arXiv subjects

Wenjie Lu

Publications and source records attributed to Wenjie Lu.

17 recordsLinked to original sources

arXiv2026

SlipSense: Multimodal Tactile Learning for Low-Latency and Generalized Slip Detection

Slip detection is fundamental to dexterous manipulation, yet existing systems often lack precise characterization of detection latency and cross-platform generalization. We present SlipSense, a multimodal tactile slip-detection framework built on TacV5, a compact sensor integrating a $32 \times 32$ piezoresistive array operating at 240 Hz and a 3-axis MEMS accelerometer operating at 8 kHz. The piezoresistive array captures spatial pressure distributions, while the accelerometer captures friction-induced vibrations, providing complementary slip cues. The framework performs modality-specific encoding, intra-sensor fusion, and cross-modal attention with causal temporal prediction at 240 Hz. Experiments on a dataset of 1.4 million frames spanning 37 objects demonstrate the complementarity of the two modalities. SlipSense achieves 96.7% Macro F1 with a false-positive rate below 1.6%, detecting 76% of slip events within 23.1 ms. When trained solely on UMI data, SlipSense generalizes zero-shot to a Tesollo dexterous hand, transferring across unseen objects, distinct sensor units, and robotic platforms without retraining.

cs.RO↗

arXiv2026

Touch2Trace: Tactile-Driven Imitation Learning for Dexterous Cable Tracing

Dexterous manipulation of deformable objects demands continuous fingertip-level regulation of pressure, friction, and incipient slip. We study one of the most challenging cases: dexterous cable tracing, feeding a cable through the hand with repeated pinch-and-curl motions of the thumb and index finger. We introduce Touch2Trace, a tactile-driven imitation-learning system for this task, and provide, to our knowledge, the first systematic real-world characterization of how encoder pretraining, control rate, temporal context, and spatial resolution each shape policy performance. The winning learning recipe combines a tactile encoder pretrained for a custom 32 x 32 piezoresistive sensor (TacV5) via self-supervised learning with a lightweight transformer policy trained on teleoperated demonstrations via behavior cloning, deployed at 60 Hz on a Tesollo DG-5F hand. Tactile feedback without vision or explicit cable-state estimation significantly improves tracing performance versus a proprioception-only baseline: from 0.2 cm to 20.1 cm mean distance and 0% to 93% success rate, with zero-shot transfer to unseen cables and routing conditions. The results quantify the influence of key parameters in tactile-driven systems for reliable dexterous deformable object manipulation.

cs.RO↗

arXiv2026

Mitigating Safety Tax via Distribution-Grounded Refinement in Large Reasoning Models

Safety alignment incurs safety tax that perturbs a large reasoning model's (LRM) general reasoning ability. Existing datasets used for safety alignment for an LRM are usually constructed by distilling safety reasoning traces and answers from an external LRM or human labeler. However, such reasoning traces and answers exhibit a distributional gap with the target LRM that needs alignment, and we conjecture such distributional gap is the culprit leading to significant degradation of reasoning ability of the target LRM. Driven by this hypothesis, we propose a safety alignment dataset construction method, dubbed DGR. DGR transforms and refines an existing out-of-distributional safety reasoning dataset to be aligned with the target's LLM inner distribution. Experimental results demonstrate that i) DGR effectively mitigates the safety tax while maintaining safety performance across all baselines, i.e., achieving \textbf{+30.2\%} on DirectRefusal and \textbf{+21.2\%} on R1-ACT improvement in average reasoning accuracy compared to Vanilla SFT; ii) the degree of reasoning degradation correlates with the extent of distribution shift, suggesting that bridging this gap is central to preserving capabilities. Furthermore, we find that safety alignment in LRMs may primarily function as a mechanism to activate latent knowledge, as a mere \textbf{10} samples are sufficient for activating effective refusal behaviors. These findings not only emphasize the importance of distributional consistency but also provide insights into the activation mechanism of safety in reasoning models.

cs.AI↗

arXiv2026

D-Artemis: A Deliberative Cognitive Framework for Mobile GUI Multi-Agents

Graphical User Interface (GUI) agents aim to automate a wide spectrum of human tasks by emulating user interaction. Despite rapid advancements, current approaches are hindered by several critical challenges: data bottleneck in end-to-end training, high cost of delayed error detection, and risk of contradictory guidance. Inspired by the human cognitive loop of Thinking, Alignment, and Reflection, we present D-Artemis -- a novel deliberative framework in this paper. D-Artemis leverages a fine-grained, app-specific tip retrieval mechanism to inform its decision-making process. It also employs a proactive Pre-execution Alignment stage, where Thought-Action Consistency (TAC) Check module and Action Correction Agent (ACA) work in concert to mitigate the risk of execution failures. A post-execution Status Reflection Agent (SRA) completes the cognitive loop, enabling strategic learning from experience. Crucially, D-Artemis enhances the capabilities of general-purpose Multimodal large language models (MLLMs) for GUI tasks without the need for training on complex trajectory datasets, demonstrating strong generalization. D-Artemis establishes new state-of-the-art (SOTA) results across both major benchmarks, achieving a 75.8% success rate on AndroidWorld and 96.8% on ScreenSpot-V2. Extensive ablation studies further demonstrate the significant contribution of each component to the framework.

cs.AI↗

arXiv2025

UltraHorizon: Benchmarking Agent Capabilities in Ultra Long-Horizon Scenarios

Autonomous agents have recently achieved remarkable progress across diverse domains, yet most evaluations focus on short-horizon, fully observable tasks. In contrast, many critical real-world tasks, such as large-scale software development, commercial investment, and scientific discovery, unfold in long-horizon and partially observable scenarios where success hinges on sustained reasoning, planning, memory management, and tool use. Existing benchmarks rarely capture these long-horizon challenges, leaving a gap in systematic evaluation. To bridge this gap, we introduce \textbf{UltraHorizon} a novel benchmark that measures the foundational capabilities essential for complex real-world challenges. We use exploration as a unifying task across three distinct environments to validate these core competencies. Agents are designed in long-horizon discovery tasks where they must iteratively uncover hidden rules through sustained reasoning, planning, memory and tools management, and interaction with environments. Under the heaviest scale setting, trajectories average \textbf{200k+} tokens and \textbf{400+} tool calls, whereas in standard configurations they still exceed \textbf{35k} tokens and involve more than \textbf{60} tool calls on average. Our extensive experiments reveal that LLM-agents consistently underperform in these settings, whereas human participants achieve higher scores, underscoring a persistent gap in agents' long-horizon abilities. We also observe that simple scaling fails in our task. To better illustrate the failure of agents, we conduct an in-depth analysis of collected trajectories. We identify eight types of errors and attribute them to two primary causes: in-context locking and functional fundamental capability gaps. \href{https://github.com/StarDewXXX/UltraHorizon}{Our code will be available here.}

cs.AI↗

arXiv2025

RUSSO: Robust Underwater SLAM with Sonar Optimization against Visual Degradation

Visual degradation in underwater environments poses unique and significant challenges, which distinguishes underwater SLAM from popular vision-based SLAM on the ground. In this paper, we propose RUSSO, a robust underwater SLAM system which fuses stereo camera, inertial measurement unit (IMU), and imaging sonar to achieve robust and accurate localization in challenging underwater environments for 6 degrees of freedom (DoF) estimation. During visual degradation, the system is reduced to a sonar-inertial system estimating 3-DoF poses. The sonar pose estimation serves as a strong prior for IMU propagation, thereby enhancing the reliability of pose estimation with IMU propagation. Additionally, we propose a SLAM initialization method that leverages the imaging sonar to counteract the lack of visual features during the initialization stage of SLAM. We extensively validate RUSSO through experiments in simulator, pool, and sea scenarios. The results demonstrate that RUSSO achieves better robustness and localization accuracy compared to the state-of-the-art visual-inertial SLAM systems, especially in visually challenging scenarios. To the best of our knowledge, this is the first time fusing stereo camera, IMU, and imaging sonar to realize robust underwater SLAM against visual degradation.

cs.RO↗

arXiv2025

CIRCUIT: A Benchmark for Circuit Interpretation and Reasoning Capabilities of LLMs

The role of Large Language Models (LLMs) has not been extensively explored in analog circuit design, which could benefit from a reasoning-based approach that transcends traditional optimization techniques. In particular, despite their growing relevance, there are no benchmarks to assess LLMs' reasoning capability about circuits. Therefore, we created the CIRCUIT dataset consisting of 510 question-answer pairs spanning various levels of analog-circuit-related subjects. The best-performing model on our dataset, GPT-4o, achieves 48.04% accuracy when evaluated on the final numerical answer. To evaluate the robustness of LLMs on our dataset, we introduced a unique feature that enables unit-test-like evaluation by grouping questions into unit tests. In this case, GPT-4o can only pass 27.45% of the unit tests, highlighting that the most advanced LLMs still struggle with understanding circuits, which requires multi-level reasoning, particularly when involving circuit topologies. This circuit-specific benchmark highlights LLMs' limitations, offering valuable insights for advancing their application in analog integrated circuit design.

cs.LG↗

arXiv2024

Singularity Structures of Linear Inviscid Damping in a Channel

This paper studies singularity structures of the linear inviscid damping of two-dimensional Euler equations in a finite periodic channel. We introduce a recursive definition of singularity structures which characterize the singularities of the spectrum density function from different sources: the free part and the boundary part of the Green function. As an application, we demonstrate that the stream function exhibits smoothness away from the channel's boundary, yet it presents singularities in close proximity to the boundary. The singularities arise due to the interaction of boundary and interior singularities of the spectrum density function. We also show that the behavior of the initial data and background flow have an impact on the regularity of different components of the stream function.

math.AP↗

arXiv2022

Non-cascaded Control Barrier Functions for the Safe Control of Quadrotors

Researchers have developed various cascaded controllers and non-cascaded controllers for the navigation and control of quadrotors in recent years. It is vital to ensure the safety of a quadrotor both in normal state and in abnormal state if a controller tends to make the quadrotor unsafe. To this end, this paper proposes a non-cascaded Control Barrier Function (CBF) for a quadrotor controlled by either cascaded controllers or a non-cascaded controller. Incorporated with a Quadratic Programming (QP), the non-cascaded CBF can simultaneously regulate the magnitude of the total thrust and the torque of the quadrotor determined a controller, so as to ensure the safety of the quadrotor both in normal state and in abnormal state. The non-cascaded CBF establishes a non-conservative forward invariant safe region, in which the controller of a quadrotor is fully or partially effective in the navigation or the pose control of the quadrotor. The non-cascaded CBF is applied to a quadrotor performing trajectory tracking and a quadrotor performing aggressive roll maneuvers in simulations to evaluate the effectiveness of the non-cascaded CBF.

eess.SY↗

arXiv2022

Safe Reinforcement Learning for a Robot Being Pursued but with Objectives Covering More Than Capture-avoidance

Reinforcement Learning (RL) algorithms show amazing performance in recent years, but placing RL in real-world applications such as self-driven vehicles may suffer safety problems. A self-driven vehicle moving to a target position following a learned policy may suffer a vehicle with unpredictable aggressive behaviors or even being pursued by a vehicle following a Nash strategy. To address the safety issue of the self-driven vehicle in this scenario, this paper conducts a preliminary study based on a system of robots. A safe RL framework with safety guarantees is developed for a robot being pursued but with objectives covering more than capture-avoidance. Simulations and experiments are conducted based on the system of robots to evaluate the effectiveness of the developed safe RL framework.

eess.SY↗

arXiv2021

A Dynamics Perspective of Pursuit-Evasion Games of Intelligent Agents with the Ability to Learn

Pursuit-evasion games are ubiquitous in nature and in an artificial world. In nature, pursuer(s) and evader(s) are intelligent agents that can learn from experience, and dynamics (i.e., Newtonian or Lagrangian) is vital for the pursuer and the evader in some scenarios. To this end, this paper addresses the pursuit-evasion game of intelligent agents from the perspective of dynamics. A bio-inspired dynamics formulation of a pursuit-evasion game and baseline pursuit and evasion strategies are introduced at first. Then, reinforcement learning techniques are used to mimic the ability of intelligent agents to learn from experience. Based on the dynamics formulation and reinforcement learning techniques, the effects of improving both pursuit and evasion strategies based on experience on pursuit-evasion games are investigated at two levels 1) individual runs and 2) ranges of the parameters of pursuit-evasion games. Results of the investigation are consistent with nature observations and the natural law - survival of the fittest. More importantly, with respect to the result of a pursuit-evasion game of agents with baseline strategies, this study achieves a different result. It is shown that, in a pursuit-evasion game with a dynamics formulation, an evader is not able to escape from a slightly faster pursuer with an effective learned pursuit strategy, based on agile maneuvers and an effective learned evasion strategy.

eess.SY↗

arXiv2021

Safe Multi-Agent Reinforcement Learning through Decentralized Multiple Control Barrier Functions

Multi-Agent Reinforcement Learning (MARL) algorithms show amazing performance in simulation in recent years, but placing MARL in real-world applications may suffer safety problems. MARL with centralized shields was proposed and verified in safety games recently. However, centralized shielding approaches can be infeasible in several real-world multi-agent applications that involve non-cooperative agents or communication delay. Thus, we propose to combine MARL with decentralized Control Barrier Function (CBF) shields based on available local information. We establish a safe MARL framework with decentralized multiple CBFs and develop Multi-Agent Deep Deterministic Policy Gradient (MADDPG) to Multi-Agent Deep Deterministic Policy Gradient with decentralized multiple Control Barrier Functions (MADDPG-CBF). Based on a collision-avoidance problem that includes not only cooperative agents but obstacles, we demonstrate the construction of multiple CBFs with safety guarantees in theory. Experiments are conducted and experiment results verify that the proposed safe MARL framework can guarantee the safety of agents included in MARL.

cs.MA↗

arXiv2020

Modular Transfer Learning with Transition Mismatch Compensation for Excessive Disturbance Rejection

Underwater robots in shallow waters usually suffer from strong wave forces, which may frequently exceed robot's control constraints. Learning-based controllers are suitable for disturbance rejection control, but the excessive disturbances heavily affect the state transition in Markov Decision Process (MDP) or Partially Observable Markov Decision Process (POMDP). Also, pure learning procedures on targeted system may encounter damaging exploratory actions or unpredictable system variations, and training exclusively on a prior model usually cannot address model mismatch from the targeted system. In this paper, we propose a transfer learning framework that adapts a control policy for excessive disturbance rejection of an underwater robot under dynamics model mismatch. A modular network of learning policies is applied, composed of a Generalized Control Policy (GCP) and an Online Disturbance Identification Model (ODI). GCP is first trained over a wide array of disturbance waveforms. ODI then learns to use past states and actions of the system to predict the disturbance waveforms which are provided as input to GCP (along with the system state). A transfer reinforcement learning algorithm using Transition Mismatch Compensation (TMC) is developed based on the modular architecture, that learns an additional compensatory policy through minimizing mismatch of transitions predicted by the two dynamics models of the source and target tasks. We demonstrated on a pose regulation task in simulation that TMC is able to successfully reject the disturbances and stabilize the robot under an empirical model of the robot system, meanwhile improve sample efficiency.

cs.RO↗

arXiv2020

DOB-Net: Actively Rejecting Unknown Excessive Time-Varying Disturbances

This paper presents an observer-integrated Reinforcement Learning (RL) approach, called Disturbance OBserver Network (DOB-Net), for robots operating in environments where disturbances are unknown and time-varying, and may frequently exceed robot control capabilities. The DOB-Net integrates a disturbance dynamics observer network and a controller network. Originated from conventional DOB mechanisms, the observer is built and enhanced via Recurrent Neural Networks (RNNs), encoding estimation of past values and prediction of future values of unknown disturbances in RNN hidden state. Such encoding allows the controller generate optimal control signals to actively reject disturbances, under the constraints of robot control capabilities. The observer and the controller are jointly learned within policy optimization by advantage actor critic. Numerical simulations on position regulation tasks have demonstrated that the proposed DOB-Net significantly outperforms a conventional feedback controller and classical RL algorithms.

cs.RO↗

arXiv2019

A2: Extracting Cyclic Switchings from DOB-nets for Rejecting Excessive Disturbances

Reinforcement Learning (RL) is limited in practice by its gray-box nature, which is responsible for insufficient trustiness from users, unsatisfied interpretation for human intervention, inadequate analysis for future improvement, etc. This paper seeks to partially characterize the interplay between dynamical environments and the DOB-net. The DOB-net obtained from RL solves a set of Partially Observable Markovian Decision Processes (POMDPs). The transition function of each POMDP is largely determined by the environments, which are excessive external disturbances in this research. This paper proposes an Attention-based Abstraction (A${}^2$) approach to extract a finite-state automaton, referred to as a Key Moore Machine Network (KMMN), to capture the switching mechanisms exhibited by the DOB-net in dealing with multiple such POMDPs. This approach first quantizes the controlled platform by learning continuous-discrete interfaces. Then it extracts the KMMN by finding the key hidden states and transitions that attract sufficient attention from the DOB-net. Within the resultant KMMN, this study found three patterns of cyclic switchings (between key hidden states), showing controls near their saturation are synchronized with unknown disturbances. Interestingly, the found switching mechanism has appeared previously in the design of hybrid control for often-saturated systems. It is further interpreted via an analogy to the discrete-event subsystem in the hybrid control.

cs.RO↗

arXiv2014

An Information Value Function for Nonparametric Gaussian Processes

This paper presents a novel information value function that can be used in online sensor planning to monitor a spatial phenomenon in which the spatial phenomenon is modeled by nonparametric Gaussian processes. The information value function is derived from the Kullback-Leibler (KL) divergence and represents the information value brought by sensor decision. The sensor decision at every time step is to select the sensing location that maximizes the information value function associated with the measurement taken. Gaussian processes (GPs) are employed to obtain the posterior distribution of the spatial phenomenon given a number of sensor measurements, because GPs have sufficient flexibilities to adopt the complexity from data. Furthermore, a greedy algorithm is designed based on the information value function. By comparing the greedy algorithm with the random algorithm, it is shown that the error decreases faster defined as the difference between the estimated posterior distribution and the true distribution of the spatial phenomenon via the greedy algorithm.

cs.IT↗

arXiv2014

Teleparallel Poincaré Cosmology and $Λ$CDM Model

We apply the teleparallelism condition to the Poincaré gauge theory of gravity. The resultant teleparallelized cosmology is completely equivalent to the Friedmann cosmology derived from Einstein's general theory of relativity. The torsion is shown to play the role of the cosmological constant driving the cosmic acceleration. We then extend such theory to include the effect of spin and explore the possibility of accounting for the current accelerating universe by a spinning dark energy.

gr-qc↗