Dynamic Task Research Articles

Dynamic Task Planning for Multi-Arm Harvesting Robots Under Multiple Constraints Using Deep Reinforcement Learning

Global fruit production costs are increasing amid intensified labor shortages, driving heightened interest in robotic harvesting technologies. Although multi-arm coordination in harvesting robots is considered a highly promising solution to this issue, it introduces technical challenges in achieving effective coordination. These challenges include mutual interference among multi-arm mechanical structures, task allocation across multiple arms, and dynamic operating conditions. This imposes higher demands on task coordination for multi-arm harvesting robots, requiring collision-free collaboration, optimization of task sequences, and dynamic re-planning. In this work, we propose a framework that models the task planning problem of multi-arm operation as a Markov game. First, considering multi-arm cooperative movement and picking sequence optimization, we employ a two-agent Markov game framework to model the multi-arm harvesting robot task planning problem. Second, we introduce a self-attention mechanism and a centralized training and execution strategy in the design and training of our deep reinforcement learning (DRL) model, thereby enhancing the model’s adaptability in dynamic and uncertain environments and improving decision accuracy. Finally, we conduct extensive numerical simulations in static environments; when the harvesting targets are set to 25 and 50, the execution time is reduced by 10.7% and 3.1%, respectively, compared to traditional methods. Additionally, in dynamic environments, both operational efficiency and robustness are superior to traditional approaches. The results underscore the potential of our approach to revolutionize multi-arm harvesting robotics by providing a more adaptive and efficient task planning solution. We will research improving the positioning accuracy of fruits in the future, which will make it possible to apply this framework to real robots.

Neuromuscular Strategies in Dominant and Non-Dominant Legs in Dancers During Dynamic Balance Tasks.

Introduction: Ballet-based dance training emphasizes the equal development of both legs. However, dancers often perceive differences between their legs during balance or landing. There still needs to be more consensus on the functional difference between dominant (D) and non-dominant legs (ND). Therefore, this study investigated both legs' neuromuscular strategy in single-leg balance and landing based on leg dominance. Methods: Thirteen female dancers (age: 22.2 ± 2.8 years) with no history of ankle injuries in the past year participated in the study. Based on the questionnaire, the dominant leg was set, which legs were preferred to balance, generate strength, and land. Joint kinematics and ground reaction force (GRF) were analyzed using a three-dimensional motion analysis system and force plates during single-leg balance (SLB), passé balance (PB), sissone simple (SS), and sissone ouverte (SO). The tibialis anterior (TA), medial gastrocnemius (MG), peroneus longus (PL), and gluteus medius (GM) activation were measured by wireless surface electromyography (EMG). Displacement (cm) of the center of pressure and the dynamic postural stability index (DPSI), a balanced score post dynamic tasks, were calculated. Results: Bilateral leg balance ability was observed based on joint kinematics and DPSI during SLB, PB, SS, and SO. Higher TA activity was noted during PB in ND legs than in D legs (P = .038). PL activation was significantly increased in ND legs (69.3 ± 34.4%) than in D legs (45.6 ± 19.2%) before contact during SS and SO (P < .05). After landing with ND legs, dancers regulated postural stability with increasing TA activation (P < .05). Conclusions: Pre-activation of PL before landing with ND legs increases ankle stiffness, enhancing stability. Conversely, D legs achieve balance with lower activation levels. The findings highlight significant differences between legs in dancers, suggesting that leg dominance should be considered in future training and performance strategies.

Effects of kinesio taping on lower limb biomechanical characteristics during dynamic postural control tasks in individuals with chronic ankle instability.

Previous studies have demonstrated significant biomechanical differences between individuals with chronic ankle instability (CAI) and healthy controls during the Y-balance test. This study aimed to examine the effects of kinesio taping (KT) on lower limb biomechanical characteristics during the Y-balance anterior reach task in individuals with CAI. A total of 30 participants were recruited, comprising 15 individuals with CAI and 15 healthy controls. All participants were randomly assigned three taping conditions: no taping (NT), placebo taping (PT), and KT, followed by the Y-balance anterior reach task. Each condition was separated by one-week intervals. Kinematic and kinetic data of the lower limbs during the movement phase were collected using the Vicon motion capture system (Vicon, T40, 200 Hz) and two Kistler force platforms (Kistler, 1000 Hz). KT significantly improved the Y-balance anterior reach distance (P = 0.003) and peak ankle eversion angle (P = 0.019) compared to NT. Additionally, KT resulted in increased peak knee flexion angle (P = 0.002, P = 0.011) and peak ankle dorsiflexion angle (P <0.001, P = 0.005) relative to both NT and PT. KT also significantly reduced mediolateral center of pressure (COP) displacement (P = 0.001) and average velocity of mediolateral COP displacement (P = 0.033) in comparison to NT. Furthermore, KT decreased mediolateral center of gravity displacement (P = 0.002, P = 0.003) relative to both NT and PT. KT significantly improved abnormal ankle posture by promoting greater ankle dorsiflexion and eversion angles. Additionally, KT reduced mediolateral COP displacement and average velocity to improve postural stability. These changes may contribute to reduced risk of ankle sprains. Therefore, KT may serve as an effective tool for managing recurrent ankle sprains in individuals with CAI.

Intelligent Data-Driven Task Offloading Framework for Internet of Vehicles Using Edge Computing and Reinforcement Learning

Introduction: The Internet of Vehicles (IoV) was enabled through innovative developments featuring advanced automotive networking and communication to fulfill the need for real-time applications that are latency-sensitive, such as autonomous driving and emergency management. Given that the servers were much farther away from the actual site of operation, traditional cloud computing faced huge delays in processing. Mobile Edge Computing (MEC) resolved this challenge by enabling localized data processing, reducing latency and enhancing resource utilization.Methods: This study proposed an Efficient Mobile Edge Computing-based Internet of Vehicles Task Offloading Framework (EMEC-IoVTOF). The framework integrated deep reinforcement learning (DRL) to optimize task offloading decisions, focusing on minimizing latency and energy consumption while accounting for bandwidth and computational constraints. Offloading costs were calculated using mathematical modeling and further optimized through Particle Swarm Optimization (PSO). An adaptive inertia weight mechanism was implemented to avoid local optimization and enhance task allocation decisions.Result: The proposed framework was thus proved effective for any latency reduction and energy consumption optimization in efficiently improving the overall system performance. DRL and MEC together facilitate scalability in task distribution by ensuring robust performance in dynamic vehicular environments. Integration with PSO further enhances the decision-making process and makes the system highly adaptable to dynamic task demands and network conditions.Discussion:The findings highlighted the potential of EMEC-IoVTOF to address key challenges in IoV systems, including latency, energy efficiency, and bandwidth utilization. Future research could explore real-world deployment and adaptability to complex vehicular scenarios, further validating its scalability and reliability.

Multi-scale convolution and dynamic task interaction detection head for efficient lightweight plum detection

Multi-scale convolution and dynamic task interaction detection head for efficient lightweight plum detection

Scalable and energy-efficient task allocation in industry 4.0: Leveraging distributed auction and IBPSO.

Industry 4.0 has transformed manufacturing with the integration of cutting-edge technology, posing crucial issues in the efficient task assignment to multi-tasking robots within smart factories. The paper outlines a unique method of decentralizing auctions to handle basic tasks. It also introduces an improved variant of the improved Binary Particle Swarm Optimization (IBPSO) algorithm to manage complicated tasks that require multi-robot collaboration. The main contributions we make are: the design of an auction decentralization algorithm (AOCTA) which allows for an efficient and flexible task distribution in dynamic contexts, the optimization of coalition formation in complex jobs by using IBPSO and improves the efficiency of energy and decreases the cost of computation as well as thorough simulations that show that our proposed method significantly surpasses conventional methods for efficiency, task completion rates in terms of energy usage, task completion rate, and scaling of the system. This research contributes to the development of smart manufacturing through providing an effective solution that aligns with the sustainability objectives and addresses operational efficiency as well as environmental impacts. Addressing the challenges posed by dynamic task allocation in distributed multi-robot systems, these advanced technologies provide a comprehensive solution, facilitating the evolution of innovative manufacturing systems.

A Study of Two-Wheel Self-Balancing Robot Dynamics for Control Systems Learning

The two-wheeled self-balancing robot (TWSBR), utilizing an inverted pendulum arrangement, exhibits dynamic stability while being statically unstable. Due to this behavior, the TWSBR system has been used to demonstrate fundamental principles of stability, nonlinear dynamics, and control theory. In this work, a computational study of the TWSBR was carried out to develop a mathematical model using Newtonian mechanics that serves as an educational tool in control education for demonstrating the impact of different control algorithms such as Proportional (P), ProportionalDerivative (PD), and Proportional-Integral-Derivative (PID) on system stability. The comparison of P, PD, and PID controllers was undertaken to demonstrate the practical advantages of each controller in improving the robot's stability and responsiveness. The PID controller was designed and optimized using the Root-locus techniques, with a damping ratio equal to 1 which exhibits a fast-settling time with minimum overshoots. When an impulse response was applied to the PID controller in the simulation environment it demonstrated system can reach a dynamic balance within 1.2 seconds demonstrating the effectiveness of the proposed PID controller. The optimal gain parameters K value and Kp, Ki, and Kd parameter values were determined using root-locus analysis, which allowed the system to operate with great stability and minimum settling time. These findings highlight the TWSBR's educational significance as a hands-on tool for teaching practical applications of control theory, as well as providing valuable insights into the design and optimization of control systems for dynamic balance tasks. This study emphasizes the relevance of control algorithms in achieving rapid and stable dynamic behaviors, making it a valuable learning resource for both students and instructors. Keywords: Two-wheeled self-balancing robot, Inverted pendulum, P, PD and PID controllers, Stability analysis, Control Education

An Improved Method for Enhancing the Accuracy and Speed of Dynamic Object Detection Based on YOLOv8s.

Accurate detection and tracking of dynamic objects are critical for enabling skill demonstration and effective skill generalization in robotic skill learning and application scenarios. To further improve the detection accuracy and tracking speed of the YOLOv8s model in dynamic object tracking tasks, this paper proposes a method to enhance both detection precision and speed based on YOLOv8s architecture. Specifically, a Focused Linear Attention mechanism is introduced into the YOLOv8s backbone network to enhance dynamic object detection accuracy, while the Ghost module is incorporated into the neck network to improve the model's tracking speed for dynamic objects. By mapping the motion of dynamic objects across frames, the proposed method achieves accurate trajectory tracking. This paper provides a detailed explanation of the improvements made to YOLOv8s for enhancing detection accuracy and speed in dynamic object detection tasks. Comparative experiments on the MS-COCO dataset and the custom dataset demonstrate that the proposed method has a clear advantage in terms of detection accuracy and processing speed. The dynamic object detection experiments further validate the effectiveness of the proposed method for detecting and tracking objects at different speeds. The proposed method offers a valuable reference for the field of dynamic object detection, providing actionable insights for applications such as robotic skill learning, generalization, and artificial intelligence-driven robotics.

A detailed inquiry of the differences between headphones and loudspeakers influences on dynamic postural task performance.

We examined the impact of auditory stimuli and their methods on a dynamic balance task performance. Twenty-four young adults wore an HTC Vive headset and dodged a virtual ball to the right or left based on its color (blue to the left, red to the right, and vice versa). We manipulated the environment by introducing congruent (auditory stimuli from the correct direction) or incongruent (auditory stimuli played randomly from either side) and comparing a multimodal (visual and congruent auditory stimuli) to unimodal (visual or auditory stimuli) presentation. We tested four apparatuses: loudspeakers, headphones, passthrough, (wearing headphones while auditory stimuli come from loudspeakers) and room simulation (externalization via headphones). We quantified reaction time (RT) and accuracy (choosing the correct direction to dodge) from the head movement. We hypothesized that the weight of the headset will slow RT, and that externalization of the auditory stimuli will make it more usable when no visual cues are provided. Interestingly, both hypotheses were refuted. In silent conditions, RT was faster with headphones compared to loudspeakers, but this difference disappeared when auditory stimuli were introduced. Participants used congruent auditory stimuli to improve accuracy but disregarded incongruent auditory stimuli across all apparatuses except for room simulation. In conclusion, this study confirmed that healthy young adults can use congruent auditory stimuli to enhance accuracy and disregard incongruent auditory stimuli such that accuracy is not harmed. RT was either faster or the same with headphones compared to loudspeakers. Notably, this specific room simulation did not enhance performance.

Development of a Fleet Management System for Multiple Robots’ Task Allocation Using Deep Reinforcement Learning

This paper presents a fleet management system (FMS) for multiple robots, utilizing deep reinforcement learning (DRL) for dynamic task allocation and path planning. The proposed approach enables robots to autonomously optimize task execution, selecting the shortest and safest paths to target points. A deep Q-network (DQN)-based algorithm evaluates path efficiency and safety in complex environments, dynamically selecting the optimal robot to complete each task. Simulation results in a Gazebo environment demonstrate that Robot 2 achieved a path 20% shorter than other robots while successfully completing its task. Training results reveal that Robot 1 reduced its cost by 50% within the first 50 steps and stabilized near-optimal performance after 1000 steps, Robot 2 converged after 4000 steps with minor fluctuations, and Robot 3 exhibited steep cost reduction, converging after 10,000 steps. The FMS architecture includes a browser-based interface, Node.js server, rosbridge server, and ROS for robot control, providing intuitive monitoring and task assignment capabilities. This research demonstrates the system’s effectiveness in multi-robot coordination, task allocation, and adaptability to dynamic environments, contributing significantly to the field of robotics.

Neuroticism's link to threat sensitivity: Evidence from a dynamic affect reactivity task.

The personality trait of neuroticism has been theoretically linked to threat sensitivity, but this perspective of neuroticism has resulted in mixed findings, arguably because mood states, rather than emotional reactions, have been examined. The present studies (total N = 519) administered a task capable of assessing emotional reactions-to appetitive versus aversive images-in a nearly continuous manner, parsing threat sensitivity in terms of emotional onsets, peak amplitudes, and prototypicality in responding. In the context of this tight temporal focus, higher levels of neuroticism tended to be associated with faster emotional onsets when aversive images were involved. In addition, neuroticism by valence interactions occurred with respect to peak amplitudes and prototypic patterning, with negativity effects for these parameters being amplified at higher, relative to lower, levels of neuroticism. These results link the neuroticism and emotion dynamics literatures while providing novel support for perspectives that emphasize neuroticism's link to threat sensitivity, in the present context defined in terms of faster, stronger, and more prototypical reactions to acute aversive events. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Goal-Directed Learning in Cortical Organoids.

Experimental neuroscience techniques are advancing rapidly, with major recent developments in high-density electrophysiology and targeted electrical stimulation. In combination with these techniques, cortical organoids derived from pluripotent stem cells show great promise as in vitro models of brain development and function. Although sensory input is vital to neurodevelopment in vivo , few studies have explored the effect of meaningful input to in vitro neural cultures over time. In this work, we demonstrate the first example of goal-directed learning in brain organoids. We developed a closed-loop electrophysiology framework to embody mouse cortical organoids into a simulated dynamical task (the inverted pendulum problem known as 'Cartpole') and evaluate learning through high-frequency training signals. Longitudinal experiments enabled by this framework illuminate how different methods of selecting training signals enable improvement on the tasks. We found that for most organoids, training signals chosen by artificial reinforcement learning yield better performance on the task than randomly chosen training signals or the absence of a training signal. This systematic approach to studying learning mechanisms in vitro opens new possibilities for therapeutic interventions and biological computation.

A decomposition-based dynamic constrained multi-objective task assignment for heterogeneous crowdsensing

Efficient task allocation is a crucial issue of Mobile crowdsensing (MCS). Generally, only homogeneous mobile users like human are selected as the participants, causing a difficulty to meet the spatiotemporal coverage demand on human-unreachable regions. To overcome this drawback, unmanned aerial vehicles are introduced to form heterogeneous MCS, which can be formulated into a dynamic constrained multi-objective task allocation model. Taking the maximum average sensing quality of all tasks and the maximum average remaining budget for each subtask as the optimization objectives, an improved decomposition-based multi-objective evolutionary algorithm is presented to find the optimal allocation scheme. Specifically, the problem is first decomposed into a set of dynamic constrained scalar subproblems. For each subproblem, a stochastic configuration network (SCN)-based initialization is developed to produce the promising population, in which SCNs learn the probabilities of mobile users being allocated to each task. Following that, a reinforcement learning-based autonomous evolutionary strategy is adopted to recommend the most appropriate solvers in terms of the state of current population. A hybrid population update mechanism is then employed to form the high-quality offspring, with the purpose of balancing the feasibility, convergence and diversity. The extensive experiments on 20 dynamic instances are conducted to demonstrate the effectiveness of proposed algorithm compared to other task allocation algorithms.

A Complex-Former Tracker With Dynamic Polar Spatio-Temporal Encoding.

Recently, the excellent performance of transformer has attracted the attention of the visual community. Visual transformer models usually reshape images into sequence format and encode them sequentially. However, it is difficult to explicitly represent the relative relationship in distance and direction of visual data with typical 2-D spatial structures. Also, the temporal motion properties of consecutive frames are hardly exploited when it comes to dynamic video tasks like tracking. Therefore, we propose a novel dynamic polar spatio-temporal encoding for video scenes. We use spiral functions in polar space to fully exploit the spatial dependences of distance and direction in real scenes. We then design a dynamic relative encoding mode for continuous frames to capture the continuous spatio-temporal motion characteristics among video frames. Finally, we construct a complex-former framework with the proposed encoding applied to video-tracking tasks, where the complex fusion mode (CFM) realizes the effective fusion of scenes and positions for consecutive frames. The theoretical analysis demonstrates the feasibility and effectiveness of our proposed method. The experimental results on multiple datasets validate that our method can improve tracker performance in various video scenarios.

Dynamic Task Offloading for Multi-UAVs in Vehicular Edge Computing With Delay Guarantees: A Consensus ADMM-Based Optimization

Dynamic Task Offloading for Multi-UAVs in Vehicular Edge Computing With Delay Guarantees: A Consensus ADMM-Based Optimization

A Dynamic Task Allocation Method for Unmanned Aerial Vehicle Swarm Based on Wolf Pack Labor Division Model

A Dynamic Task Allocation Method for Unmanned Aerial Vehicle Swarm Based on Wolf Pack Labor Division Model

DAWN: Dynamic Task Planning of Multi-UAV With Two-Layer Optimization Mechanism in Uncertain Environments

DAWN: Dynamic Task Planning of Multi-UAV With Two-Layer Optimization Mechanism in Uncertain Environments

Identifying Parkinson’s disease and its stages using static standing balance

The current assessment of Parkinson’s disease (PD) relies on dynamic motor tasks, limiting accessibility. This study aimed to propose an innovative approach to identifying PD and its stages using static standing balance and machine learning. A total of 210 participants were recruited, including a control group and five PD groups categorized by stage. Each participant completed a 10-s static standing balance task in which center of pressure trajectory data in the medial-lateral and anterior-posterior directions were collected. Features were extracted from these trajectory data and the data derived from them using both representation learning and handcrafting methods. A Transformer encoder-based classifier was trained on these features and achieved an F1-score of 0.963 in classifying the six study groups. This approach enhances the accessibility of PD assessment, enabling earlier detection and timely intervention. The novel data mining framework introduced in this study heralds a new era of time-series data-driven digital healthcare.

Hybrid Agile-Kanban frameworks for workflow adaptability: A proposed solution for innovation in project management

Agile and Kanban methodologies have become indispensable in modern project management, emphasizing workflow adaptability, iterative processes, and streamlined task prioritization to meet dynamic project requirements. However, traditional applications of these methodologies often face challenges such as workflow inefficiencies, misaligned task priorities, and bottlenecks caused by static configurations and manual adjustments. A hybrid approach that combines Agile methodologies with Kanban principles, including visualization tools and flow control, offers a promising solution to these challenges. This paper presents a comprehensive exploration of a Hybrid Agile-Kanban Framework (HAKF), designed to enhance project workflow optimization, team collaboration, and task execution. By integrating artificial intelligence (AI) for real-time feedback loops and dynamic task prioritization, the proposed framework addresses common bottlenecks and resource imbalances, enabling more efficient and scalable project management practices. Despite its transformative potential, challenges such as tool integration, scalability, and initial adoption barriers must be addressed for widespread implementation. To mitigate these issues, this paper introduces a conceptual HAKF model that seamlessly integrates with existing project management tools, providing actionable insights for optimizing workflows. This study lays a foundation for future research into hybrid methodologies, advancing the potential of Agile-Kanban integration to create more efficient and adaptive project management systems.

Monitoring Anthropogenically Disturbed Parcels with Soil Erosion Dynamics Change Based on an Improved SegFormer

Amidst burgeoning socioeconomic development, anthropogenic activities have exacerbated soil erosion. This erosion, characterized by its brief duration, high frequency, and considerable environmental degradation, presents a major challenge to ecological systems. Therefore, it is imperative to regulate and remediate erosion–prone, anthropogenically disturbed parcels, with dynamic change detection (CD) playing a crucial role in enhancing management efficiency. Currently, traditional methods for change detection, such as field surveys and visual interpretation, suffer from time inefficiencies, complexity, and high resource consumption. Meanwhile, despite advancements in remote sensing technology that have improved the temporal and spatial resolution of images, the complexity and heterogeneity of terrestrial cover types continue to limit large–scale dynamic monitoring of anthropogenically disturbed soil erosion parcels (ADPSE) using remote sensing techniques. To address this, we propose a novel ISegFormer model, which integrates the SegFormer network with a pseudo–residual multilayer perceptron (PR–MLP), cross–scale boundary constraint module (CSBC), and multiscale feature fusion module (MSFF). The PR–MLP module improves feature extraction by capturing spatial contextual information, while the CSBC module enhances boundary prediction through high– and low–level semantic guidance. The MSFF module fuses multiscale features with attention mechanisms, boosting segmentation precision for diverse change types. Model performance is evaluated using metrics, such as precision, recall, F1–score, intersection over union (IOU), and mean intersection over union (mIOU). The results demonstrate that our improved model performs exceptionally well in dynamic monitoring tasks for ADPSE. Compared to five other models, our model achieved an mIOU of 72.34% and a Macro–F1 score of 83.55% across twelve types of ADPSE changes, surpassing the other models by 1.52–2.48% in mIOU and 2.25–3.64% in Macro–F1 score. This work provides a theoretical and methodological foundation for policy–making in soil and water conservation departments.