodic modulations. Long-wavelength, incommensurate modulations are of particular interest, exemplified recently by moir\'e patterned van der Waals (vdW) heterostructures. Bulk vdW superlattices hosting interfaces between clean 2D layers represent scalable bulk analogs of vdW heterostructures and present a complementary venue to explore incommensurately modulated 2D states. Here we report the bulk vdW superlattice SrTa$_2$S$_5$ realizing an incommensurate 1D modulation of 2D transition metal dichalcogenide (TMD) $H$-TaS$_2$ layers. High-quality electronic transport in the $H$-TaS$_2$ layers, evidenced by quantum oscillations, is made anisotropic by the modulation and shows commensurability oscillations akin to lithographically modulated 2D systems. We also find unconventional, clean-limit superconductivity (SC) in SrTa$_2$S$_5$ with a pronounced suppression of interlayer coherence relative to intralayer coherence. Such a hierarchy can arise from pair-density wave (PDW) SC with mismatched spatial arrangement in adjacent superconducting layers. Examining the in-plane magnetic field $H_{ab}$ dependence of interlayer critical current density $J_c$, we find anisotropy with respect to $H_{ab}$ orientation: $J_c$ is maximized (minimized) when $H_{ab}$ is perpendicular (parallel) to the stripes, consistent with 1D PDW SC. From diffraction we find the structural modulation is shifted between adjacent $H$-TaS$_2$ layers, suggesting mismatched 1D PDW is seeded by the striped structure. With a high-mobility Fermi liquid in a coherently modulated structure, SrTa$_2$S$_5$ is a promising host for novel phenomena anticipated in clean, striped metals and superconductors. More broadly, SrTa$_2$S$_5$ establishes bulk vdW superlattices as macroscopic platforms to address long-standing predictions for modulated electronic phases., Comment: 19 pages, 4 figures

Recent advancements in manufacturing have a growing demand for fast, automatic prototyping (i.e. assembly and disassembly) capabilities to meet users' needs. This paper studies automatic rapid LEGO prototyping, which is devoted to constructing target LEGO objects that satisfy individual customization needs and allow users to freely construct their novel designs. A construction plan is needed in order to automatically construct the user-specified LEGO design. However, a freely designed LEGO object might not have an existing construction plan, and generating such a LEGO construction plan requires a non-trivial effort since it requires accounting for numerous constraints (e.g. object shape, colors, stability, etc.). In addition, programming the prototyping skill for the robot requires the users to have expert programming skills, which makes the task beyond the reach of the general public. To address the challenges, this paper presents a simulation-aided learning from demonstration (SaLfD) framework for easily deploying LEGO prototyping capability to robots. In particular, the user demonstrates constructing the customized novel LEGO object. The robot extracts the task information by observing the human operation and generates the construction plan. A simulation is developed to verify the correctness of the learned construction plan and the resulting LEGO prototype. The proposed system is deployed to a FANUC LR-mate 200id/7L robot. Experiments demonstrate that the proposed SaLfD framework can effectively correct and learn the prototyping (i.e. assembly and disassembly) tasks from human demonstrations. And the learned prototyping tasks are realized by the FANUC robot.

We present a novel method for fabricating highly customizable three-dimensional structures hosting quantum sensors based on Nitrogen Vacancy (NV) centers using two-photon polymerization. This approach overcomes challenges associated with structuring traditional single-crystal quantum sensing platforms and enables the creation of complex, fully three-dimensional, sensor assemblies with sub-microscale resolutions (down to 400 nm) and large fields of view (>1 mm). By embedding NV center-containing nanoparticles in exemplary structures, we demonstrate high sensitivity optical sensing of temperature and magnetic fields at the microscale. Our work showcases the potential for integrating quantum sensors with advanced manufacturing techniques, facilitating the incorporation of sensors into existing microfluidic and electronic platforms, and opening new avenues for widespread utilization of quantum sensors in various applications.

We show that a generalized version of the 4-point string theory KLT double-copy map is the most general solution to the minimal-rank double-copy bootstrap in effective field theory. This follows from significant restrictions of the 4-point map resulting from the 6-point bootstrap analysis. The generalized 4-point double-copy map is defined by a function with only two parameters times a simple function that is symmetric in $s,t,u$. The two parameters can be interpreted as independent choices of $\alpha'$, one for each of the two sets of amplitudes double-copied with the map. Specifically, each of those two sets of amplitudes must obey either the string monodromy relations or the field theory KK & BCJ relations; there are no other options. We propose a closed form of the new double-copy map that interpolates between the original KLT string double-copy and the open & closed string period integrals. The construction clarifies the "single-valued projection" property of the Riemann zeta-function values for the 4-point string theory double copy., Comment: 8 pages, LCTP-23-01

String monodromy is a set of linear relations among open string tree amplitudes with different orderings of the vertex operators. In this Letter, we show how these intrinsically stringy relations emerge in low-energy effective field theory from the assumptions of locality and the field theory Kleiss-Kuijf (KK) and Bern-Carrasco-Johansson (BCJ) relations. Specifically, we study the bi-adjoint scalar model effective field theory (BAS EFT). We impose the field theory KK and BCJ relations on one of the two color-orderings of the BAS EFT amplitudes and show that the second color-ordering has an emergent set of linear relations that, as checked at 4-point to 36-derivative order, are exactly the monodromy relations. The 4-point results depend on a delicate interplay between consistent factorization of 6-point BAS EFT amplitudes and the 6-point KK BCJ conditions. As a consequence of the analysis, the 4-point Z-theory tree amplitudes are bootstrapped up to a symmetric function in $s,t,u$ whose simple exponentiated form has free parameters that capture all the odd zeta values when matched to Z-theory., Comment: 5 pages, 1 figure

Encoder-decoder-based recurrent neural network (RNN) has made significant progress in sequence-to-sequence learning tasks such as machine translation and conversational models. Recent works have shown the advantage of this type of network in dealing with various time series forecasting tasks. The present paper focuses on the problem of multi-horizon short-term load forecasting, which plays a key role in the power system's planning and operation. Leveraging the encoder-decoder RNN, we develop an attention model to select the relevant features and similar temporal information adaptively. First, input features are assigned with different weights by a feature selection attention layer, while the updated historical features are encoded by a bi-directional long short-term memory (BiLSTM) layer. Then, a decoder with hierarchical temporal attention enables a similar day selection, which re-evaluates the importance of historical information at each time step. Numerical results tested on the dataset of the global energy forecasting competition 2014 show that our proposed model significantly outperforms some existing forecasting schemes.

The origin of many elements of the periodic table remains an unsolved problem. While many nucleosynthetic channels are broadly understood, significant uncertainties remain regarding certain groups of elements such as the intermediate and rapid neutron-capture processes, the p-process, or the origin of odd-Z elements in the most metal-poor stars. Canada has a long tradition of leadership in nuclear astrophysics, dating back to the work of Alastair Cameron in the 1950s. Recent faculty hires have further boosted activity in the field, including transient observation and theory, survey science on galactic nucleosynthesis, and nuclear experiments. This white paper contains a brief overview of recent activity in the community, highlighting strengths in each sub-field, and provides recommendations to improve interdisciplinary collaboration. Sustaining Canadian leadership in the next decade will require, on the observational side, access to transient and non-transient surveys like LSST, SKA, or MSE, support for target-of-opportunity observing in current and future Canadian telescopes, and participation in next-generation X-ray telescopes such as ATHENA. State-of-the-art theoretical predictions will require an ambitious succession plan for the Niagara supercomputer to support large parallel jobs. We propose a funding instrument for postdoctoral training that reflects the interdisciplinary nature of nuclear astrophysics research, and the creation of a national collaborative funding program that allows for joint projects and workshop organization., Comment: White paper submitted to the Canadian Long Range Plan 2020. Minor formatting changes relative to submitted version

The present work reports the results of 30S radioactive beam development for a future experiment directly measuring data to extrapolate the 30S(alpha,p) stellar reaction rate in Type I X-ray bursts, a phenomena where nuclear explosions occur repeatedly on the surface of accreting neutron stars. We produce the radioactive ion 30S via the 3He(28Si,30S)n reaction, by bombarding a cryogenically cooled target of 3He at 400 Torr and 80 K with 28Si beams of 6.9 and 7.54 MeV/u. In order to perform a successful future experiment which allows us to calculate the stellar 30S(alpha, p) reaction rate, Hauser-Feshbach calculations indicate we require a 30S beam of ~10^5 particles per second at ~32 MeV. Based on our recent beam development experiments in 2006 and 2008, it is believed that such a beam may be fabricated in 2009 according to the results presented. We plan to measure the 4He(30S,p) cross-section at astrophysical energies in 2009, and some brief remarks on the planned (alpha,p) technique are also elucidated., Comment: 5 pages, 3 figures. Submitted on Dec. 25, 2008 for the Proceedings of the 10th Symposium on Nuclei in the Cosmos, Mackinac Island, Michigan, USA, 27 July - 1 August, 2008. Accepted for publication Jan. 09, 2009

Future Mars rover missions will require more advanced onboard autonomy for increased scientific productivity and reduced mission operations cost. One such form of autonomy can be achieved by targeting precise science measurements to be made in a single command uplink cycle. In this paper we present an overview of our solution to the subproblems of navigating a rover into place for microscopic imaging, mapping an instrument target point selected by an operator using far away science camera images to close up hazard camera images, verifying the safety of placing a contact instrument on a sample or finding nearby safe points, and analyzing the data that comes back from the rover. The system developed includes portions used in the Multiple Target Single Cycle Instrument Placement demonstration at NASA Ames in October 2004, and portions of the MI Toolkit delivered to the Athena Microscopic Imager Instrument Team for the MER mission still operating on Mars today. Some of the component technologies are also under consideration for MSL mission infusion.