Task Planning Research Articles

A-star algorithm vs. Q-learning algorithm for path planning: A comparative analysis

In robotics research, the core focus is on path planning, as it's essential for robots to navigate according to human objectives to be valuable. This paper provides an in-depth analysis of the commonly used A-Star algorithm and Q-learning algorithm in mobile robot path planning. It specifically examines the advantages and disadvantages of both the A-Star algorithm and the Q-learning algorithm, highlighting their roles in the advancement of robotic intelligence. The A-star algorithm is an algorithm with a heuristic function to guide path selection direction. Also, the A-star algorithm is guaranteed to identify the most efficient path in terms of traffic cost when facing static obstacles. However, its performance degrades when dealing with trap nodes. Specifically, the algorithm occupies a large amount of memory and takes a long computational time. In contrast, the Q-learning algorithm is effective for learning continuous actions with the same network data structure for local path planning tasks.

A Lightweight Multi-Label Classification Method for Urban Green Space in High-Resolution Remote Sensing Imagery

Urban green spaces are an indispensable part of the ecology of cities, serving as the city’s “purifier” and playing a crucial role in promoting sustainable urban development. Therefore, the refined classification of urban green spaces is an important task in urban planning and management. Traditional methods for the refined classification of urban green spaces heavily rely on expert knowledge, often requiring substantial time and cost. Hence, our study presents a multi-label image classification model based on MobileViT. This model integrates the Triplet Attention module, along with the LSTM module, to enhance its label prediction capabilities while maintaining its lightweight characteristic for standalone operation on mobile devices. Trial outcomes in our UGS dataset in this study demonstrate that the approach we used outperforms the baseline by 1.64%, 3.25%, 3.67%, and 2.71% in mAP,F1,precision, and recall, respectively. This indicates that the model can uncover the latent dependencies among labels to enhance the multi-label image classification device’s performance. This study provides a practical solution for the intelligent and detailed classification of urban green spaces, which holds significant importance for the management and planning of urban green spaces.

A deep reinforcement learning with dynamic spatio-temporal graph model for solving urban logistics delivery planning problems

ABSTRACT The urban logistics delivery planning problems are a crucial component of urban spatial decision analysis. Most studies typically focus on traditional urban logistics delivery planning problems and ignore real-time traffic information. With the advancement of urbanization, real-time traffic networks play a critical role. However, previous studies have utilized heuristic methods to solve urban logistics delivery planning with real-time traffic information problems, and few studies have applied deep reinforcement learning methods to tackle this problem. Deep reinforcement learning methods solving traditional logistics delivery planning problems overlook the impact of dynamic spatio-temporal features on route planning. In this study, we propose a new deep reinforcement method called DRLDSTG. The method introduces the dynamic spatio-temporal graph model into a deep reinforcement learning method to capture these dynamic features from urban logistics delivery planning tasks. The actor-critic with maximum entropy method is employed to train the model and determine the optimal policy function. The experimental results indicated that the proposed method can achieve a superior solution with faster computational efficiency compared to commercial software and heuristic methods. Compared to other deep reinforcement learning methods, our method can more effectively learn dynamic spatio-temporal features from environments, demonstrating promising applications in cities.

The Impact of Regional Carbon Emission Reduction on Corporate ESG Performance in China

The integrated planning of central and local emission reduction tasks is crucial for achieving sustainable economic development, and corporate ESG performance aligns with the principles of sustainable development, having become a prominent topic in academic research. This paper empirically investigates the impact of regional carbon emission reductions on the ESG performance of local enterprises from 2009 to 2021 using provincial carbon emission data from China. The findings indicate that regional carbon emission reductions significantly enhance the ESG performance of local firms. The underlying mechanism is that regional carbon emission reductions facilitate local enterprises obtaining green credit, attracting media coverage and green investors and thus improving ESG performance. Second, heterogeneity tests reveal that regional carbon emission reductions enhance the ESG performance of local firms more significantly in regions with stricter environmental regulations, within heavily polluted industries, and among less digitized enterprises. Finally, further analysis demonstrates that regional residents’ carbon emission reductions can enhance the ESG performance of local enterprises, with regional carbon emission reductions exerting a dual effect after improving ESG performance. The findings of this study provide valuable insights into the low-carbon development of various economic entities and the collaborative promotion of economic green transformation.

THE ROUTES OF UNMANNED AERIAL VEHICLE PLANNING METHODS ONTOLOGY

The article analyzes modern methods of the route of unmanned aerial vehicles (UAVs) planning. An ontology for the routes of the UAVs group planning algorithms classification has been developed, which includes two elements: global planning and local planning. The article notes that the purpose of UAV flight routes planning is the optimal (rational) solution of specific tasks under appropriate weather conditions and taking into account other factors of the external environment. According to the premise of taking into account a significant number of constraints, the objective function of planning is to reach each UAV from the group to the end point (mission point). Planning the UAVs group route is a complex task of its multi-criteria optimization and solution in the presence of constraints. With the increase in the UAVs number, the analytical space of the route planning task will grow exponentially. The existing methods of planning the of the UAVs group movement route have a number of disadvantages, therefore, the development of a comprehensive method that reduces computational requirements, saves time, allows to plan routes in real time, and is also more effective from the group's mission performance point of view. The article defines the requirements for the UAVs group routes planning promising methods: complex modeling of many factors of the external environment influence; planning in real time; integration of planning methods. Combining the characteristics of heuristic algorithms and machine learning methods also allows you to obtain complex methods that have a higher adaptability to the conditions of the external environment and expand the range of applications.

On some aspects of studying the topic «Russia: trials and heroes» within the framework of the discipline «Fundamentals of Russian Statehood»

The article considers the issue of teaching the new course «Fundamentals of Russian Statehood», which has been introduced everywhere into the programs of universities in the country only since 2023, which determines its relevance. The authors are trying to solve the problem: how to awaken a sense of patriotism in young people, convey to them the values of Russian civilization and do this in a way that does not cause rejection or resistance?The goal of the research is to find a theoretical and methodological basis for sensitive, but at the same time effective teaching of the discipline «Fundamentals of Russian Statehood» and to demonstrate the chosen pedagogical strategy using the example of studying the topic «Russia: Trials and Heroes».The authors of the article, using methods of semantic and structural analysis, have formed an ideological framework and set the logic for selecting material, which allows them to provide the necessary information within the chosen topic in a targeted and orderly manner.The result of the research was methodological recommendations and materials for studying the topic «Russia: trials and heroes». In order to avoid a moralizing tone and excessive pathos, it has been proposed to adhere to such pedagogical principles and techniques as focusing on dialogue, relying on previously formed knowledge and ideas, searching for connections between this topic and modernity and professional interests that are relevant for certain groups of students.The Conclusions. The article has a clear practical focus. It presents methodological developments, including, among other things, a logically structured plan and specific tasks on the topic «Russia: trials and heroes».

Comfort navigation improvement of path planning task in human-robot interaction

Comfort navigation improvement of path planning task in human-robot interaction

Unlocking underrepresented use-cases for large language model-driven human-robot task planning

Large language models (LLM) are now the de facto task planners for Embodied AI (EAI) systems. This shift can be attributed to LLMs' powerful, emergent properties which enable their adaptation to downstream tasks with minimal to no fine tuning via prompting. However, we find that LLM-driven task planning is not a solved problem. In this work we measure the extent to which these models can be adapted to complex and domain-specific task planning via few-shot prompting. Additionally, we contribute quantitative and qualitative analysis on prompt robustness. Lastly, to meet the challenges of adapting EAI systems to real-world, industrial domains, we adopt a human-in-the-loop approach to guarantee safe and interpretable task planning and execution. We successfully demonstrate co-located, human-robot teaming where an Augmented Reality (AR) headset mediates information exchanged between an EAI agent and human operator for a variety of inspection tasks. To our knowledge the use of an AR headset for multimodal grounding and the application of EAI to industrial tasks are novel contributions within Embodied AI research.

A CLOUD COMPUTING PARADIGM IN ROBOT MOTION PLANNING: INNOVATIONS AND PROSPECTS

The paper presents a new approach to the simulation of robot motion planning using cloud computing technologies. Robot motion planning plays a crucial role in robotics and enables robots to efficiently control and perform tasks in various environments. Using cloud computing, the proposed simulation framework offers significant advantages in terms of scalability, resource utilization, and availability. The paper begins with an overview of the potential applications of cloud computing in robot motion planning and simulation tasks. It then presents the architecture and components of a cloud-based simulation framework, highlighting its advantages over traditional methods. Overall, this paper provides valuable insights into the integration of cloud technologies into robotics, paving the way for more efficient and scalable simulation solutions.

ArchGPT: harnessing large language models for supporting renovation and conservation of traditional architectural heritage

The renovation of traditional architecture contributes to the inheritance of cultural heritage and promotes the development of social civilization. However, executing renovation plans that simultaneously align with the demands of residents, heritage conservation personnel, and architectural experts poses a significant challenge. In this paper, we introduce an Artificial Intelligence (AI) agent, Architectural GPT (ArchGPT), designed for comprehensively and accurately understanding needs and tackling architectural renovation tasks, accelerating and assisting the renovation process. To address users’ requirements, ArchGPT utilizes the reasoning capabilities of large language models (LLMs) for task planning. Operating under the use of tools, task-specific models, and professional architectural guidelines, it resolves issues within the architectural domain through sensible planning, combination, and invocation. Ultimately, ArchGPT achieves satisfactory results in terms of response and overall satisfaction rates for customized tasks related to the conservation and restoration of traditional architecture.

Computational Mechanisms Underlying Multi-Step Planning Deficits in Methamphetamine Use Disorder.

Current theories suggest individuals with methamphetamine use disorder (iMUDs) have difficulty considering long-term outcomes in decision-making, which could contribute to risk of relapse. Aversive interoceptive states (e.g., stress, withdrawal) are also known to increase this risk. The present study analyzed computational mechanisms of planning in iMUDs, and examined the potential impact of an aversive interoceptive state induction. A group of 40 iMUDs and 49 healthy participants completed two runs of a multi-step planning task, with and without an anxiogenic breathing resistance manipulation. Computational modeling revealed that iMUDs had selective difficulty identifying the best overall plan when this required enduring negative short-term outcomes - a mechanism referred to as aversive pruning. Increases in reported craving before and after the induction also predicted greater aversive pruning in iMUDs. These results highlight a novel mechanism that could promote poor choice in recovering iMUDs and create vulnerability to relapse.

An SD-LV Calculation Model for the Scale of the Urban Rail Transit Network

The planning for the scale of the urban rail transit network (URTN) is one of the key tasks of URTN planning. The scale should match the urban development (UD). A reasonable scale can improve travel efficiency, increase economic activities, and promote UD, while an unreasonable scale may consume more urban resources, fail to meet urban transportation demands, and even inhibit UD. Currently, the URTN scale is primarily determined by qualitative analyses and static indicators, which leads to the scale does not match UD perfectly. To determine a reasonable scale, a System Dynamics–Lotka–Volterra (SD-LV) model is constructed. The SD model is adopted to simulate the dynamic interaction between the URT and UD. The LV (Lotka–Volterra) model is employed to calculate the scale, in which the mutualism coefficients are proposed to characterize the mutualistic relationships between the URT and UD. The model is validated by using a dataset of the Beijing URTN from 2017 to 2021. The simulation errors of the URTN scale range from −4.3% to 1.32%, which demonstrates the robustness and effectiveness of the proposed model. The study offers quantitative theoretical insights for determining the reasonable scale of the URTN.

DEFINITION OF THE RECONSTRUCTION METHOD OF A PUBLIC BUILDING OR STRUCTURE

A systematic approach to the organization and management of various types of work within a particular industry is what makes the research relevant, including the creation of a well-established method for selecting methodological solutions. The various industries that make up the structure of the national economy have certain distinctive features. Construction, as one of the most complex industries from an organizational and technological point of view, is of particular interest to study. The various types and areas of activity in its composition define narrow areas of research. The definition of the reconstruction method as an integral part of the organizational and technological system is a serious and urgent task of urban planning. The construction of new buildings and structures for cultural, educational purposes and spiritual and ethical education is currently carried out much less frequently than residential buildings. Accordingly, it is often necessary to reconstruct buildings and structures that have already been built. The reconstruction process includes a significant number of types of work, has its own specifics that distinguish it from new construction. The choice of the method of reconstruction of a public building or structure involves making decisions based on the established tasks for each specific object of reconstruction. The article considers the concept of reconstruction, systematizes the tasks of reconstruction of public buildings and structures, develops an algorithm for choosing a reconstruction method, and considers the factors that determine decision-making at certain stages of work.

A Real-Time Resource Dispatch Approach for Edge Computing Devices in Digital Distribution Networks Considering Burst Tasks

Edge computing technology can effectively solve huge challenges posed by the large number of terminal devices accessing and massive data processing in digital distribution networks. Burst tasks, such as faults and data requests from the cloud, can occur at any time for edge computing devices in distribution networks. These tasks are unpredictable and usually hold the highest priority and must be completed as soon as possible. Although resources can be reserved partially at each period in the pre-scheduled operation plan, they may still be insufficient to handle burst tasks adequately. A real-time resource dispatch approach for burst tasks is developed in this study to address the above problems. The concept of flexibility for edge computing devices is presented, determining the real-time dispatch duration. Real-time resource dispatch and task handling processing are analyzed in detail, considered as task real-time dispatch models, computation process real-time dispatch constraints, and resource limitation constraints. The proposed real-time resource dispatch approach takes full advantage of the transferable characteristics for partial original plan tasks to adjust the pre-scheduled operation plan and release flexible resources for immediate processing of the burst task, completing burst tasks quickly and minimizing the impact for previous planned tasks on the edge computing device. The capability of the proposed method to efficiently deal with the burst tasks is also verified by the case study.

Hierarchical online automated planning for a flexible manufacturing system

Task planning and action planning for workshop machines are essential for modern manufacturing. Traditionally, these two problems are solved independently with elaborate manual methods. However, personalized customization introduces more dynamic exogenous events into the manufacturing system, and it is then impossible to consider all possible dynamic scenarios manually. This paper focuses on online automated planning, generating new plans automatically in response to new dynamic situations. We first formulate the planning problem for a flexible manufacturing system as a fully observable nondeterministic planning problem. Second, a hierarchical automated online planning approach is presented. Finally, the effectiveness of the proposed approach is verified by an ARIAC 2022 competition environment.

Real-time local path planning strategy based on deep distributional reinforcement learning

Local path planning and obstacle avoidance in complex environments are two challenging problems in the research of intelligent robots. In this study, we develop a novel approach grounded in deep distributional reinforcement learning to address these challenges. Within this methodology, agents instantiated by deep neural networks perceive real-time local environmental information through sensor data, addressing inherent stochasticity and local path planning tasks in complex environments. End-to-end training is facilitated via distributional reinforcement learning algorithms and reward functions informed by heuristic knowledge. Optimal actions for path planning are determined through return value distributions. Finally, the simulation results show that the success rate of the proposed distributed algorithm is 98% in a random environment and 94% in a dynamic environment. This proves that the algorithm has better generalization and flexibility than the non-distributed algorithm.

Synthesis and Verification of Mission Plans for Multiple Autonomous Agents under Complex Road Conditions

Mission planning for multi-agent autonomous systems aims to generate feasible and optimal mission plans that satisfy given requirements. In this article, we propose a tool-supported mission-planning methodology that combines (i) a path-planning algorithm for synthesizing path plans that are safe in environments with complex road conditions, and (ii) a task-scheduling method for synthesizing task plans that schedule the tasks in the right and fastest order, taking into account the planned paths. The task-scheduling method is based on model checking, which provides means of automatically generating task execution orders that satisfy the requirements and ensure the correctness and efficiency of the plans by construction. We implement our approach in a tool named MALTA, which offers a user-friendly GUI for configuring mission requirements, a module for path planning, an integration with the model checker UPPAAL, and functions for automatic generation of formal models, and parsing of the execution traces of models. Experiments with the tool demonstrate its applicability and performance in various configurations of an industrial case study of an autonomous quarry. We also show the adaptability of our tool by employing it in a special case of an industrial case study.

Quality-centered production and maintenance scheduling for multi-machine manufacturing systems under variable operating conditions

Production, maintenance, and quality represent the three primary tasks in manufacturing systems. Previous studies have predominantly focused on optimizing these tasks separately or jointly optimizing two of them. However, it is worth noting that these tasks exhibit mutual interaction, and models failing to consider them simultaneously cannot achieve global optimality. Therefore, this paper aims to develop a quality-centered production and maintenance scheduling methodology (QPMSM) for multi-machine and multi-product manufacturing systems. Specifically, a delivery time-based production scheduling strategy is introduced to emphasize the impact of maintenance activities. A quality loss model is developed to characterize the mapping relationship between quality deterioration and component degradation. Furthermore, a well-designed maintenance strategy determines the optimal timing and frequency of maintenance activities to improve system performance and reduce quality deterioration. Building upon this foundation, an integrated model is proposed to concurrently optimize production scheduling, maintenance planning, and quality control tasks to minimize the overall system cost. The effectiveness of the proposed QPMSM in terms of cost-savings has been verified using a designed hybrid meta-heuristic algorithm, providing a viable approach for simultaneous optimization.

A framework for neurosymbolic robot action planning using large language models.

Symbolic task planning is a widely used approach to enforce robot autonomy due to its ease of understanding and deployment in engineered robot architectures. However, techniques for symbolic task planning are difficult to scale in real-world, highly dynamic, human-robot collaboration scenarios because of the poor performance in planning domains where action effects may not be immediate, or when frequent re-planning is needed due to changed circumstances in the robot workspace. The validity of plans in the long term, plan length, and planning time could hinder the robot's efficiency and negatively affect the overall human-robot interaction's fluency. We present a framework, which we refer to as Teriyaki, specifically aimed at bridging the gap between symbolic task planning and machine learning approaches. The rationale is training Large Language Models (LLMs), namely GPT-3, into a neurosymbolic task planner compatible with the Planning Domain Definition Language (PDDL), and then leveraging its generative capabilities to overcome a number of limitations inherent to symbolic task planners. Potential benefits include (i) a better scalability in so far as the planning domain complexity increases, since LLMs' response time linearly scales with the combined length of the input and the output, instead of super-linearly as in the case of symbolic task planners, and (ii) the ability to synthesize a plan action-by-action instead of end-to-end, and to make each action available for execution as soon as it is generated instead of waiting for the whole plan to be available, which in turn enables concurrent planning and execution. In the past year, significant efforts have been devoted by the research community to evaluate the overall cognitive capabilities of LLMs, with alternate successes. Instead, with Teriyaki we aim to providing an overall planning performance comparable to traditional planners in specific planning domains, while leveraging LLMs capabilities in other metrics, specifically those related to their short- and mid-term generative capabilities, which are used to build a look-ahead predictive planning model. Preliminary results in selected domains show that our method can: (i) solve 95.5% of problems in a test data set of 1,000 samples; (ii) produce plans up to 13.5% shorter than a traditional symbolic planner; (iii) reduce average overall waiting times for a plan availability by up to 61.4%.

Understanding Sample Generation Strategies for Learning Heuristic Functions in Classical Planning

We study the problem of learning good heuristic functions for classical planning tasks with neural networks based on samples represented by states with their cost-to-goal estimates. The heuristic function is learned for a state space and goal condition with the number of samples limited to a fraction of the size of the state space, and must generalize well for all states of the state space with the same goal condition. Our main goal is to better understand the influence of sample generation strategies on the performance of a greedy best-first heuristic search (GBFS) guided by a learned heuristic function. In a set of controlled experiments, we find that two main factors determine the quality of the learned heuristic: the algorithm used to generate the sample set and how close the sample estimates to the perfect cost-to-goal are. These two factors are dependent: having perfect cost-to-goal estimates is insufficient if the samples are not well distributed across the state space. We also study other effects, such as adding samples with high-value estimates. Based on our findings, we propose practical strategies to improve the quality of learned heuristics: three strategies that aim to generate more representative states and two strategies that improve the cost-to-goal estimates. Our practical strategies result in a learned heuristic that, when guiding a GBFS algorithm, increases by more than 30% the mean coverage compared to a baseline learned heuristic.