This paper explores Particle Swarm Optimization (PSO) based control of a morphing robot arm mounted on a Controlled Floating Space Robot (CFSR). CFSR is a new study spacecraft model allowing controlled adjustments of ...
详细信息
ISBN:
(数字)9798350307566
ISBN:
(纸本)9798350307573
This paper explores Particle Swarm Optimization (PSO) based control of a morphing robot arm mounted on a Controlled Floating Space Robot (CFSR). CFSR is a new study spacecraft model allowing controlled adjustments of the satellite base's pose. The dynamic model of the chaser spacecraft is derived relative to the target in an inertial frame, using the Clohessy- Wiltshire model or Hills equations for defining their relative motion. It is assumed that both spacecraft are in an Inplane-elliptical formation. The manipulator dynamics are obtained through Lagrangian formulation. The PSO algorithm optimizes controller gain values while a quintic polynomial tra-jectory is provided as input to the arm joints. Results show that the PSO-based controller effectively reduces tracking error and power requirements, and it is also compared with a conventional proportional-derivative (PD) controller.
In the modern age of industry 4.0 and state-of-the-art manufacturing, automated inspection plays an essential role. The bearing in the rotary machine is a vital and critical component as it provides support and stabil...
详细信息
ISBN:
(数字)9798350354218
ISBN:
(纸本)9798350354225
In the modern age of industry 4.0 and state-of-the-art manufacturing, automated inspection plays an essential role. The bearing in the rotary machine is a vital and critical component as it provides support and stability to the rotary elements of machines. The defect in the bearing is responsible for the increase in vibrations, noise and reduction in reliability and safety of the rotating machines. It is very challenging to humanly detect different faults in these bearings. This work highlights the methods of multi-class bearing fault classification using Deep Learning (DL) and Transfer Learning (TL) models. The different DL models such as Artificial Neural Network (ANN), and Convolutional Neural Network (CNN) are used along with the TL models such as VGG16 and MobileNet. The models perform with good classification accuracy ranging from 88% to 98%. The Transfer learning models perform better for both binary and multi-class fault classification of the bearings.
Reinforcement learning (RL) has recently enjoyed significant success in games, robotics, bioinformatics, etc. Soon, it will not be uncommon to see AI models employing RL agents integrated with various hardware and sof...
详细信息
ISBN:
(数字)9783030955021
ISBN:
(纸本)9783030955014
Reinforcement learning (RL) has recently enjoyed significant success in games, robotics, bioinformatics, etc. Soon, it will not be uncommon to see AI models employing RL agents integrated with various hardware and software solutions. Due to its generality and robustness, RL is applied in several disciplines such as game theory, control theory, multi-agent systems, swarm intelligence, robotics, and NLP. Despite these advances and successes, reinforcement learning faces many challenges for real-world adoption. Some of the major difficulties being, operator’s trust and ability of an agent to explain the actions taken in a human-understandable manner. Traditionally the AI systems are black-box models. With the advent of various legal regulations worldwide, notably the European General Data Protection Regulation (GDPR) [29], it has started becoming mandatory that the AI models be transparent, interpretable, and secure. If an RL agent can effectively and accurately explain the actions carried out by the RL system to the observers/operators, it will be a tangible step towards developing the ART (accountable, reliable and trustworthy) RL agent. This can effectively facilitate the adoption of RL systems in real-world domains. Various explainable AI (XAI) methods have been reported in the literature. However, there is a considerable lacuna in the availability of Explainable RL (XRL) methods. This paper introduces a novel RL algorithm agnostic approach of generating human-understandable explanations using the probabilistic graphical model. This method is based on Probabilistic Graphical models (PGM) [36]. It is algorithm agnostic in that it is not dependent on any specific RL method and can be integrated with any RL algorithm. We also introduce a PGM model, which is learned along with an agent’s training via classic methods and used for generating explanations at run time. Specific case studies are considered, and results are presented which demonstrate our approach. Our exper
As the demand for online retail and food delivery services continues to grow, it becomes increasingly important to find efficient last-mile delivery solutions that can support the sustainability of urban areas. Autono...
详细信息
ISBN:
(数字)9798350361230
ISBN:
(纸本)9798350361247
As the demand for online retail and food delivery services continues to grow, it becomes increasingly important to find efficient last-mile delivery solutions that can support the sustainability of urban areas. Autonomous Delivery Robots (ADRs) could potentially offer a solution to these challenges. However, long-term real-world studies are still required. This work presents a concept for a flexible ADR-based logistics system for urban areas, with a focus on the design of a robust ADR with detachable transport boxes and accompanying infrastructure. Here, we draw on well-established methods from autonomous mobile robotics in intralogistics. We present a two-mode delivery concept that removes the need for the recipient to be present at the time of delivery. This concept is currently being tested in the ongoing efeuCampus living lab in Bruchsal, Germany. We provide first findings and recommendations for further research in this area.
Permanent magnet synchronous machines (PMSMs) are of great interest in automation and robotics applications due to their small size, high efficiency, and low maintenance requirements. However, these motors face challe...
详细信息
ISBN:
(数字)9798331542726
ISBN:
(纸本)9798331542733
Permanent magnet synchronous machines (PMSMs) are of great interest in automation and robotics applications due to their small size, high efficiency, and low maintenance requirements. However, these motors face challenges in controlling their speed and position, such as precise rotor detection. This situation is essential in many Industrial applications. Conventional controllers such as PID controllers are commonly used, but their performance is often affected by changes in system dynamics, necessitating updating of controller parameters based on the accurate mathematical model of the motor. This research aims to investigate several intelligent control strategies that do not rely on the mathematical model of the motor. The performance of the drive system will be evaluated using three intelligent controllers: a neural network controller, a neural fuzzy inference system (ANFIS) controller, and a deep reinforcement learning (DRL) controller. The performance of these smart controllers will be compared with that of the PID controller. The results presented in this paper demonstrate the ability of the proposed smart controllers to regulate the speed of PMSMs compared to a PID controller. Specifically, the ANFIS-based controller shows the best performance in terms of maximum overshoot, while the DRL controller demonstrates the best settling time.
Instance segmentation is a long-standing task for supporting robotic bin picking. However, objects of diverse classes can be closely packed with occlusions in cluttered and chaotic scenes, hence, even recent methods c...
Instance segmentation is a long-standing task for supporting robotic bin picking. However, objects of diverse classes can be closely packed with occlusions in cluttered and chaotic scenes, hence, even recent methods could have difficulty in locating clear and precise boundaries to distinguish nearby objects. In this work, we aim to improve the boundary quality of the instance masks for robust and precise instance segmentation in these challenging scenarios. Technical-wise, we first formulate an IoU-based Boundary-aware Mask head (IBM head) for predicting the instance-level mask, boundary, and their corresponding IoU scores. With this core module, we then follow the coarse-to-fine strategy and design our pipeline with two stages: an 1IoUNet to learn localization-based objectness cue and a hierarchical mask refiner to produce sharper and cleaner boundaries. We deploy the IBM head throughout the framework. Extensive experimental results on three grasping benchmarks manifest that our method attains the best instance segmentation performance, compared with the state-of-the-art approaches. Practically, we conduct real-world picking tests to show that with the objectness and boundary IoU scores as guidance, we are able to filter invalid (occluded) instances and select high-fidelity (exposed) instances for grasping.
Perceptive deep reinforcement learning (DRL) has lead to many recent breakthroughs for complex AI systems leveraging image-based input data. Applications of these results range from super-human level video game agents...
详细信息
ISBN:
(数字)9798350384574
ISBN:
(纸本)9798350384581
Perceptive deep reinforcement learning (DRL) has lead to many recent breakthroughs for complex AI systems leveraging image-based input data. Applications of these results range from super-human level video game agents to dexterous, physically intelligent robots. However, training these perceptive DRL-enabled systems remains incredibly compute and memory intensive, often requiring huge training datasets and large experience replay buffers. This poses a challenge for the next generation of field robots that will need to be able to learn on the edge in order to adapt to their environments. In this paper, we begin to address this issue through differentially encoded observation spaces. By reinterpreting stored imagebased observations as a video, we leverage lossless differential video encoding schemes to compress the replay buffer without impacting training performance. We evaluate our approach with three state-of-the-art DRL algorithms and find that differential image encoding reduces the memory footprint by as much as 14.2× and 16.7× across tasks from the Atari 2600 benchmark and the DeepMind Control Suite (DMC) respectively. These savings also enable large-scale perceptive DRL that previously required paging between flash and RAM to be run entirely in RAM, improving the latency of DMC tasks by as much as 32%.
The Rating Scale method has been long deemed the standard for measuring subjective perceptions. However, in the field of physical human-robot collaboration (pHRC), its aptness should be put under scrutiny due to inher...
详细信息
ISBN:
(数字)9798350384574
ISBN:
(纸本)9798350384581
The Rating Scale method has been long deemed the standard for measuring subjective perceptions. However, in the field of physical human-robot collaboration (pHRC), its aptness should be put under scrutiny due to inherent challenges such as response bias, between-subject variations, and the granularity *** variances can introduce significant bias in the rating scale results. A high granularity in the scale could overwhelm participants, leading to unclear and biased responses, while a low granularity may gloss over the fine nuances of human feelings. Additionally, there’s a notable risk of receiving careless responses, which compromise data reliability. Recognizing these challenges, this paper proposes the application of Pairwise Comparison (PC) in pHRC — an alternative survey technique that emphasizes direct comparisons between items on the defined criteria. By using the NASA Task Load Index (NASA-TLX) as a template, RS and PC questionnaires are designed and used in a series of pHRC experiments. Our preliminary findings suggest that PC is more precise and robust than the rating scale method. Compared to RS, PC fosters authentic participant interests in the experiment by intuitive question design and reducing the experimental duration. Besides, the accuracy and reliability of PC are also found to be consistent regardless of the variations in our experimental procedure design.
Automatic visual object recognition has gained great popularity in the world and is successfully applied to various areas such as robotics, security or commerce using deep learning techniques. Training in machine lear...
详细信息
ISBN:
(纸本)9781665401272
Automatic visual object recognition has gained great popularity in the world and is successfully applied to various areas such as robotics, security or commerce using deep learning techniques. Training in machine learning models based on deep learning requires an enormous amount of supervised data, which is expensive to obtain. An alternative is to use semi-supervised models as co-training where the views given by deep networks are differentiated using models that incorporate lateral information from each training object. In this document, we describe and test a co-training model for deep networks, adding as auxiliary inputs to self-supervised network features. The results show that the proposed model managed to converge using a few dozen iterations, exceeding 2 % in precision compared to recent models. This model, despite its simplicity, manages to be competitive with more complex recent works. As future work, we plan to modify deep self-supervised networks to increase diversity in co-training learning.
Exoskeleton technologies have numerous potential applications, ranging from improving human motor skills to aiding individuals in their daily activities. While exoskeletons are increasingly viewed, for example, as pro...
详细信息
ISBN:
(数字)9798350384574
ISBN:
(纸本)9798350384581
Exoskeleton technologies have numerous potential applications, ranging from improving human motor skills to aiding individuals in their daily activities. While exoskeletons are increasingly viewed, for example, as promising tools in industrial ergonomics, the effect of using them on human motor control, particularly on inter-joint coordination, remains relatively uncharted. This paper investigates the effects of generic low-amplitude force fields applied by an exoskeleton on motor strategies in asymptomatic users. The force fields mimic common perturbations encountered in exoskeletons, such as residual friction, over/under-tuned assistance, or structural elasticity. Fifty-five participants performed reaching tasks while connected to an arm exoskeleton, experiencing one of five tested force fields. Their movements before and after exposure to the exoskeleton force field were compared. The study focuses both on spatial and temporal changes in coordination using specific metrics. The results reveal that even brief exposure to a low- amplitude force field, or to uncompensated residual friction and dynamic forces, applied at the joint level can alter the interjoint coordination, while task performance remains unaffected. The tested force fields induced varying degrees of changes in joint contributions and synchronization. This study highlights the importance of monitoring coordination changes to fully understand the impact of exoskeletons on human motor control and thus enable safe and widespread adoption of those devices.
暂无评论