Autonomous Robots最新文献

In collective motion, perceptually-limited individuals move in an ordered manner, without centralized control. The perception of each individual is highly localized, as is its ability to interact with others. While natural collective motion is robust, most artificial swarms are brittle. This particularly occurs when vision is used as the sensing modality, due to ambiguities and information-loss inherent in visual perception. This paper presents mechanisms for robust collective motion inspired by studies of locusts. First, we develop a robust distance estimation method that combines visually perceived horizontal and vertical sizes of neighbors. Second, we introduce intermittent locomotion as a mechanism that allows robots to reliably detect peers that fail to keep up, and disrupt the motion of the swarm. We show how such faulty robots can be avoided in a manner that is robust to errors in classifying them as faulty. Through extensive physics-based simulation experiments, we show dramatic improvements to swarm resilience when using these techniques. We show these are relevant to both distance-based Avoid–Attract models, as well as to models relying on Alignment, in a wide range of experiment settings.

New settlements in remote environments require terrain modification, a task well suited for autonomous multi-robot systems. Simple, robust earthmover robots offer an inexpensive and scalable alternative to sophisticated construction robots. We present a mathematical model for such robots modifying continuous granular structures in 2D and develop both centralized and decentralized planning algorithms to achieve user-defined construction goals. These algorithms decompose long-horizon tasks into subtasks solvable using optimal transport theory and Wasserstein geodesics. Simulations across 100 randomly generated tasks show that a centralized controller with global information achieves on average 85% and 92% construction progress on untraversable and traversable terrains respectively, even with action noise. Multiple robots reduce overall travel distance by 70%, important because motion over the structure also disturbs it. The distributed algorithm—without global information—matches centralized performance on traversable terrain, reaching 93% progress. Increasing robot numbers accelerates convergence, lowers moved material, and raises convergence rates, though congestion can increase total travel distance. These results indicate that simple earthmover robots hold promise for construction tasks ranging from extraterrestrial habitat preparation to coastal protective berms.

This paper considers the problem of autonomous mobile robot navigation in unknown environments with moving obstacles. We propose a new method to achieve environment-aware safe tracking (EAST) of robot motion plans that integrates an obstacle clearance cost for path planning, a convex reachable set for robot motion prediction, and safety constraints for dynamic obstacle avoidance. EAST adapts the motion of the robot according to the locally sensed environment geometry and dynamics, leading to fast motion in wide open areas and cautious behavior in narrow passages or near moving obstacles. Our control design uses a reference governor, a virtual dynamical system that guides the robot’s motion and decouples the path tracking and safety objectives. While reference governor methods have been used for safe tracking control in static environments, our key contribution is an extension to dynamic environments using convex optimization with control barrier function (CBF) constraints. Thus, our work establishes a connection between reference governor techniques and CBF techniques for safe control in dynamic environments. We validate our approach in simulated and real-world environments, featuring complex obstacle configurations and natural dynamic obstacle motion.

When legged robots impact their environment executing dynamic motions, they undergo large changes in their velocities in a short amount of time. Measuring and applying feedback to these velocities is challenging, further complicated by uncertainty in the impact model and impact timing. This work proposes a general framework for adapting feedback control during impact by projecting the control objectives to a subspace that is invariant to the impact event. The resultant controller is robust to uncertainties in the impact event while maintaining maximum control authority over the impact-invariant subspace. We demonstrate the improved performance of the projection over other commonly used heuristics on a walking controller for a planar five-link-biped. The projection is also applied to jumping, box jumping, and running controllers for the compliant 3D bipedal robot, Cassie. The modification is easily applied to these various controllers and is a critical component to deploying on the physical robot. Code and video of the experiments are available at https://impact-invariant-control.github.io/.

Maintaining a robust communication network is crucial for the success of multi-robot online task planning. A key capability of such systems is the ability to repair the communication topology in the event of robot failures, thereby ensuring continued coordination. In this paper, we address the Fast k-Connectivity Restoration (FCR) problem, which seeks to restore a network’s k-connectivity with minimal robot movement. Here, a k-connected network refers to a topology that remains connected despite the removal of up to (k-1) nodes. We first formulate the FCR problem as a Quadratically Constrained Program (QCP), which yields optimal solutions but is computationally intractable for large-scale instances. To overcome this limitation, we propose EA-SCR, a scalable algorithm grounded in graph-theoretic principles, which leverages global network information to guide robot movements. Furthermore, we develop a learning-based approach, GNN-EA-SCR, which employs aggregation graph neural networks to learn a decentralized counterpart of EA-SCR, relying solely on local information exchanged among neighboring robots. Through empirical evaluation, we demonstrate that EA-SCR achieves solutions within 10% of the optimal while being orders of magnitude faster. Additionally, EA-SCR surpasses existing methods by 30% in terms of the FCR distance metric. For the learning-based solution, GNN-EA-SCR, we show it attains a success rate exceeding 90% and exhibits comparable maximum robot movement to EA-SCR.

Learning from Demonstration (LfD) algorithms seek to enable end-users to teach robots new skills through human demonstration of a task. Previous studies have analyzed how robot failure affects human trust, but not in the context of the human teaching the robot. In this paper, we investigate how human teachers react to robot failure in an LfD setting. We conduct a study in which participants teach a robot how to complete three tasks, using one of three instruction methods, while the robot is pre-programmed to either succeed or fail at the task. We find that when the robot fails, people trust the robot less ((p<.001)) and themselves less ((p=.003)) and they believe that others will trust them less ((p<.001)). Human teachers also have a lower impression of the robot and themselves ((p<.001)) and found the task more difficult when the robot fails ((p<.001)). Motion capture was found to be a less difficult instruction method than teleoperation ((p=.012)), while kinesthetic teaching gave the teachers the lowest impression of themselves compared to teleoperation ((p=.035)) and motion capture ((p=.001)). Importantly, a mediation analysis showed that people’s trust in themselves is heavily mediated by what they think that others—including the robot—think of them ((p<.001)). These results provide valuable insights to improving the human–robot relationship for LfD.

We introduce GEOTACT, the first robotic system capable of grasping and retrieving objects of potentially unknown shapes buried in a granular environment. While important in many applications, ranging from mining and exploration to search and rescue, this type of interaction with granular media is difficult due to the uncertainty stemming from visual occlusion and noisy contact signals. To address these challenges, we use a learning method relying exclusively on touch feedback, trained end-to-end with simulated sensor noise. We show that our problem formulation leads to the natural emergence of learned pushing behaviors that the manipulator uses to reduce uncertainty and funnel the object to a stable grasp despite spurious and noisy tactile readings. We introduce a training curriculum that bootstraps learning in simulated granular environments, enabling zero-shot transfer to real hardware. Despite being trained only on seven objects with primitive shapes, our method is shown to successfully retrieve 35 different objects, including rigid, deformable, and articulated objects with complex shapes.

This paper reports an attempt to model the system dynamics and estimate both the unknown internal control input and the state of a recently developed marine autonomous vehicle, the Jaiabot. Although the Jaiabot has shown promise in many applications, process and sensor noise necessitates state estimation and noise filtering. In this work, we present the first surge and heading linear dynamical model for Jaiabots derived from real data collected during field testing. An adaptive input estimation algorithm is implemented to accurately estimate the control input and hence the state. For validation, this approach is compared to the classical Kalman filter, highlighting its advantages in handling unknown control inputs.

This work addresses the challenge of grasping a target object in cluttered environments, even when it is partially visible or fully occluded. The proposed approach enables the manipulator to learn a sequence of strategic pushing actions that rearrange the scene to make the target object graspable. Our pipeline integrates image morphological processing with deep reinforcement learning (DRL), using GR-ConvNet to predict grasp points for the target. When the object is considered ungraspable, a soft actor-critic (SAC) model guides optimal pushing actions. A novel clutter map is introduced, encoding environmental clutter into a quantitative score that informs the decision-making process. The system shows improved performance with a discount factor ((gamma )) of 0.9, demonstrated through comparative analysis with and without the clutter map. We also compare models trained in discrete versus continuous action spaces to evaluate the impact of action space on DRL effectiveness. The pipeline generalizes well to diverse objects and integrates directly with hardware, requiring no additional training for real-world deployment.