Robotics (cs.RO)

Abstract: In dynamic motion generation tasks, including contact and collisions, small changes in policy parameters can lead to extremely different returns. For example, in soccer, the ball can fly in completely different directions with a similar heading motion by slightly changing the hitting position or the force applied to the ball or when the friction of the ball varies. However, it is difficult to imagine that completely different skills are needed for heading a ball in different directions. In this study, we proposed a multitask reinforcement learning algorithm for adapting a policy to implicit changes in goals or environments in a single motion category with different reward functions or physical parameters of the environment. We evaluated the proposed method on the ball heading task using a monopod robot model. The results showed that the proposed method can adapt to implicit changes in the goal positions or the coefficients of restitution of the ball, whereas the standard domain randomization approach cannot cope with different task settings.

Abstract: The ability to grasp and manipulate small objects in cluttered environments remains a significant challenge. This paper introduces a novel approach that utilizes a tactile sensor-equipped gripper with eight degrees of freedom to overcome these limitations. We employ DenseTact 2.0 for the gripper, enabling precise control and improved grasp success rates, particularly for small objects ranging from 5mm to 25mm. Our integrated strategy incorporates the robot arm, gripper, and sensor to manipulate and orient small objects for subsequent classification effectively. We contribute a specialized dataset designed for classifying these objects based on tactile sensor output and a new control algorithm for in-hand orientation tasks. Our system demonstrates 88% of successful grasp and successfully classified small objects in cluttered scenarios.

Abstract: As multi-robot systems continue to advance and become integral to various applications, managing conflicts and ensuring secure access control are critical challenges that need to be addressed. Access control is essential in multi-robot systems to ensure secure and authorized interactions among robots, protect sensitive data, and prevent unauthorized access to resources. This paper presents a novel framework for customizable conflict resolution and attribute-based access control in multi-robot systems for ROS 2 leveraging the Hyperledger Fabric blockchain. We introduce an attribute-based access control (ABAC) Fabric-ROS 2 bridge to enable secure communication and control between users and robots. By defining conflict resolution policies based on task priorities, robot capabilities, and user-defined constraints, our framework offers a flexible way to resolve conflicts. Additionally, it incorporates attribute-based access control, granting access rights based on user and robot attributes. ABAC offers a modular approach to control access compared to existing access control approaches in ROS 2, such as SROS2. Through this framework, multi-robot systems can be managed efficiently, securely, and adaptably, ensuring controlled access to resources and managing conflicts. Our experimental evaluation shows that our framework marginally improves latency and throughput over exiting Fabric and ROS 2 integration solutions. At higher network load, it is the only solution to operate reliably without a diverging transaction commitment latency. We also demonstrate how conflicts arising from simultaneous control or a robot by two users are resolved in real-time and motion distortion is effectively eliminated.

Abstract: Autonomous exploration requires the robot to explore an unknown environment while constructing an accurate map with the SLAM (Simultaneous Localization and Mapping) techniques. Without prior information, the exploratory performance is usually conservative due to the limited planning horizon. This paper exploits a prior topo-metric graph of the environment to benefit both the exploration efficiency and the pose graph accuracy in SLAM. Based on recent advancements in relating pose graph reliability with graph topology, we are able to formulate both objectives into a SLAM-aware path planning problem over the prior graph, which finds a fast exploration path with informative loop closures that globally stabilize the pose graph. Furthermore, we derive theoretical thresholds to speed up the greedy algorithm to the problem, which significantly prune non-optimal loop closures in iterations. The proposed planner is incorporated into a hierarchical exploration framework, with flexible features including path replanning and online prior map update that adds additional information to the prior graph. Extensive experiments indicate that our method has comparable exploration efficiency to others while consistently maintaining higher mapping accuracy in various environments. Our implementations will be open-source on GitHub.

Abstract: We propose augmenting the empathetic capacities of social robots by integrating non-verbal cues. Our primary contribution is the design and labeling of four types of empathetic non-verbal cues, abbreviated as SAFE: Speech, Action (gesture), Facial expression, and Emotion, in a social robot. These cues are generated using a Large Language Model (LLM). We developed an LLM-based conversational system for the robot and assessed its alignment with social cues as defined by human counselors. Preliminary results show distinct patterns in the robot's responses, such as a preference for calm and positive social emotions like 'joy' and 'lively', and frequent nodding gestures. Despite these tendencies, our approach has led to the development of a social robot capable of context-aware and more authentic interactions. Our work lays the groundwork for future studies on human-robot interactions, emphasizing the essential role of both verbal and non-verbal cues in creating social and empathetic robots.

Abstract: Game theory offers an interpretable mathematical framework for modeling multi-agent interactions. However, its applicability in real-world robotics applications is hindered by several challenges, such as unknown agents' preferences and goals. To address these challenges, we show a connection between differential games, optimal control, and energy-based models and demonstrate how existing approaches can be unified under our proposed Energy-based Potential Game formulation. Building upon this formulation, this work introduces a new end-to-end learning application that combines neural networks for game-parameter inference with a differentiable game-theoretic optimization layer, acting as an inductive bias. The experiments using simulated mobile robot pedestrian interactions and real-world automated driving data provide empirical evidence that the game-theoretic layer improves the predictive performance of various neural network backbones.

Abstract: Shared benchmark problems have historically been a fundamental driver of progress for scientific communities. In the context of academic conferences, competitions offer the opportunity to researchers with different origins, backgrounds, and levels of seniority to quantitatively compare their ideas. In robotics, a hot and challenging topic is sim2real-porting approaches that work well in simulation to real robot hardware. In our case, creating a hybrid competition with both simulation and real robot components was also dictated by the uncertainties around travel and logistics in the post-COVID-19 world. Hence, this article motivates and describes an aerial sim2real robot competition that ran during the 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems, from the specification of the competition task, to the details of the software infrastructure supporting simulation and real-life experiments, to the approaches of the top-placed teams and the lessons learned by participants and organizers.

Abstract: Target detection is a basic task to divide the object types in the orchard point cloud global map, which is used to count the overall situation of the orchard. And provide necessary information for unmanned navigation planning of agricultural vehicles. In order to divide the fruit trees and the ground in the point cloud global map of the standardized orchard, and provide the orchard overall information for the path planning of autonomous vehicles in the natural orchard environment. A fruit tree detection method based on the Yolo-V7 network is proposed, which can effectively detect fruit tree targets from multi-sensor fused radar point cloud, reduce the 3D point cloud information of the point cloud map to 2D for the fruit tree point cloud in the Yolo-V7 network detection map, and project the prediction results into the point cloud map. Generally, the target detection network based on PointNet has the problem of low speed and large computational load. The method proposed in this paper is fast and low computational load and is suitable for deployment in mobile robots. From the experimental results, the recall rate and accuracy rate of the proposed method in orchard fruit tree detection are 0.4 and 0.696 respectively, and its weight and reasoning time are 7.4 M and 28 ms respectively. The experimental results show that this method can achieve the robustness and efficiency of real-time detection of orchard fruit trees.

Abstract: We present our approach for the development, validation and deployment of a data-driven decision-making function for the automated control of a vehicle. The decisionmaking function, based on an artificial neural network is trained to steer the mobile robot SPIDER towards a predefined, static path to a target point while avoiding collisions with obstacles along the path. The training is conducted by means of proximal policy optimisation (PPO), a state of the art algorithm from the field of reinforcement learning. The resulting controller is validated using KPIs quantifying its capability to follow a given path and its reactivity on perceived obstacles along the path. The corresponding tests are carried out in the training environment. Additionally, the tests shall be performed as well in the robotics situation Gazebo and in real world scenarios. For the latter the controller is deployed on a FPGA-based development platform, the FRACTAL platform, and integrated into the SPIDER software stack.

Abstract: We propose a learning-based system for enabling quadrupedal robots to manipulate large, heavy objects using their whole body. Our system is based on a hierarchical control strategy that uses the deep latent variable embedding which captures manipulation-relevant information from interactions, proprioception, and action history, allowing the robot to implicitly understand object properties. We evaluate our framework in both simulation and real-world scenarios. In the simulation, it achieves a success rate of 93.6 % in accurately re-positioning and re-orienting various objects within a tolerance of 0.03 m and 5 {\deg}. Real-world experiments demonstrate the successful manipulation of objects such as a 19.2 kg water-filled drum and a 15.3 kg plastic box filled with heavy objects while the robot weighs 27 kg. Unlike previous works that focus on manipulating small and light objects using prehensile manipulation, our framework illustrates the possibility of using quadrupeds for manipulating large and heavy objects that are ungraspable with the robot's entire body. Our method does not require explicit object modeling and offers significant computational efficiency compared to optimization-based methods. The video can be found at $\href{https://youtu.be/fO_PVr27QxU}{this \ http \ URL}$.

Abstract: This paper describes a framework for learning Automated Vehicles (AVs) driver models via knowledge sharing between vehicles and personalization. The innate variability in the transportation system makes it exceptionally challenging to expose AVs to all possible driving scenarios during empirical experimentation or testing. Consequently, AVs could be blind to certain encounters that are deemed detrimental to their safe and efficient operation. It is then critical to share knowledge across AVs that increase exposure to driving scenarios occurring in the real world. This paper explores a method to collaboratively train a driver model by sharing knowledge and borrowing strength across vehicles while retaining a personalized model tailored to the vehicle's unique conditions and properties. Our model brings a federated learning approach to collaborate between multiple vehicles while circumventing the need to share raw data between them. We showcase our method's performance in experimental simulations. Such an approach to learning finds several applications across transportation engineering including intelligent transportation systems, traffic management, and vehicle-to-vehicle communication. Code and sample dataset are made available at the project page https://github.com/wissamkontar.

Abstract: Visual active tracking is a growing research topic in robotics due to its key role in applications such as human assistance, disaster recovery, and surveillance. In contrast to passive tracking, active tracking approaches combine vision and control capabilities to detect and actively track the target. Most of the work in this area focuses on ground robots, while the very few contributions on aerial platforms still pose important design constraints that limit their applicability. To overcome these limitations, in this paper we propose D-VAT, a novel end-to-end visual active tracking methodology based on deep reinforcement learning that is tailored to micro aerial vehicle platforms. The D-VAT agent computes the vehicle thrust and angular velocity commands needed to track the target by directly processing monocular camera measurements. We show that the proposed approach allows for precise and collision-free tracking operations, outperforming different state-of-the-art baselines on simulated environments which differ significantly from those encountered during training.

Abstract: It is a long-standing problem in robotics to develop agents capable of executing diverse manipulation tasks from visual observations in unstructured real-world environments. To achieve this goal, the robot needs to have a comprehensive understanding of the 3D structure and semantics of the scene. In this work, we present $\textbf{GNFactor}$, a visual behavior cloning agent for multi-task robotic manipulation with $\textbf{G}$eneralizable $\textbf{N}$eural feature $\textbf{F}$ields. GNFactor jointly optimizes a generalizable neural field (GNF) as a reconstruction module and a Perceiver Transformer as a decision-making module, leveraging a shared deep 3D voxel representation. To incorporate semantics in 3D, the reconstruction module utilizes a vision-language foundation model ($\textit{e.g.}$, Stable Diffusion) to distill rich semantic information into the deep 3D voxel. We evaluate GNFactor on 3 real robot tasks and perform detailed ablations on 10 RLBench tasks with a limited number of demonstrations. We observe a substantial improvement of GNFactor over current state-of-the-art methods in seen and unseen tasks, demonstrating the strong generalization ability of GNFactor. Our project website is https://yanjieze.com/GNFactor/ .

Abstract: Contact is at the core of robotic manipulation. At times, it is desired (e.g. manipulation and grasping), and at times, it is harmful (e.g. when avoiding obstacles). However, traditional path planning algorithms focus solely on collision-free paths, limiting their applicability in contact-rich tasks. To address this limitation, we propose the domain of Language-Conditioned Path Planning, where contact-awareness is incorporated into the path planning problem. As a first step in this domain, we propose Language-Conditioned Collision Functions (LACO) a novel approach that learns a collision function using only a single-view image, language prompt, and robot configuration. LACO predicts collisions between the robot and the environment, enabling flexible, conditional path planning without the need for manual object annotations, point cloud data, or ground-truth object meshes. In both simulation and the real world, we demonstrate that LACO can facilitate complex, nuanced path plans that allow for interaction with objects that are safe to collide, rather than prohibiting any collision.