Large-scale Robot Research Articles

Research on coupling dynamics modeling and composite control of the multi-flexible space robot

Space robots can assist astronauts in accomplishing a variety of space missions in high-risk environments. As spacecraft become increasingly lightweight, large-scale, and integrated, space robots are a type of large-scale robots with flexible links and flexible joints. However, the flexibility of the components complicates robot modeling and controller design. Therefore, to address the issue of stable trajectory tracking for multi-flexible space robots, a dynamics model of space robots with multiple flexibilities is established, and a composite controller is designed to enhance control stability. In modeling, comprehensively considering the rigid joints, joint flexibility and link flexibility, the kinematics is consistently described by using the component deformation matrix and recursive method. An accurate dynamics model is established by Lagrange’s equations. In control, using fuzzy neural network sliding mode control, a rigid subsystem controller is designed to suppress the system jitter caused by sudden parameter changes. And it implements the robust tracking of the robot trajectory. Meanwhile, the weighted average fuzzy controller is designed to control the link deformation. The regulation of the flexible joint by the fast feedback controller is implemented using the virtual torque feedback control. Finally, the validity of the dynamics and control models is verified using Matlab and Adams. Compared with previous studies, this study considers multiple flexibility factors more fully. The vibration of multiple flexible components is suppressed during the stable trajectory tracking of the robot.

Deep Reinforcement Learning-Based Large-Scale Robot Exploration

Deep Reinforcement Learning-Based Large-Scale Robot Exploration

Dexterous manipulation of construction tools using anthropomorphic robotic hand

Emerging studies are utilizing reinforcement learning (RL) and imitation learning (IL) to control large-scale robots in heavy construction tasks. There is limited attention given to the automation of delicate tasks typically performed manually. This paper proposes the control of an anthropomorphic robotic hand with a high degree of freedom for the manipulation of construction tools in a learning-based approach. For controlling the robotic hand, a simulation-based policy learning framework based on pretraining policies through IL is presented, subsequently fine-tuning them for construction tool manipulation using RL. In experimental trials, six policies are trained for the robotic hand to grasp six different construction tools. The results indicate that each of the learned policies enables the robotic hand to manipulate the corresponding tool with an almost 100% success rate, demonstrating resilience when confronted with tools of different scales. Additionally, the paper showcases the potential for scaling up the fundamental policy for downstream applications.

Smooth Path Planning for Redundant Large-Scale Robots using Measured Reference Points

Assistance systems generating paths for large-scale robots automatically ensure comfort for the operator and save time compared to operation by hand levers. For an industrial use case, a previously saved path with multiple measurement points needs to be replayed on demand in a time efficient fashion. For this purpose, an existing algorithm for smooth path planning is extended by modifications to allow its application to redundant large-scale robots. Due to low computational effort, it is possible to implement the proposed algorithm on the control unit of the system online. The resulting path fulfills the requirement of being continuously differentiable and all actuator limits are respected. A numerical simulation shows the effectiveness of the proposed path planning algorithm. The smooth path of the tool center point position in world coordinates is still very close to the measurements to prevent collisions, but brings the opportunity to save time while reaching for the same target position.

Integrated Task Allocation and Path Coordination for Large-Scale Robot Networks With Uncertainties

Artificial intelligence-enhanced autonomous unmanned systems, such as large-scale autonomous robot networks, are widely used in logistic and industrial applications. In this article, we address the integrated task assignment, path planning, and coordination problem applied for large-scale robot networks with the existence of uncertainties. In particular, a novel generalized conflict graph is designed which encodes the traveling time cost of the subsequent path planning result of each task-robot assignment and also includes the predicted path conflicts of each two assignments. An integrated optimization problem which aims to minimize the total traveling cost and potential path conflicts simultaneously is first formulated and then transformed into a linear programming instance to obtain the optimal solution. In particular, to satisfy the real-time requirement in large-scale systems, a greedy solution is presented which has the near-optimal performance but can decrease the computational complexity by orders of magnitude. The optimality, scalability, robustness, and efficiency of our approach are demonstrated by comprehensive comparisons with existing state-of-the-art approaches. Note to Practitioners—With the development of artificial intelligence techniques, large-scale autonomous robot networks are increasingly used in the logistic warehouses, unmanned container terminals, and intelligence transportation systems. This article considers the large-scale networks with hundreds or even thousands of unmanned robots which are implemented in lifelong transportation systems with uncertainties existed in practical execution process. Our main concept is to simultaneously minimize the total time cost of all the tasks and the potential motion conflicts among all the robots in the subsequent execution stage, thus alleviating robot congestions, balancing traffic distributions, increasing system efficiency, and improving the robustness and scalability. Lifelong simulations with thousand robots illustrate that our approach can reduce more than 30% of the time steps consumed for coordinating robot motion conflicts, and in the meantime, the throughput, and overall system efficiency are improved. However, simulation results show that the system improvement decreases in the presence of extreme high uncertainties (such as temporary motion and communication failures of the robot), due to the inaccurate conflict prediction in the integrated optimization stage. Our future work includes the deep-learning-based traffic evolution prediction and the online reallocation and planning in highly dynamic scenarios.

Equivalent-Input-Disturbance-Based Dynamic Tracking Control for Soft Robots via Reduced-Order Finite-Element Models

This article develops a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">systematic</i> framework for dynamic tracking control of soft robots. To this end, we propose a new projector for the proper orthogonal decomposition algorithm to significantly reduce the large-scale robot models, obtained from finite-element methods (FEM), while preserving their structure and stability properties. Such a property preservation enables an effective equivalent-input-disturbance-based scheme for dynamic tracking control of elastic soft robots with various geometries and materials. The proposed control scheme is composed of three key components, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> , feedforward control, disturbance-estimator control, and feedback control. To account for the trajectory reference, the feedforward action is designed from the dynamic FEM reduced-order robot model. The disturbance-estimator control action is obtained from an unknown input observer, which also provides the estimates of the reduced states for the feedback control design. The feedback gains of the observer-based controller are computed from an optimization problem under linear matrix inequality constraints. The closed-loop tracking properties are guaranteed using the Lyapunov stability theory. The effectiveness of the proposed dynamic control framework has been demonstrated via both high-fidelity SOFA simulations and experimental validations, performed on two soft robots with different natures. In particular, comparative studies with state-of-the-art control methods have been also carried out to highlight the interests of the new soft robot control results. This article is complemented with a video: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://bit.ly/2VVwtLn</uri> .

Distributed Fleet Management in Noisy Environments via Model-Predictive Control

We consider dynamic route planning for a fleet of Autonomous Mobile Robots (AMRs) doing fetch and carry tasks on a shared factory floor. In this paper, we propose Stochastic Work Graphs (SWG) as a formalism for capturing the semantics of such distributed and uncertain planning problems. We encode SWGs in the form of a Euclidean Markov Decision Process (EMDP) in the tool Uppaal Stratego, which employs Q-Learning to synthesize near-optimal plans. Furthermore, we deploy the tool in an online and distributed fashion to facilitate scalable, rapid replanning. While executing their current plan, each AMR generates a new plan incorporating updated information about the other AMRs positions and plans. We propose a two-layer Model Predictive Controller-structure (waypoint and station planning), each individually solved by the Q-learning-based solver. We demonstrate our approach using ARGoS3 large-scale robot simulation, where we simulate the AMR movement and observe an up to 27.5% improvement in makespan over a greedy approach to planning. To do so, we have implemented the full software stack, translating observations into SWGs and solving those with our proposed method. In addition, we construct a benchmark platform for comparing planning techniques on a reasonably realistic physical simulation and provide this under the MIT open-source license.

Visual Interaction Force Estimation Based on Time-Sensitive Dual-Resolution Learning Network

Haptic force feedback is an important perception method for humans to understand the surrounding environment. It can estimate tactile force in real time and provide appropriate feedback. It has important research value for robot-assisted minimally invasive surgery, interactive tactile robots, and other application fields. However, most of the existing noncontact visual power estimation methods are implemented using traditional machine learning or 2D/3D CNN combined with LSTM. Such methods are difficult to fully extract the contextual spatiotemporal interaction semantic information of consecutive multiple frames of images, and their performance is limited. To this end, this paper proposes a time-sensitive dual-resolution learning network-based force estimation model to achieve accurate noncontact visual force prediction. First, we perform continuous frame normalization processing on the robot running the video captured by the camera and use the hybrid data augmentation to improve the data diversity; secondly, a deep semantic interaction model is constructed based on the time-sensitive dual-resolution learning network, which is used to automatically extract the deep spatiotemporal semantic interaction information of continuous multiframe images; finally, we construct a simplified prediction model to realize the efficient estimation of interaction force. The results based on the large-scale robot hand interaction dataset show that our method can estimate the interaction force of the robot hand more accurately and faster. The average prediction MSE reaches 0.0009 N, R 2 reaches 0.9833, and the average inference time for a single image is 6.5532 ms; in addition, our method has good prediction generalization performance under different environments and parameter settings.

A flexible scaling self-healing method for morphology of swarm robots

The paper studies the formed body after completing its autonomous formation task and then proposes a self-healing method that flexibly scales with the scale of the loss in the situation that a large-scale swarm robot is locally missing due to external disturbances. With the stratified mechanism that includes layer-by-layer separation, layer-by-layer filling and iteration loop, the macro-level behavior of a swarm robot is transformed to the local action of an individual robot within the edge layers of the current aggregate. Based on the existing formation technique, driven by the shape genes of the formed body and the changes in the scale of population, the self-healing goal of the large scale of population is transformed into the similar scale of the formed body after scaling. Through the self-healing movement control method centered on individual behavior rules, the bottom-up self-healing configuration is presented on a scaled scale. The paper verifies the feasibility and scalability of this novel method through simulations. The maximum number of robots can reach tens of thousands. Finally, the self-healing method is used to implement the Cilibot robot, a hardware system developed by our laboratory.

Improving Software Robot Maintenance in Large-Scale Environments–is Center of Excellence a Solution?

Robotic Process Automation (RPA) performs high-volume tasks such as checking invoices. However, the governance and maintenance of large-scale software robot environments can be challenging when robot servers perform automatized tasks simultaneously for customer organizations with complex programming rules, dedicated parameters, and dependencies on timetables. A multivocal literature review (MLR) was conducted to explore whether there are 1) mechanisms to improve software robot maintenance in large-scale robot environments, 2) or software robot maintenance practices for scalable RPA in organizations providing shared services, 3) or governance models for optimizing the performance of software robot maintenance, and 4) is the Center of Excellence (CoE) one of the success factors concerning large-scale robot environments. By doing this, 5) we found eleven functional requirements for the monitoring tool to support maintenance in a large-scale environment. In addition, we adapted them to the RPA monitoring tool abilities for the Finnish Government Shared Services Centre for Finance and HR (Palkeet). As a result, the eleven functional requirements and the monitoring tool abilities are adaptable for other large-scale environments to improve software robot maintenance. However, commercial monitoring tools for RPA maintenance do not fulfil functional requirements, and organizations in large-scale environments must develop their monitoring tools. Based on MLR, either in-house or outsourced CoE seems to be one of the success factors in RPA maintenance in large-scale environments.

3D-Printed Miniature Robots with Piezoelectric Actuation for Locomotion and Steering Maneuverability Applications

The miniaturization of robots with locomotion abilities is a challenge of significant technological impact in many applications where large-scale robots have physical or cost restrictions. Access to hostile environments, improving microfabrication processes, or advanced instrumentation are examples of their potential use. Here, we propose a miniature 20 mm long sub-gram robot with piezoelectric actuation whose direction of motion can be controlled. A differential drive approach was implemented in an H-shaped 3D-printed motor platform featuring two plate resonators linked at their center, with built-in legs. The locomotion was driven by the generation of standing waves on each plate by means of piezoelectric patches excited with burst signals. The control of the motion trajectory of the robot, either translation or rotation, was attained by adjusting the parameters of the actuation signals such as the applied voltage, the number of applied cycles, or the driving frequency. The robot demonstrated locomotion in bidirectional straight paths as long as 65 mm at 2 mm/s speed with a voltage amplitude of only 10 V, and forward and backward precise steps as low as 1 µm. The spinning of the robot could be controlled with turns as low as 0.013 deg. and angular speeds as high as 3 deg./s under the same conditions. The proposed device was able to describe complex trajectories of more than 160 mm, while carrying 70 times its own weight.

Bioinspired soft microrobots actuated by magnetic field

In contrast to traditional large-scale robots, which require complicated mechanical joints and material rigidity, microrobots made of soft materials have exhibited amazing features and great potential for extensive applications, such as minimally invasive surgery. However, microrobots are faced with energy supply and control issues due to the miniaturization. Magnetic field actuation emerges as an appropriate approach to tackle with these issues. This review summarizes the latest progress of biomimetic soft microrobots actuated by magnetic field. Starting with an overview of the soft material and magnetic material adopted in the magnetic field actuated soft microrobots, the various fabrication methods and design structures of soft microrobots are summarized. Subsequently, practical and potential applications, such as targeted therapy, surgical operation, and the transportation of microscopic objects, in the fields of biomedicine and environmental remediation are presented. In the end, some current challenges, and the future development trends of magnetic soft microrobots are briefly discussed. This review is expected to offer a helpful guidance for the new researchers of biomimetic soft microrobots actuated by magnetic field.

Robotification & ethical cleansing

Robotics is currently not only a cutting-edge research area, but is potentially disruptive to all domains of our lives—for better and worse. While legislation is struggling to keep pace with the development of these new artifacts, our intellectual limitations and physical laws seem to present the only hard demarcation lines, when it comes to state-of-the-art R&D. To better understand the possible implications, the paper at hand critically investigates underlying processes and structures of robotics in the context of Heidegger’s and Nishitani’s accounts of science and technology. Furthermore, the analysis draws on Bauman’s theory of modernity in an attempt to assess the potential risk of large-scale robot integration. The paper will highlight undergirding mechanisms and severe challenges imposed upon our socio-cultural lifeworlds by massive robotic integration. Admittedly, presenting a mainly melancholic account, it will, however, also explore the possibility of robotics forcing us to reassess our position and to solve problems, which we seem unable to tackle without facing existential crises.

Lightweight 3-D Localization and Mapping for Solid-State LiDAR

The LIght Detection And Ranging (LiDAR) sensor has become one of the most important perceptual devices due to its important role in simultaneous localization and mapping (SLAM). Existing SLAM methods are mainly developed for mechanical LiDAR sensors, which are often adopted by large scale robots. Recently, the solid-state LiDAR is introduced and becomes popular since it provides a cost-effective and lightweight solution for small scale robots. Compared to mechanical LiDAR, solid-state LiDAR sensors have higher update frequency and angular resolution, but also have smaller field of view (FoV), which is very challenging for existing LiDAR SLAM algorithms. Therefore, it is necessary to have a more robust and computationally efficient SLAM method for this new sensing device. To this end, we propose a new SLAM framework for solid-state LiDAR sensors, which involves feature extraction, odometry estimation, and probability map building. The proposed method is evaluated on a warehouse robot and a hand-held device. In the experiments, we demonstrate both the accuracy and efficiency of our method using an Intel L515 solid-state LiDAR. The results show that our method is able to provide precise localization and high quality mapping. We made the source codes public at https://github.com/wh200720041/SSL_SLAM.

PRIMAL$_2$: Pathfinding Via Reinforcement and Imitation Multi-Agent Learning - Lifelong

Multi-agent path finding (MAPF) is an indispensable component of large-scale robot deployments in numerous domains ranging from airport management to warehouse automation. In particular, this work addresses lifelong MAPF (LMAPF) - an online variant of the problem where agents are immediately assigned a new goal upon reaching their current one - in dense and highly structured environments, typical of real-world warehouse operations. Effectively solving LMAPF in such environments requires expensive coordination between agents as well as frequent replanning abilities, a daunting task for existing coupled and decoupled approaches alike. With the purpose of achieving considerable agent coordination without any compromise on reactivity and scalability, we introduce PRIMAL2, a distributed reinforcement learning framework for LMAPF where agents learn fully decentralized policies to reactively plan paths online in a partially observable world. We extend our previous work, which was effective in low-density sparsely occupied worlds, to highly structured and constrained worlds by identifying behaviors and conventions which improve implicit agent coordination, and enable their learning through the construction of a novel local agent observation and various training aids. We present extensive results of PRIMAL2 in both MAPF and LMAPF environments and compare its performance to state-of-the-art planners in terms of makespan and throughput. We show that PRIMAL2 significantly surpasses our previous work and performs comparably to these baselines, while allowing real-time re-planning and scaling up to 2048 agents.

Flexible multibody system modelling of an aerial rescue ladder using Lagrange’s equations

ABSTRACT Due to limited weight and stiffness, large-scale robots are susceptible to structural oscillations during operation. In this publication, a model for the vertical dynamics of an aerial rescue ladder as an application example for large scale flexible robots is derived based on the Lagrange formalism. The model consists of five flexible segments, each using an arbitrary number of flexible modes. Kinematic loops occurring due to the type of interconnection between the telescopic elements are implicitly solved by the chosen assumed modes. Linearized system matrices are extracted from the model directly and very efficiently. In the resulting modelling process, the manipulator is solely described by its position and orientation kinematics and thus an adaption to different kinds of manipulators is straightforward. A validation against real world measurement data confirms the high accuracy of the derived model.

Prediction, Planning, and Coordination of Thousand-Warehousing-Robot Networks With Motion and Communication Uncertainties

In this article, we focus on resolving the traffic flow prediction, robot path planning, and motion coordination problems in large-scale warehousing robotics systems with thousand-robot networks. The warehousing environment is partitioned into several sectors, and a hierarchical framework is developed, which includes a centralized prediction and planning level and a decentralized local coordination level. In the centralized level, a traffic flow prediction algorithm is first proposed to predict the evolution of the robot density distribution in a future horizon and estimate the future traffic heat value of each sector. Based on this, the sector-level robot path can be generated in the time-expended sector graph by comprehensively considering the traveling distance and the predicted traffic heat value and will be dynamically updated by considering the most recent traffic information. In the coordination level, local cooperative A* algorithm, incorporated with the conflict-based searching strategy, is implemented within each sector to generate conflict-free road-level paths for all the robots in the sector simultaneously, and the rolling planning scheme is utilized in order to immediately react to robot motion uncertainties and communication disconnections. The effectiveness and practical applicability of the proposed approach are validated by large-scale simulations with more than one 1000 robots and real laboratory experiments. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —Considering practical situations and requirements in industrial warehouses and automated logistics systems, this article resolves the life-long planning and coordination problems of large-scale robot networks and ensures the practical execution performance in the presence of robot motion uncertainties and temporary communication disconnections. Our main idea is to reduce robot congestions and improve warehouse working efficiency by balancing the traffic flow in the whole environment. To achieve this, we present a traffic flow prediction algorithm to estimate the robot density distribution in a future horizon and take this information into consideration in sector-level path planning. The reliability, scalability, and the real-time performance of the proposed solution are achieved by the presented hierarchical system framework and the dynamic planning scheme. The proposed concept and approach can also be used to coordinate other large-scale systems with multirobot or multi-AGV networks. Simulation and experimental results suggest that the proposed solution is effective and practically applicable, but a saturation phenomenon of the system capacity can be observed under a very heavy workload. In the future, we will investigate the relation between the maximum system capacity and the environment structure and make further efforts to optimize the environment structure and road layout in order to improve the warehouse working efficiency.

Navigation in Large Groups of Robots

As robots become more ubiquitous, research in large-scale robot navigation has gained more focus given their potential real-world applications. This review intends to provide a summary of recent advances on the field of multi-robot navigation, focusing on cases where scale is an important attribute of the system. Experimental evaluation of a large number of robots is complex due to size and cost constraints. Successful applications of large-scale robot navigation approaches are in structured and known environments where robots are centrally controlled. Advances in the standardization of robot hardware and software are increasing interest for research in navigation for large groups of robots. We present a review on navigation approaches for multi-robot systems. Then, we focus on recent articles that deal with the problem of moving a large number of robots in virtual or physical space. Finally, we summarize our main findings and emphasize the challenges.

A Multi-Class Neural Network Model for Rapid Detection of IoT Botnet Attacks

The tremendous number of Internet of Things (IoT) devices and their widespread use have made our lives considerably more manageable. At the same time, however, , the vulnerability of these innovations means that our day-to-day existence is surrounded by insecure devices, thereby facilitating ways for cybercriminals to launch various attacks by large-scale robot networks (botnets) through IoT. This problem is further heightened by the constraints of the IoT on security techniques due to limited resources including central processing units (CPUs), memory, and power consumption. In consideration of these issues, we propose a lightweight neural network-based model to rapidly detect IoT botnet attacks. The model was developed using FastGRNN algorithm which is a lightweight and fast version of the recurrent neural network. In addition, it is independent and does not require any specific equipment or software to fetch the required features for learning and detection processes. Therefore, only packet headers are required to complete learning and detection. Furthermore, the model provides multi-classification, which is necessary for taking appropriate countermeasures to understand and stop the attacks. According to the conducted experiments, the proposed model is accurate and achieves 99.99%, 99.04% as F1 score for MedBIoT and Mirai-RGU datasets in addition, to fulfilling IoT constraints regarding complexity and speed. It is less complicated in terms of computations, and it provides real-time detection that outperformed the state-of-the-art, achieving a detection time ratio of 1:5 for MedBIoT dataset and a ratio of 1:8 for Mirai-RGU dataset.

Evaluation of a Large Scale Event Driven Robot Skin

This letter evaluates and describes the large-scale integration of our multi-modal event-driven robot skin system on our humanoid robot H1 (REEM-C, PAL robotics). The robot skin is powered by the robot and all processing of tactile perception and control are executed onboard the robot. The robot skin system largely covers the humanoid and employs 1260 skin cells in 47 skin patches, in total 7560 multi-modal tactile sensors. The robot skin system is driven by events, i.e., the occurrence of novel information drives the acquisition, transmission, and processing of information rather than the synchronous sampling clock as in clock-driven systems. The event-driven robot skin system enables efficient multi-modal large area tactile perception, which is a prerequisite for realizing reactive whole-body control. We analyze the new robot skin system and evaluate its performance in clock-driven mode and event-driven mode. This analysis includes modeling the required CPU load and investigating the scaling of the system in multi-core systems. We evaluate the robot skin system with an experiment, where a large number of skin cells is activated (>680 cells), and a large number of events is generated. The obtained results demonstrate the superior performance of the event-driven system. In the clock-driven mode, the robot skin system constantly produces 315 000 packets/s, while in event-driven mode the system at most produces 40 000 packets/s (13%). The CPU load reduces from constantly 270% to at most 100% (37%). In clock-driven mode, the PC drops on average 80 000 packets/s (25% of all packets), while in event-driven mode, the package loss is practically neglectable. This efficient large-scale robot skin enables the complete onboard integration into a humanoid robot without the need for additional external power or processing capabilities.