Team Of Robots Research Articles

Emergence of Collective Behaviors for the Swarm Robotics Through Visual Attention-Based Selective Interaction

Emergence of Collective Behaviors for the Swarm Robotics Through Visual Attention-Based Selective Interaction

Robot learning on the job: Human-in-the-loop autonomy and learning during deployment

With the rapid growth of computing powers and recent advances in deep learning, we have witnessed impressive demonstrations of novel robot capabilities in research settings. Nonetheless, these learning systems exhibit brittle generalization and require excessive training data for practical tasks. To harness the capabilities of state-of-the-art robot learning models while embracing their imperfections, we present Sirius, a principled framework for humans and robots to collaborate through a division of work. In this framework, partially autonomous robots are tasked with handling a major portion of decision-making where they work reliably; meanwhile, human operators monitor the process and intervene in challenging situations. Such a human–robot team ensures safe deployments in complex tasks. Further, we introduce a new learning algorithm to improve the policy’s performance on the data collected from the task executions. The core idea is re-weighing training samples with approximated human trust and optimizing the policies with weighted behavioral cloning. We evaluate Sirius in simulation and on real hardware, showing that Sirius consistently outperforms baselines over a collection of contact-rich manipulation tasks, achieving an 8% boost in simulation and 27% on real hardware than the state-of-the-art methods in policy success rate, with twice faster convergence and 85% memory size reduction. Videos and more details are available at https://ut-austin-rpl.github.io/sirius/ .

Special Issue on Robotics and Mechatronics Technology for Aerial Robots

Aerial robot technologies, such as drones and unmanned aerial vehicles, are rapidly advancing in various domains. These robots now possess enhanced autonomous flying abilities and use sensor technology, Global Navigation Satellite System (GNSS), Light Detection and Ranging (LiDAR) system, and cameras to avoid obstacles, automatically land, and track flights. Evolving technologies in airspace management and route optimization ensure that flights are efficient and safe. Aerial robots have a wide range of applications, such as agriculture, disaster response, environmental monitoring, and inspection. These devices excel at precise data collection thanks to their high-resolution cameras and high-precision sensors. Swarm robotics, which allows collaboration between multiple aerial and ground robots, is also progressing. This enables quick coverage of large areas, which benefits disaster relief and land management. In addition, research and development of flying robots modeled after flying animals is underway. The evolution of aerial robots results in their efficient use in a variety of fields and new applications. This special issue includes the latest research papers and development reports on aerial robotic technology from the aforementioned perspectives. We hope that this special issue will pique researchers’ and engineers’ interest in aerial robots, thereby encouraging additional research and development. Finally, we would like to express our gratitude to all authors, all reviewers, the editorial board of the Journal of Robotics and Mechatronics, and Fuji Technology Press Ltd.

Body orientation change of neighbors leads to scale-free correlation in collective motion

Collective motion, such as milling, flocking, and collective turning, is a common and captivating phenomenon in nature, which arises in a group of many self-propelled individuals using local interaction mechanisms. Recently, vision-based mechanisms, which establish the relationship between visual inputs and motion decisions, have been applied to model and better understand the emergence of collective motion. However, previous studies often characterize the visual input as a transient Boolean-like sensory stream, which makes it challenging to capture the salient movements of neighbors. This further hinders the onset of the collective response in vision-based mechanisms and increases demands on visual sensing devices in robotic swarms. An explicit and context-related visual cue serving as the sensory input for decision-making in vision-based mechanisms is still lacking. Here, we hypothesize that body orientation change (BOC) is a significant visual cue characterizing the motion salience of neighbors, facilitating the emergence of the collective response. To test our hypothesis, we reveal the significant role of BOC during collective U-turn behaviors in fish schools by reconstructing scenes from the view of individual fish. We find that an individual with the larger BOC often takes on the leading role during U-turns. To further explore this empirical finding, we build a pairwise interaction mechanism on the basis of the BOC. Then, we conduct experiments of collective spin and collective turn with a real-time physics simulator to investigate the dynamics of information transfer in BOC-based interaction and further validate its effectiveness on 50 real miniature swarm robots. The experimental results show that BOC-based interaction not only facilitates the directional information transfer within the group but also leads to scale-free correlation within the swarm. Our study highlights the practicability of interaction governed by the neighbor’s body orientation change in swarm robotics and the effect of scale-free correlation in enhancing collective response.

Automatic Design of Robot Swarms under Concurrent Design Criteria: A Study Based on Iterated F‐Race

Automatic design is an appealing approach to realizing robot swarms. In this approach, a designer specifies a mission that the swarm must perform, and an optimization algorithm searches for the control software that enables the robots to perform the given mission. Traditionally, research in automatic design has focused on missions specified by a single design criterion, adopting methods based on single‐objective optimization algorithms. In this study, we investigate whether existing methods can be adapted to address missions specified by concurrent design criteria. We focus on the bi‐criteria case. We conduct experiments with a swarm of e‐puck robots that must perform sequences of two missions: each mission in the sequence is an independent design criterion that the automatic method must handle during the optimization process. We consider modular and neuroevolutionary methods that aggregate concurrent criteria via the weighted sum, hypervolume, or ‐norm. We compare their performance with that of Mandarina, an original automatic modular design method. Mandarina integrates Iterated F‐race as an optimization algorithm to conduct the design process without aggregating the design criteria. Results from realistic simulations and demonstrations with physical robots show that the best results are obtained with modular methods and when the design criteria are not aggregated.

Robo-Matter towards reconfigurable multifunctional smart materials

Maximizing materials utilization efficiency via enhancing their reconfigurability and multifunctionality offers a promising avenue in addressing the global challenges in sustainability. To this end, significant efforts have been made in developing reconfigurable multifunctional smart materials, which can exhibit remarkable behaviors such as morphing and self-healing. However, the difficulty in efficiently manipulating and controlling matter at the building block level with manageable cost and complexity, which is crucial to achieving superior responsiveness to environmental clues and stimuli, has significantly hindered the further development of such smart materials. Here we introduce a concept of Robo-Matter, which can be activated and controlled through external information exchange at the building block level, to enable a high-level of controllability, mutability and versatility for reconfigurable multifunctional smart materials. Using specially designed micro-robot building blocks with symmetry-breaking active motion modes, tunable anisotropic interactions, and interactive coupling with a programmable spatial-temporal dynamic light field, we demonstrate an emergent Robot-Matter duality, which enables a spectrum of desirable behaviors spanning from matter-like properties such as ultra-fast self-assembly and adaptivity, to robot-like properties including active force output, smart healing, smart morphing and infiltration. Our work demonstrates a promising direction for designing next-generation smart materials and large-scale robotic swarms.

Robust Optimization Models for Planning Drone Swarm Missions

This article presents methods of planning unmanned aerial vehicle (UAV) missions in which individual platforms work together during the reconnaissance of objects located within a terrain. The planning problem concerns determining the flight routes of a swarm, where each UAV has the ability to recognize an object using a specific type of sensor. The experiments described in this article were carried out for drone formation; one drone works as a swarm information hub and exchanges information with the ground control station (GCS). Numerical models for mission planning are presented, which take into account the important constraints, simplifying the description of the mission without too much risk of losing the platforms. Several types of objective functions were used to optimize swarm flight paths. The mission models are presented in the form of mixed integer linear programming problems (MILPs). The experiments were carried out on a terrain model built on the basis of graph and network theory. The method of building a network on which the route plan of a drone swarm is determined is precisely presented. Particular attention was paid to the description of ways to minimize the size of the network on which the swarm mission is planned. The presented methods for building a terrain model allow for solving the optimization problem using integer programming tasks.

Evaluating Swarm Robotics for Mining Environments: Insights into Model Performance and Application

The mining industry is experiencing a transformative shift with the integration of automation, particularly through autonomous haul truck systems, and further advancements are anticipated with the application of swarm robotics. This study evaluates the performance of four swarm robot models, namely baseline, ant, firefly, and honeybee, in optimizing key mining operations such as ore detection, extraction, and transportation. Simulations replicating real-world mining environments were conducted to assess improvements in operational efficiency, scalability, reliability, selectivity, and energy consumption. The results demonstrate that these models can significantly enhance the precision and productivity of mining activities, especially in complex and dynamic settings. A case study of the Pilbara iron ore mine in Australia is presented to illustrate the practical applicability of these models in an actual mining context. The study also highlights specific enhancements in each model, including role specialization in the ant model, advanced communication in the firefly model, and improved localization combined with hybrid control in the honeybee model. While the honeybee model showed superior performance in high-precision tasks, its reliability was limited under high-error conditions, and it faced a computational resources bottleneck in large-scale operations, highlighting the need for further development. By evaluating these models against performance criteria, the study identifies the most suitable swarm models for various mining conditions, offering insights into achieving more sustainable, scalable, and efficient mining operations.

The use of digital twin for mobile robot swarm task allocation

Cyber-physical systems, with the advent of the Internet of Things (IoT), have paved the way for the rise of the Industrial Internet of Things (IIoT) in industrial environments. A significant manifestation of this synergy is the Digital Twin—a virtual representation of industrial entities, including robots. This paper examines the fusion of digital twins in swarm robotics, emphasizing communication, and task allocation among robots. This study employs several mobile robots interconnected via a wireless network, each equipped with a Raspberry Pi and specific sensors. A centralized server, acting as both a Docker host and a Docker registry, facilitates task allocation. The results indicate effective communication among robots, with the digital twin playing an instrumental role in safety assessment and task allocation, as well as deploying Docker containers using Portainer and GitLab APIs. Unity3D aids in visualizing IMU motion and orientation. The motivation behind this study stems from the need to enhance efficiency and safety in industrial settings through advanced automation and coordination in swarm robotics. This research not only demonstrates the feasibility of integrating digital twins with IoT-enabled robots but also sets the stage for future investigations into more complex task allocation and safety protocols in industrial environments. By leveraging digital twins in swarm robotics, we open new pathways for research into more adaptive, responsive, and efficient industrial systems.

Multi-agent long-distance end-to-end indoor navigation: using imitation learning pre-training and global map

Abstract To address the navigation challenges of mobile robot swarm coordination in complex environments, this paper presents a distributed navigation approach for mobile robot swarm. The study initially formulates the problem of robot swarm navigation as a partially observable Markov decision process, employing a distributed proximal policy optimization algorithm to train decision models mapping observations to actions directly. Furthermore, it adopts imitation learning pre-training to maximize expert decisions in the initial training phase, thereby reducing exploration space. Lastly, an improved path point generation algorithm along with a spatiotemporal reachability calculation method is proposed to guide the Deep Reinforcement Learning policy in accomplishing long-distance indoor navigation. Comprehensive performance evaluations are conducted in simulated environments to compare the proposed method with existing approaches. Results demonstrate that the proposed approach effectively enhances both convergence speed and model performance, while significantly improving the navigation capabilities of robot swarms in complex environments.

Kinematics-guided data-driven energy surrogate model for robotic additive manufacturing

With the increase in the usage of industrial robots for manufacturing applications, there will be a corresponding rise in energy consumption. Especially, robotic additive manufacturing (AM) has demonstrated significant potential for working autonomously in hazardous or extraterrestrial settings, as well as for producing large-scale structures. Recent advancements have made it possible for teams of robots to collaborate for printing structures even larger than their sizes together. Utilizing robots efficiently for AM is essential for achieving energy sustainability goals. The literature on energy models for robots and AM processes lacks a comprehensive energy model for robotic AM. To address this gap, a multi-point trajectory-based energy model is proposed for robotic AM. The model is created to help users manufacture parts in an energy-efficient way. These simulations require a significant amount of computation time, which limits their usage to offline purposes only. Hence, we propose two energy surrogate models for real-time predictions: a pure data-driven model and a kinematics-guided data-driven model. The kinematics-guided approach is an enhanced version of pure data-driven model which learns from inverse kinematics solutions along the trajectory to understand the robot’s dynamics and kinematics. There were two test cases conducted. The first test case consisted of paths that were feasible, printable, and within the robot’s reach, while the second test case consisted of a mixture of paths that were both feasible and infeasible. In comparison to the pure data-driven approach, the kinematics-guided approach stood out in both testing cases. Two real-world case studies have been used to demonstrate that the proposed approach can be used in lieu of simulations for real-time applications. Especially in the second case, where the part was placed very close to the robot making a portion of the part infeasible to print, the proposed approach correctly identified that it was not possible to print the geometry due to the robot’s kinematics. The proposed kinematics-guided approach is 195 and 700 times faster than the simulation results, respectively, in two case studies, and thus supporting the use of the proposed approach for real-time applications.

Automated Discovery of Efficient Behavior Strategies for Distributed Shape Formation of Swarm Robots by Genetic Programming

Automated Discovery of Efficient Behavior Strategies for Distributed Shape Formation of Swarm Robots by Genetic Programming

Energy Efficiency Deep Reinforcement Learning for URLLC in 5G Mission-Critical Swarm Robotics

Energy Efficiency Deep Reinforcement Learning for URLLC in 5G Mission-Critical Swarm Robotics

Bridging the reality gap in drone swarm development through mixed reality

AbstractSwarm algorithms promise to solve certain problems in large multi-robot systems. The evaluation of large swarms is however challenging as simulations alone often lack some properties of real systems whereas real-world experiments are costly and complex. We present a mixed reality (MR) system that connects simulated and physical robots though a 5G network, facilitating MR experiments to evaluate communication-based swarm algorithms. The effectiveness of the system is demonstrated through extensive experiments with unmanned aerial vehicles. Measurements show that the communication requirements of swarm coordination are well met by 5G but the computing power of the simulation server can be a bottleneck. However, even when the simulation slows down, communication and coordination take place in real time. In conclusion, 5G-enabled MR experiments are a feasible tool for bridging the reality gap in the development and evaluation of robot swarms.

TIP: A trust inference and propagation model in multi-human multi-robot teams

Trust is a crucial factor for effective human–robot teaming. Existing literature on trust modeling predominantly focuses on dyadic human-autonomy teams where one human agent interacts with one robot. There is little, if not no, research on trust modeling in teams consisting of multiple human and robotic agents. To fill this important research gap, we present the Trust Inference and Propagation (TIP) model to model and estimate human trust in multi-human multi-robot teams. In a multi-human multi-robot team, we postulate that there exist two types of experiences that a human agent has with a robot: direct and indirect experiences. The TIP model presents a novel mathematical framework that explicitly accounts for both types of experiences. To evaluate the model, we conducted a human-subject experiment with 15 pairs of participants (N=30\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$N=30$$\\end{document}). Each pair performed a search and detection task with two drones. Results show that our TIP model successfully captured the underlying trust dynamics and significantly outperformed a baseline model. To the best of our knowledge, the TIP model is the first mathematical framework for computational trust modeling in multi-human multi-robot teams.

Explainable machine learning model of disorganisation in swarms of drones.

The main challenges when managing a fleet of unmanned aerial vehicles are to ensure the relative stability of its formation and to minimise disorganisation, specifically when undergoing an intrusion. When planning the mission it is beneficial for the operator to set the parameters of the formation to balance the needs of the mission with the disorganisation that an intruder may cause. The model developed in this research predicts the anticipated disturbance as a function of the parameters of the formation. The effectiveness of six machine learning methods are compared with a previously established baseline, using data obtained from simulations. CatBoost (categorical boosting) delivered the best results, with an (coefficient of determination) value of 83.3%, representing an improvement of 80% over the baseline. The SHAP (Shapley Additive Explanations) method was used to extend the model beyond predictability for particular combinations of values of parameters, towards generalised recommendations for the operator of the formation.

An improved whale optimization algorithm for UAV swarm trajectory planning

For the problem of trajectory planning in the small-scale unmanned aerial vehicle (UAV) swarms, classical intelligent algorithms and newly emerged bio-inspired algorithms often suffer from being trapped in local optima, leading to suboptimal solutions. Moreover, these algorithms fail to ensure optimal solutions and convergence speed in high-speed dynamic UAV networking scenarios. In response to these challenges, this paper proposes an Improved Whale Optimization Algorithm (IWOA) that considers the global perspective. To address the issue of initial solution diversity, an opposition-based learning method is introduced by constructing a reverse population, enhancing the diversity of the initial population. To overcome the problem of the algorithm getting trapped in local optima, random convergence factors, and end-point random neighborhood perturbations are incorporated to accelerate the global convergence speed of the IWOA and prevent it from getting stuck in local optima. Additionally, a grid digital elevation model is employed when constructing the experimental environment model, considering various physical constraints of the UAV swarm. This ensures that the simulation validation of the IWOA algorithm is closer to real-world scenarios. Simulation results indicate that, compared to commonly used algorithms, the IWOA can generate more optimal trajectories for drone swarm planning under constraints resembling real-world scenarios. It exhibits better performance in trajectory evaluation, convergence speed, and stability, thereby meeting the requirements of trajectory planning for small-scale UAV swarms. The proposed IWOA enhances UAV swarm coordination and efficiency, significantly impacting real-world applications in various fields.

Drone Swarm Classification from ISAR Imaging

In today's technology, the use of drones has become very popular for they can be easily purchased over the Internet and can be easily developed. With drones that have wide usage areas, swarm structures have become popular. However, this has brought about some problems. The issue of drone detection has emerged in order to prevent the uncontrolled use of drone swarms in the airspace. Drone swarm detection is important to prevent dangerous accidents or criminal acts. In this study, a new classification algorithm is proposed with deep learning using inverse synthetic aperture radar (ISAR) images of drone swarms based on various formation swarm types. ISAR images are created using ANSYS simulation. Additionally, high frequency structural simulator (HFSS) - shooting bouncing ray (SBR+) solver is used for high-speed computation. Radar and simulation parameters to obtain ISAR images are discussed. Especially, down-range and cross-range resolution parameters are taken into account to achieve high resolution. ISAR images are classified using deep learning methods in terms of formation. Formation types include Line, Square, Cross, and Triangle. The convolutional neural network (CNN) model is used to solve classification problems. The model consists of train, validation, and test steps. Classification performance results are presented with high accuracy. The developed method can be used for anti-drone technologies.

Проблема ценности в цифровом обществе роевого интеллекта

Introduction. The emerging digital society demonstrates the problem of justifying the value, worth and price of digital products and services, which requires consideration of the specificity of value in this society. Theoretical analysis. Following the financial and art market, the boundary between “value” and “worth” is blurred in the digital society. Value, including the value of an individual, becomes variable and is conditioned by the presence of unique attributes and unique experiences that obtain market value. At the same time, under standard digital protocols and technologies, the true uniqueness of an individual experience can only be achieved through the creation of customized digital ecosystems in the form of robotic swarms that support human life activities. A set of digital devices that have sufficient resources to continuously self-renew and form a closed environment around the individual that provides a unique experience can be called a chrysalic swarm. The chrysalic swarm creates an opportunity for the individual to escape from the logic of mutual competition dictated by society. Conclusion. Studying the value problematization in the digital society of swarm intelligence opens a theoretical perspective on the problem of the crisis of social elites in the context of the fulfillment of this scenario of society development.

Exploring advancements and emerging trends in robotic swarm coordination and control of swarm flying robots: A review

Swarm Robotics (SR) is an interdisciplinary field that is rapidly advancing to address complex industrial challenges. This paper provides a comprehensive review of recent advancements and emerging trends in SR, with a specific focus on the coordination and control of Swarm Flying Robots (SFRs). The motivation behind this review is to explore scalable and robust solutions for SFRs to enhance their performance and adaptability across various applications. Key objectives include examining the characteristics and essential behaviors of SR, analyzing the challenges and so lutions for implementing SR in Flying Robots (FRs), and highlighting current and future research directions. The review delves into critical areas such as multiple robot path planning, Swarm Intelligence (SI), combinatorial optimization, and formation flying using SFR. Special attention is given to coordination and control techniques, including formation control in GPS-denied environments, to underscore their significance in advancing SR. The paper also addresses ethical, privacy, and security considerations, emphasizing the importance of responsible practices in SR development. Major takeaways from this review include the identification of key technical challenges and potential solutions in SFR, the exploration of SI algorithms, and the future research directions necessary for fully realizing the potential of SR technologies. By offering detailed insights into state-of-the-art research and its industrial implications, this paper serves as a foundational guide for future studies in the dynamic and promising domain of swarm robotics.