Zoe Betta, Davide Corongiu, Carmine Tommaso Recchiuto
et al.
Infrastructure inspection is becoming increasingly relevant in the field of robotics due to its significant impact on ensuring workers' safety. The harbor environment presents various challenges in designing a robotic solution for inspection, given the complexity of daily operations. This work introduces an initial phase to identify critical areas within the port environment. Following this, a preliminary solution using a quadruped robot for inspecting these critical areas is analyzed.
We present pyastrobee: a simulation environment and control stack for Astrobee in Python, with an emphasis on cargo manipulation and transport tasks. We also demonstrate preliminary success from a sampling-based MPC controller, using reduced-order models of NASA's cargo transfer bag (CTB) to control a high-order deformable finite element model. Our code is open-source, fully documented, and available at https://danielpmorton.github.io/pyastrobee
This paper presents a comprehensive sim-to-real pipeline for autonomous strawberry picking from dense clusters using a Franka Panda robot. Our approach leverages a custom Mujoco simulation environment that integrates domain randomization techniques. In this environment, a deep reinforcement learning agent is trained using the dormant ratio minimization algorithm. The proposed pipeline bridges low-level control with high-level perception and decision making, demonstrating promising performance in both simulation and in a real laboratory environment, laying the groundwork for successful transfer to real-world autonomous fruit harvesting.
This work develops a distributed optimization algorithm for multi-robot 3-D semantic mapping using streaming range and visual observations and single-hop communication. Our approach relies on gradient-based optimization of the observation log-likelihood of each robot subject to a map consensus constraint to build a common multi-class map of the environment. This formulation leads to closed-form updates which resemble Bayes rule with one-hop prior averaging. To reduce the amount of information exchanged among the robots, we utilize an octree data structure that compresses the multi-class map distribution using adaptive-resolution.
We introduce an analytic method for generating a parametric and constraint-aware kick for humanoid robots. The kick is split into four phases with trajectories stemming from equations of motion with constant acceleration. To make the motion execution physically feasible, the kick duration alters the step frequency. The generated kicks seamlessly integrate within a ZMP-based gait, benefitting from the stability provided by the built-in controls. The whole approach has been evaluated in simulation and on a real NimbRo-OP2X humanoid robot.
Generative Pre-Trained Transformers (GPTs) are hyped to revolutionize robotics. Here we question their utility. GPTs for autonomous robotics demand enormous and costly compute, excessive training times and (often) offboard wireless control. We contrast GPT state of the art with how tiny insect brains have achieved robust autonomy with none of these constraints. We highlight lessons that can be learned from biology to enhance the utility of GPTs in robotics.
This paper revisits the topic of rotation estimation through the lens of special unitary matrices. We begin by reformulating Wahba's problem using $SU(2)$ to derive multiple solutions that yield linear constraints on corresponding quaternion parameters. We then explore applications of these constraints by formulating efficient methods for related problems. Finally, from this theoretical foundation, we propose two novel continuous representations for learning rotations in neural networks. Extensive experiments validate the effectiveness of the proposed methods.
The research on multi-robot coverage path planning (CPP) has been attracting more and more attention. In order to achieve efficient coverage, this paper proposes an improved DARP coverage algorithm. The improved DARP algorithm based on A* algorithm is used to assign tasks to robots and then combined with STC algorithm based on Up-First algorithm to achieve full coverage of the task area. Compared with the initial DARP algorithm, this algorithm has higher efficiency and higher coverage rate.
Coverage Path Planning involves visiting every unoccupied state in an environment with obstacles. In this paper, we explore this problem in environments which are initially unknown to the agent, for purposes of simulating the task of a vacuum cleaning robot. A survey of prior work reveals sparse effort in applying learning to solve this problem. In this paper, we explore modeling a Cover Path Planning problem using Deep Reinforcement Learning, and compare it with the performance of the built-in algorithm of the Roomba, a popular vacuum cleaning robot.
This invited review discusses causal learning in the context of robotic intelligence. The paper introduced the psychological findings on causal learning in human cognition, then it introduced the traditional statistical solutions on causal discovery and causal inference. The paper reviewed recent deep causal learning algorithms with a focus on their architectures and the benefits of using deep nets and discussed the gap between deep causal learning and the needs of robotic intelligence.
Trust is one of the hallmarks of human-human and human-robot interaction. Extensive evidence has shown that trust among humans requires reciprocity. Conversely, research in human-robot interaction (HRI) has mostly relied on a unidirectional view of trust that focuses on robots' reliability and performance. The current paper argues that reciprocity may also play a key role in the emergence of mutual trust and successful collaboration between humans and robots. We will gather and discuss works that reveal a reciprocal dimension in human-robot trust, paving the way to a bidirectional and dynamic view of trust in HRI.
Policy gradient is an effective way to estimate continuous action on the environment. This paper, it about explaining the mathematical formula and code implementation. In the end, comparing between the rotation angle of the stick on CartPole , and the angle of the Autonomous vehicle when turning, and utilizing the Bicycle Model, a simple Kinematic dynamic model, are the purpose to discover the similarity between these two models, so as to facilitate the model transfer from CartPole to the F1tenth Autonomous vehicle.
The ability to produce metallic membrane materials with porosity on the nanoscale from Ni‐based superalloys, hitherto used exclusively for high temperature applications, has been discovered 15 years ago. The basic principle is to first convert the initial γ/γ′ microstructure, containing isolated γ′‐precipitates, into a bi‐continuous network where both phases are in themselves continuous and interpenetrate each other. Then, one of the two phases is selectively removed, so that a rigid structure consisting of the remaining phase with pores on the location of the removed phase results. This article reviews the progress made so far. In that time period, a number of ways to fabricate these unique materials have emerged, utilizing 1) single crystals and polycrystals as precursor materials as well as 2) coarsening of coherent and incoherent γ′‐precipitates to realize bi‐continuity of the microstructure. Consequently, a family of superalloy membranes has emerged with specific microstructures, properties, advantages, and limitations. It is the intention of this article to give an overview on these various manufacturing routes, as well as on resulting microstructures and properties. Finally, possible fields of applications are outlined. It is demonstrated that the particular manufacturing process from a solid to the porous material leads to certain advantages, such as the ability to structure the material in porous and solid areas as required by the application.
The paper presents a novel technique for the design of optimal calibration experiments for a planar anthropomorphic manipulator with n degrees of freedom. Proposed approach for selection of manipulator configurations allows essentially improving calibration accuracy and reducing parameter identification errors. The results are illustrated by application examples that deal with typical anthropomorphic manipulators.
In modern life the road safety has becomes the core issue. One single move of a driver can cause horrifying accident. The main goal of intelligent car system is to make communication with other cars on the road. The system is able to control to speed, direction and the distance between the cars the intelligent car system is able to recognize traffic light and is able to take decision according to it. This paper presents a framework of the intelligent car system. I validate several aspect of our system using simulation.
The design of a novel prismatic drive is reported in this paper. This transmission is based on Slide-o-Cam, a cam mechanism with multiple rollers mounted on a common translating follower. The design of Slide-o-Cam was reported elsewhere. This drive thus provides pure-rolling motion, thereby reducing the friction of rack-and-pinions and linear drives. Such properties can be used to design new transmissions for parallel-kinematics machines. In this paper, this transmission is intended to replace the ball-screws in Orthoglide, a three-dof parallel robot intended for machining applications.
This article synthezises the most important results on the kinematics of cuspidal manipulators i.e. nonredundant manipulators that can change posture without meeting a singularity. The characteristic surfaces, the uniqueness domains and the regions of feasible paths in the workspace are defined. Then, several sufficient geometric conditions for a manipulator to be noncuspidal are enumerated and a general necessary and sufficient condition for a manipulator to be cuspidal is provided. An explicit DH-parameter-based condition for an orthogonal manipulator to be cuspidal is derived. The full classification of 3R orthogonal manipulators is provided and all types of cuspidal and noncuspidal orthogonal manipulators are enumerated. Finally, some facts about cuspidal and noncuspidal 6R manipulators are reported.
The goal of this paper is to define the n-connected regions in the Cartesian workspace of fully-parallel manipulators, i.e. the maximal regions where it is possible to execute point-to-point motions. The manipulators considered in this study may have multiple direct and inverse kinematic solutions. The N-connected regions are characterized by projection, onto the Cartesian workspace, of the connected components of the reachable configuration space defined in the Cartesian product of the Cartesian space by the joint space. Generalized octree models are used for the construction of all spaces. This study is illustrated with a simple planar fully-parallel manipulator.
The design of a spherical wrist with parallel architecture is the object of this article. This study is part of a larger project, which aims to design and to build an eel robot for inspection of immersed piping. The kinematic analysis of the mechanism is presented first to characterize the singular configurations as well as the isotropic configurations. We add the design constraints related to the application, such as (i) the compactness of the mechanism, (ii) the symmetry of the elements in order to ensure static and dynamic balance and (iii) the possibility of the mechanism to fill the elliptic form of the ell sections.
Alborz Geramifard, Peyman Nayeri, Reza Zamani-Nasab
et al.
This paper presents a new architecture called FAIS for imple- menting intelligent agents cooperating in a special Multi Agent environ- ment, namely the RoboCup Rescue Simulation System. This is a layered architecture which is customized for solving fire extinguishing problem. Structural decision making algorithms are combined with heuristic ones in this model, so it's a hybrid architecture.