arXiv_RO
-
Impedance Cloning: Learning Equilibrium Point Parameters for Contact-Rich Manipulation
2026-09-25 05:36:42Hayato Takahashi, Ryoga Oishi, Yuki Kasuga, Toshiaki TsujiAbstract
Contact-rich manipulation requires robots to regulate force against surfaces whose geometry deviates unpredictably from training conditions. Trajectory-based imitation learning, which reproduces observable outputs, breaks down under such shifts. We propose Impedance Cloning, which instead imitates the biomechanical priors that generate motion -- the stiffness and equilibrium point -- and thereby passively absorbs contact uncertainty. Because these parameters encode intent rather than outcome, they generalize across surface geometries where trajectory reproduction does not. We extract them from bilateral teleoperation demonstrations via a particle filter without force/torque sensors and evaluate the framework on two CRANE-X7 manipulators. In a wiping task with joint-space actions, the trajectory-based baseline loses contact below -6 cm, whereas the proposed method maintains a consistent 4-5 N contact force above -6 cm, with a gradual decrease below; with Cartesian-space actions, its force-height slope over 0 to +8 cm is 0.13 +/- 0.03 N/cm, versus 0.34-0.83 N/cm for fixed-impedance baselines. In a pick-and-place task with 10 diverse cups (100 trials), the proposed method succeeds in 84 trials, outperforming the fixed-impedance baseline (74/100) and performing comparably to a variable impedance control baseline (82/100) with one demonstration instead of ten. In a grasping task, the representation reduces torque tracking error with both ILBiT and Mamba backbones, confirming its generality across architectures.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30842
PDF
-
Fast Plans, Faithful Actions: Closing the Planning-Execution Gap in Hierarchical Vision-Language-Action Models
2026-09-25 05:22:23Chuanliang Xie, Boyu Ma, Gen Li, Yizhou Liu, Houwang Chen, Xinyu Zhou, Jianfei YangAbstract
Hierarchical vision-language-action (VLA) systems consist of a high-level vision-language planner and a low-level action expert that generates continuous actions. This hierarchical design has practical value only if the planner can generate plans fast enough to meet real-time control requirements, and the resulting plans actually contribute to the generation of action. We study one such system, a waypoint hierarchy pipeline adapted from $\pi_{0.5}$, and find that neither requirement is satisfied. This baseline relies on token-level autoregressive decoding (Token-AR) to generate a waypoint plan, requiring 57 very expensive vision-language model (VLM) forward passes. However, we find that erasing the waypoint endpoints has little effect on task success. Two findings reveal the misalignment of planner-executor: the planner generates outputs at an excessively fine granularity, and the executor underuses plans as a control condition. We address the latency issue with waypoint-aligned block-autoregressive decoding (Block-AR), and plan underuse issue with normalized goal modulation (NGM), a layer-wise goal path constrained by phase gating and anti-shortcut training so that the waypoint influences action generation maintaining other signals. Our method reduces the maximum number of VLM forward passes from 57 to 8 on LIBERO, including one prefix prefill, and achieves an $8.7\times$ reduction in planning latency on a Rokae dual-arm robot. With normalized goal modulation and anti-shortcut training, Block-AR's success rate on LIBERO-Long increases from 91.0% to 96.2%, while its average success rate across the four suites increases from 95.85% to 98.45%. On three bimanual tasks with this robot, success rates remain comparable across methods.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30833
PDF
-
HIRE: History-Conditioned Interaction Reasoning and High-Rate Execution for Visually Aliased Precision Manipulation
2026-09-25 05:13:23Rongji Li, Wenhao He, Cewu Lu, Xingyu Chen, Xu-Yao ZhangAbstract
Precision manipulation with contact-critical interactions is often history-dependent: visually similar observations can correspond to different latent interaction states and therefore require different actions, while small execution errors can alter task outcomes. Policies relying on the current visual observation alone cannot resolve such ambiguity; force-aware and memory-augmented methods enrich physical or temporal context, while reactive high-rate policies improve local contact response, yet long-horizon temporal reasoning and precision execution remain largely decoupled in existing methods, limiting reliable progression in visually aliased precision manipulation. To bridge this gap, we introduce History-Conditioned Interaction Reasoning and Execution (HIRE), a cross-rate framework comprising a history-conditioned Interaction-State Reasoner (ISR) and a high-rate Interaction-Manifold Executor (IME). ISR encodes ordered wrench history with a temporal wrench encoder and Force Perceiver as persistent physical evidence for state-consistent action generation, while IME structures contact-critical motion into intrinsic progress and transverse correction for precise execution; their cross-rate loop allows the resulting physical traces to inform subsequent reasoning. In real-robot experiments across surface, insertion, and rotational interactions, HIRE achieves at least 90% completion across all evaluated task stages while improving interaction-state disambiguation, execution precision, and generalization. More broadly, HIRE provides a unified reasoning--execution perspective on precision manipulation under history-dependent partial observability, where physical interaction both realizes task intent and reveals latent-state evidence for future decisions. Code will be released upon publication.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30828
PDF
-
SeA-RVINS: Semantic-Aware Tightly Coupled RTK-Visual-Inertial System with Correlation-Preserving Robust Estimation for Urban Navigation
Abstract
Reliable absolute pose estimation in urban environments is undermined by outlier measurements and incorrect temporal associations that can persist in tightly coupled estimators. Global Navigation Satellite System (GNSS) observations provide globally referenced measurements but are prone to multipath effects. Visual-inertial sensing supplies local motion constraints, but false visual associations can corrupt the estimator. We present SeA-RVINS, a fixed-lag factor-graph Real-Time Kinematic (RTK) visual-inertial system for robust urban pose estimation. A semantic-aware learned stereo frontend rejects unreliable tracks before persistent landmarks enter the graph. For double-differenced GNSS measurements, SeA-RVINS applies Dynamic Covariance Scaling through configurable batch, scalar, and latent-pivot robust formulations while retaining the shared-pivot correlation structure. We propose a hybrid ambiguity-continuation strategy that shares one ambiguity state over short arcs with verified continuity and softly links successive arcs through random-walk factors. On an approximately 20-km route from the public TEX-CUP dataset, including about 50\% deep-urban driving, the latent-pivot configuration achieves 100\% availability and a 1.6-m maximum horizontal error, with 96.16\% and 99.90\% of epochs below 1.0 and 1.5 m, respectively. The implementation is released as open-source software
Abstract (translated)
URL
https://arxiv.org/abs/2609.30814
PDF
-
Moving Horizon Estimation for Quadrotors: An $mathcal{L}_1$ Adaptive Optimizer Approach
2026-09-25 04:01:22Thinh Nguyen, Minkyung Kim, Sandeep Banik, Jinrae Kim, Naira HovakimyanAbstract
Moving Horizon Estimation (MHE) is a state estimation method based on finite-horizon optimization that can offer higher accuracy at the cost of increased computation compared to Kalman filter-based approaches. We present a linear smoothing MHE formulation as a dense Quadratic Program (QP), and a solver consisting of a continuous-time Newton's method augmented with the $\mathcal{L}_1$ Adaptive Optimizer ($\mathcal{L}_1$-AO). While MHE is inherently time-varying, conventional approaches treat it as a sequence of independent, time-invariant problems and employ iterative solvers at each time step, which can be both inaccurate and computationally burdensome. In contrast, time-varying solvers track the optimal solution with fewer iterations by exploiting the temporal evolution of the problem, thereby reducing the computational load. In this research, we enhance both the performance and efficiency of MHE through a time-varying solver with an $\mathcal{L}_1$-AO augmentation that compensates for the prediction inaccuracy, which is common in practice due to noisy sensors and the lack of prior knowledge of the system. Simulation results on a quadrotor platform show that the $\mathcal{L}_1$-AO-augmented approach solves the MHE optimization problem more efficiently than the baseline time-invariant solver and achieves higher estimation accuracy under challenging conditions, compared with both the Extended Kalman Filter and the standard MHE.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30777
PDF
-
Design and Characterization of a Variable-Length Continuum Mechanism with Force Locking
2026-09-25 03:37:30Katelyn King, Veronica Fish, Allison M. OkamuraAbstract
The utility of flexible continuum mechanisms for dexterous navigation is often impaired by their low stiffness, making them ineffective at manipulation in high-force scenarios. To address this challenge, we propose a novel continuum mechanism that achieves both flexible and rigid behavior by antagonistic extension and contraction of a rod-driven continuum helical structure. The helical design combines variable-length capacity with force locking for workspace and stiffness enhancement. In this article, we present the detailed design of the proposed mechanism and characterize its performance through experiments that quantify bending and stiffness. The results demonstrate 180 degree bending range of motion with an average distal positioning error of <10%. Further tests demonstrate that force locking directly improves axial stiffness and thus indirectly increases bending stiffness anisotropically, with maximum bending stiffness along load paths with a large axial component. Tensioning the driving rods provides additional stiffness tunability in the force-locked state, where increasing rod tension proportionally increases bending stiffness with a dimensionless gain of 0.56.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30759
PDF
-
Praxis: Distilling Physical Interaction Priors from Egocentric Videos for Generalizable Whole-Body Manipulation
2026-09-25 03:04:09Shuliang He, Ruiyan Xu, Bo Yue, Hengming Zhang, Huayi Zhou, Shuai Wang, Wei-Shi Zheng, Guiliang LiuAbstract
Mobile humanoid manipulation requires both reaching a usable workspace and preserving precise hand-object interactions as object poses and contact conditions change. Learning these behaviors from limited task-specific data remains challenging. To bridge this gap, we introduce Praxis, a whole-body manipulation framework that combines physical interaction priors from one-shot egocentric video demonstrations with closed-loop posture calibration and online perception. The framework coordinates three stages: vision-language-guided navigation toward target objects, closed-loop posture calibration to align the arm-hand workspace, and dexterous manipulation with synchronized upper- and lower-body control. Online visual feedback re-grounds demonstrated interaction geometry under new object poses and scene configurations, while tactile feedback adapts hand motions to actual contact conditions. Each manipulation skill is specified by one human demonstration, without task-specific manipulation-policy retraining. Experiments across five long-horizon manipulation tasks demonstrate spatial, visual, and cross-object generalization, as well as recovery from external physical disturbances across all three stages.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30735
PDF
-
RoboMonitor: Label-Efficient Runtime Monitoring of Robot Task Execution via Predictive Representation Learning
2026-09-25 02:43:24Abhiroop Ajith, Gokul Narayanan, Kyle Coelho, Tingji Zhao, Yash Shahapurkar, Brian Zhu, Melih Erdogan, Ted Krubasik, Constantinos Chamzas, Eugen SolowjowAbstract
Learned robot policies produce actions, but their outputs alone do not establish whether execution is progressing as intended. Robot execution monitoring requires identifying the current execution phase, detecting failures, and recognizing task completion from observations available during execution. Training such monitors requires annotations that are scarce in datasets collected for robot-policy learning. We present RoboMonitor, a label-efficient vision--language execution monitor that learns from these datasets before introducing monitoring supervision. We pre-train on 25 hours of multi-camera trajectories spanning 12 manipulation tasks and two robot embodiments, using action-conditioned future-feature prediction, inverse dynamics, and masked-present prediction. We then transfer the learned visual and context encoders to a causal monitor and apply temporal supervised fine-tuning (Temporal SFT), which combines supervision throughout each observation window with consistency objectives within and across overlapping windows. At deployment, RoboMonitor requires only the task instruction and camera observations. On a four-task monitoring benchmark, RoboMonitor trained with 52 labeled episodes achieves 93.1% mean phase accuracy and 85.9% macro recall over two fine-tuning seeds, exceeding Qwen3-VL and Robometer trained with the same monitoring supervision. Its phase accuracy also exceeds that of both Qwen3-VL and Robometer trained with 100 episodes. A Qwen3-VL ablation shows that Temporal SFT reduces mean spurious phase switching from 15.23% to 4.95%. In closed-loop deployment, the integrated system completes 39 of 40 simulated Toolbox Sorting trials and 35 of 40 real-world Reel Packing trials, with no false recovery triggers observed.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30715
PDF
-
From Visual Search to Movement Control: A Priority Field for Artificial Agents
Abstract
Human spatial attention is widely conceptualized as being guided by a priority map that integrates perceptual salience, current goals, and past experiences. Here, we extend priority-based computation to movement control in artificial agents. We first introduce a lightweight model of visual search based on an integrated priority map. Trained on human saccades, it reproduced key behavioral patterns, including oculomotor suppression and history-driven selection. Extending the search model, we equipped an artificial agent with a priority field and evaluated its performance in a reach-avoid task that required reaching a goal destination while avoiding moving obstacles. Compared with alternative architectures, priority-field agents trained more efficiently and performed better in unseen, complex scenarios, even from simple demonstrations. Adding a simple memory mechanism also produced human-like, history-driven effects in anticipating the likely location of the upcoming goal. These findings suggest that priority-based computation may provide a promising foundation for movement control in artificial agents.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30704
PDF
-
Learning Vision-Based Agile Gap Traversal: Differentiable Simulation with a Warm-Started Critic
2026-09-25 02:07:30Nuthasith Gerdpratoom, Tianchen Sun, Yichao Gao, Lin ZhaoAbstract
Traversing narrow gaps is challenging for autonomous quadrotors, especially when control commands come directly from high-dimensional visual observations. Existing end-to-end methods often rely on behavior cloning or full-rollout backpropagation through time (BPTT) via differentiable simulation, which can limit policy performance or incur high training costs. We propose a two-stage reinforcement learning framework for more efficient ego-centric visuomotor gap-traversal policy training, leveraging quasi-analytical policy gradients (QPG) via differentiable simulation and critic warm-starting. The framework utilizes QPG to avoid backpropagation through visual rendering, reducing computation and memory costs while improving sample efficiency. In the first stage, an expert actor and critic are trained using privileged observations, including gap geometry. Unlike prior gap-traversal approaches, our training utilizing QPG does not require resetting the agent along optimized reference trajectories. In the second stage, a visual policy is trained using binary gap masks from two ego-centric cameras and low-dimensional observations, while its privileged critic is warm-started from the first stage. This substantially improves training efficiency and traversal success compared with cold-starting the critic or using full-rollout BPTT. Our framework does not require retraining the expert actor when system parameters change, enabling more efficient generalization across drone platforms than state-of-the-art visual gap-traversal methods based on action supervision. The learned visual policy also generalizes to gaps with unseen shapes. Extensive real-world experiments further demonstrate robust gap traversal using binary masks rendered online. Beyond gap traversal, the proposed framework is generic and can be extended to other visuomotor robot learning tasks.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30696
PDF
-
Multi-Objective Human-in-the-Loop Bayesian Optimization of a Lower-Limb Exoskeleton
2026-09-25 02:03:29Neil Janwani, Matthew T. Lerner, Aaron J. Young, Maegan TuckerAbstract
Human-in-the-loop optimization (HILO) is a common approach for optimizing the control of assistive devices to account for the wearer's unique biomechanics and subjective preferences. However, despite research suggesting that a person may have a different prioritization of objectives depending on time-varying factors such as the environment, their mood, or energy levels, existing HILO approaches only consider a single objective or enforce a fixed weighting on a set of objectives. Neither approach is capable of representing an individual's preferences over objectives. In this work, we propose Multi-Objective Human-in-the-loop Bayesian Optimization (MO-HILBO), which builds on explicit multi-objective Bayesian optimization to efficiently infer a personalized set of Pareto-optimal controllers. We compare our approach with an existing multi-objective HILO method and experimentally demonstrate MO-HILBO on a lower-limb exoskeleton across two objectives: metabolic cost (efficiency) and ordinal human feedback (comfort). We find that MO-HILBO (1) discovers Pareto-optimal controllers, and (2) that the pairwise ordering of points on the Pareto front itself is consistent with validation trials. Lastly, we open-source mohilo, a Python package for running both HILO and MO-HILBO on wearable devices: this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30695
PDF
-
Can a Robot Read Braille? - Learning to Adapt Contact via Imitation Learning for Tactile Braille Recognition
2026-09-25 01:30:23Xi Chen, Yunlong Shan, Sihan Chen, Jun Hu, Zhongxuan Li, Shiyao Zhang, Sichao Liu, Zhong Zhao, Kosta Jovanovic, Peng ZhouAbstract
For people who are blind, touch provides an essen-tial channel for accessing written information through Braille. Bringing a similar capability to robots requires them not only to recognize tactile patterns, but also to actively establish physical contact that makes those patterns readable. Yet existing robotic Braille readers largely focus on recognition after contact, leaving contact establishment itself insufficiently addressed. We present an adaptive-contact framework for robotic tactile Braille reading that assesses contact quality and physically corrects unsuitable contact before recognition and reconstruc-tion. Multi-Head Policy Learning uses expert-guided contact-adjustment demonstrations to jointly learn contact acceptability and pose corrections. During deployment, the robot iteratively evaluates and re-establishes contact, retaining reliable tactile observations for pose-aware fusion and Braille reconstruction. Across 20 physical Braille plates used for learning and eval-uation, the proposed approach achieves 94.0% tactile quality and 88.6% tactile reconstruction on the ten online-evaluation plates. These results demonstrate the importance of actively establishing readable contact, rather than relying solely on recognition under imperfect tactile observations, for reliable robotic Braille reading.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30676
PDF
-
MR. POP: Multi-Robot Parallel Optimizing Planner for Almost-Surely Asymptotically Optimal Planning
2026-09-25 00:33:33Chih H. Huang, Roy Xing, Brian Plancher, Zachary KingstonAbstract
Finding globally optimal paths remains a fundamental challenge in multi-robot motion planning. Despite acceleration of almost-surely asymptotically optimal (a.s.a.o.) planners via CPU-based parallelism, achieving both probabilistic convergence guarantees and strong computational performance, these algorithms still struggle to scale to multi-robot settings. As such, we introduce MR. POP, a GPU-based a.s.a.o. multi-robot planner based on dRRT and the AO-x meta-algorithm. MR. POP uses large-scale GPU-based SIMT-parallelism to simultaneously run hundreds of roadmap construction and tree search iterations with underlying parallel nearest neighbor search and collision checking operations. We show that this enables MR. POP to become the only planner achieving a 100% solve rate while being faster than state-of-the-art a.s.a.o. planners in multi-robot systems up to 35-DOF. MR. POP also raises the success rate of downstream motion optimizers (e.g., from 4% to 72%), by creating high-quality, diverse seeds that help avoid local minima.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30644
PDF
-
Frequency-Modulated Piezoelectric Haptic Display
Abstract
We present a frequency-modulated (FM) haptic display based on piezoelectric vibrating actuators. Existing haptic displays commonly encode haptic intensity through the deformation amplitude of individual haptic pixels. Although amplitude-modulated (AM) approaches have enabled compact haptic pixels, independently controlling the deformation amplitude of a large number of pixels can require increasingly complex and bulky driving systems, posing challenges for scaling toward high-density, large-area wearable displays. To address this scaling challenge, we investigate an FM design principle in which haptic intensity is encoded through vibration frequency. We further develop a \textit{Shared-Source Frequency Modulation} (SSFM) structure in which multiple haptic pixels are powered by a common power amplifier while their vibration spectrum are controlled individually, reducing the need for independent high-power amplification at each pixel. A proof-of-concept piezoelectric haptic display was built and evaluated on rendering spatial and temporal haptic patterns through volunteer tests. The results show that participants reliably distinguished spatial and temporal patterns encoded using FM principles within the investigated operating range. These findings demonstrate the feasibility of FM-based distributed haptic rendering and suggest a potential pathway toward more compact driving architectures for future high-density, large-area wearable haptic displays.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30626
PDF
-
Audit Before You Commit: Locating Belief Failures in Active Identification for One-Shot Manipulation
Abstract
A robot that probes a few times before one irreversible action, such as tapping a surface before inserting a peg, must decide when the evidence is enough to commit. We argue that this decision rests on two conditions that existing methods do not separate: the belief must still cover the truth in the coordinate that decides the action, and the failure model that scores actions must track realized failure. We audit both conditions separately, offline and with ground truth, on a deployed probe-then-commit pipeline: a particle belief, a scenario failure score, and one commit. On simulated insertion, more taps sharpen the belief while the truth leaves its support on 16.9% of episodes and the failure score turns optimistic by 0.31. Conformal calibration restores coverage but not the decision: confidently wrong instances still pass a confidence gate. The audit's signatures instead point at the observation model, where a hand scan finds a 2.1 mm error in the tap boundary. Correcting that one number cuts failure from 0.354 to 0.112 on untouched instances and transfers unrefitted to a second engine, while in a third engine the same audit suggests an execution-model mismatch instead. Across seven task families in three engines, a few probes at a fixed executor reduce miss or failure. On a physical arm inserting a tool into a rigid pocket by touch, the gain and the audit's two conditions reproduce, and replaying the recorded taps under an injected model error shows the audit's signature on real data. Additional materials are available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30608
PDF
-
Learning-Accelerated Narrow-Phase Collision Detection via Check Ordering for Sampling-Based Motion Planning
2026-09-24 22:28:08Hao Jiang, Yinghan Wang, Jianping He, Xiaoming DuanAbstract
Collision detection is critical for ensuring the safety of planned paths. However, it imposes a non-negligible computational burden on motion planners, motivating extensive studies on collision-detection acceleration. In commonly used phase-based collision-detection methods, the broad phase employs hierarchical structures to rapidly discard object pairs that are clearly collision-free, while the subsequent narrow phase performs detailed collision checks on the remaining object pairs whose collision status cannot be determined by the broad phase. Although these methods effectively reduce the number of detailed checks through broad-phase pruning, the narrow phase is usually executed in the default order returned by the broad phase, with little explicit optimization of the check order. This leaves room for further acceleration, especially in cluttered environments where many object pairs may remain after the broad phase and the narrow phase can account for a significant portion of the total detection time. In this work, we propose a learning-based method to accelerate phase-based collision detection by optimizing the check order in the narrow phase. We first formulate the expected time cost of the narrow phase and derive an optimal check-ordering criterion that minimizes this expectation. Since the priors required by this criterion are difficult to obtain in advance, we design a hypernetwork-based model to predict collision probabilities, which are then used to approximate the optimal check order. The resulting order guides the execution of exact mesh checks in the narrow phase, thereby reducing detection time without replacing the underlying geometric collision checker. Simulation results show that our method effectively accelerates phase-based collision detection and improves the efficiency and success rate of sampling-based motion planning, especially in cluttered environments.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30599
PDF
-
HuGo: LLMs as Whole-Body Policy Code Designers for Humanoid Loco-Manipulation
2026-09-24 22:17:00Seoyeon Choi, Shizhao Ye, Nicholas Bui, Aayushi Shrivastava, Kanghyun Ryu, Dhruva Tirumala, Markus Wulfmeier, Negar MehrAbstract
For humanoids to be useful in everyday environments, they must perform a wide range of tasks that couple locomotion and manipulation. Existing approaches commonly acquire a loco-manipulation policy through reward engineering or demonstrations followed by task-specific training, making it costly to scale to new tasks. In this work, we propose a hierarchical approach to humanoid loco-manipulation that eliminates these per-task requirements. HuGo, Humanoid policy code Generation, uses a Large Language Model (LLM) to generate executable, closed-loop high-level policy code from a task description on top of a frozen low-level whole-body policy. Given the task, observation, and command specifications, the LLM constructs the task logic in code. HuGo then refines the policy from its rollouts using numerical trajectories and selected video frames to produce feedback and targeted code updates. Across five simulation tasks, using two different low-level policies, HuGo substantially outperforms a high-level reinforcement learning baseline and approaches the performance of a demonstration-based baseline. We achieve this level of performance without task-specific reward design or demonstration collection. We further demonstrate zero-shot transfer of simulation-generated policies to hardware and show that applying the same refinement loop to real-world rollouts can further improve transfer performance without expert demonstrations or policy retraining. Project website is this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2609.30594
PDF
-
MOCHA: Multi-Objective Co-Design using Hypernetwork Architectures
2026-09-24 21:31:19Varun Madabushi, Neil Janwani, Maegan TuckerAbstract
In this work, we present MOCHA, the first, to our knowledge, reinforcement learning based approach to computing a family of Pareto-optimal policies across the design space of a robot using a single network. Specifically, MOCHA leverages the hypernetwork architecture to learn a network that produces specialized network parameters optimized for a given objective and parameterized robot design; we term this a multi-objective design hypernetwork (MDH). We demonstrate the capabilities of MDHs to represent a complex family of design-dependent strategies on two distinct robot morphologies, each with six design dimensions and across 2-3 objectives. Moreover, we propose an approach for efficiently producing a Design Pareto set using evolutionary search of the learned policy network, generating the optimal design-policy combination for each objective prioritization. Lastly, we provide an efficient method for computing generalist robot designs which achieve the best cumulative performance across the entire set of objectives.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30570
PDF
-
SoGuDiff: Socially Guided Diffusion for Steerable, Norm-Grounded Robot Navigation
2026-09-24 21:11:47Christian Schaible, Haoran Ji, Yash Vardhan Pant, Stephen L. SmithAbstract
Beyond collision avoidance, socially competent robot navigation requires adherence to implicit social conventions that vary across contexts, cultures, and deployment requirements. Many conventional navigation policies learn a single normative behavior, either through reinforcement learning against a fixed reward function or imitation of human demonstrations, exposing no interface for adjusting that conduct at runtime. We present a diffusion-based navigation framework whose social behavior can be tuned at deployment: a desired style is specified, such as how closely the robot passes, which side it yields to, or how much it defers to groups, and the planner adapts accordingly. Continuous style axes can be followed independently or composed, spanning a behavioral space rather than discrete, primitive-based specifications. A feasibility projection layer separates learned social behavior from kinematic feasibility and collision avoidance. A single-axis sweep illustrates a tradeoff curve that strictly dominates the evaluated fixed-behavior baseline configurations, and stylistic differences are replicated in real-world demonstrations.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30560
PDF
-
Privacy-Preserving Prompted Policy Search for Robotic Control
2026-09-24 21:03:32Ali Irshayyid, Feng Lin, Chong Li, Jun ChenAbstract
Large language models (LLMs) have recently demonstrated promising capabilities as in-context policy optimizers for Reinforcement Learning (RL), enabling policy search driven by both numerical reward signals and natural language reasoning. However, deploying such methods in practice requires transmitting raw policy parameters and rewards history to cloud-based LLM APIs, exposing proprietary control strategies to third-party service providers. To address this issue, this paper introduces Privacy-Preserving Prompted Policy Search (PP-ProPS), a framework that enables LLM-guided policy optimization while keeping policy and environmental parameters confidential. PP-ProPS encodes policy parameters and reward values using secret client-side transformations before they are included in each API request, ensuring that the LLM provider observes only encoded policy parameters and scaled reward information. Furthermore, unlike Vanilla ProPS, the proposed framework does not require the true optimal episodic return to be known or disclosed to the LLM. Beyond protecting the optimization data, PP-ProPS improves the search process in two ways. First, it provides the LLM with individual reward components instead of only a single total return, offering more informative feedback about each candidate policy. Second, it uses a bounded history that prevents the prompt from growing indefinitely, improving search with high-dimensional policies and supporting the use of open-weight LLMs. The proposed PP-ProPS is evaluated on both continuous and discrete control problems spanning Multi-Joint dynamics with Contact (MuJoCo) locomotion, classic control, highway driving, and robotic arm manipulation. Compared to Vanilla ProPS, the proposed PP-ProPS outperforms ProPS in seven of the ten evaluated tasks, and surpasses conventional RL methods including PPO, SAC, and TRPO, in five of the six tasks.
Abstract (translated)
URL
https://arxiv.org/abs/2609.30554
PDF
