Multi-objective optimal trajectory planning for robot manipulator attention to end-effector path limitation

Robotica, Год журнала: 2024, Номер 42(6), С. 1761 - 1780

Опубликована: Апрель 17, 2024

Abstract In the process of trajectory optimization for robot manipulator, path that is generated may deviate from intended because adjustment parameters, if there limitation end-effector in Cartesian space specific tasks, this phenomenon dangerous. This paper proposes a methodology based on Pareto front to address issue, and takes into account both multi-objective robotic arm quality path. Based dung beetle optimizer, research improved non-dominated sorting optimizer. interpolates manipulator with quintic B -spline curves, achieves simultaneously optimizes traveling time, energy consumption, mean jerk, selection strategy solution set by introducing concept Fréchet distance, enables approach desired space. Simulation experimental results validate effectiveness practicability proposed Sawyer manipulator.

Язык: Английский

---

Implementation of an Enhanced Crayfish Optimization Algorithm

Biomimetics, Год журнала: 2024, Номер 9(6), С. 341 - 341

Опубликована: Июнь 4, 2024

This paper presents an enhanced crayfish optimization algorithm (ECOA). The ECOA includes four improvement strategies. Firstly, the Halton sequence was used to improve population initialization of algorithm. Furthermore, quasi opposition-based learning strategy is introduced generate opposite solution population, increasing algorithm’s searching ability. Thirdly, elite factor guides predation stage avoid blindness in this stage. Finally, fish aggregation device effect increase ability jump out local optimal. performed tests on widely IEEE CEC2019 test function set verify validity proposed method. experimental results show that has a faster convergence speed, greater performance stability, and stronger optimal compared with other popular algorithms. applied two real-world engineering problems, verifying its solve practical problems superiority

Язык: Английский

---

Short-Term Electricity Load Forecasting Based on Improved Data Decomposition and Hybrid Deep-Learning Models

Applied Sciences, Год журнала: 2024, Номер 14(14), С. 5966 - 5966

Опубликована: Июль 9, 2024

Short-term power load forecasting plays a key role in daily scheduling and ensuring stable system operation. The problem of the volatility sequence poor prediction accuracy is addressed. In this study, learning model integrating intelligent optimization algorithms proposed, which combines an ensemble-learning based on long short-term memory (LSTM), variational modal decomposition (VMD) multi-strategy dung beetle algorithm (MODBO). aim to address shortcomings optimizer (DBO) forecasting, such as its time-consuming nature, low accuracy, ease falling into local optimum. paper, firstly, initialized using lens-imaging reverse-learning strategy avoid premature convergence algorithm. Second, spiral search used update dynamic positions breeding beetles balance global capabilities. Then, foraging are updated optimal value bootstrapping Finally, dynamic-weighting coefficients position stealing improve ability proposed new named MVMO-LSTM. Compared traditional algorithms, four-quarter averages RMSE, MAE R2 MVMO-LSTM improved by 0.1147–0.7989 KW, 0.09799–0.6937 1.00–13.05%, respectively. experimental results show that paper not only solves DBO but also enhances stability, capability information utilization model.

Язык: Английский

---

Enhancing Swarm Intelligence for Obstacle Avoidance with Multi-Strategy and Improved Dung Beetle Optimization Algorithm in Mobile Robot Navigation

Electronics, Год журнала: 2023, Номер 12(21), С. 4462 - 4462

Опубликована: Окт. 30, 2023

The Dung Beetle Optimization (DBO) algorithm is a powerful metaheuristic that widely used for optimization problems. However, the DBO has limitations in balancing global exploration and local exploitation capabilities, often leading to getting stuck optima. To overcome these address problems, this study introduces Multi-Strategy Improved (MSIDBO) Algorithm. MSIDBO incorporates several advanced computational techniques enhance its performance. Firstly, it random reverse learning strategy improve population diversity mitigate early convergence or stagnation issues present algorithm. Additionally, fitness-distance employed better manage trade-off between within population. Furthermore, utilizes spiral foraging precision, promote strong exploratory prevent being trapped further search ability particle utilization of algorithm, combines Optimal Dimension-Wise Gaussian Mutation strategy. By minimizing premature convergence, increased, accelerated. This expansion space reduces likelihood optima during evolutionary process. demonstrate effectiveness extensive experiments are conducted using benchmark test functions, comparing performance against other well-known algorithms. results highlight feasibility superiority solving Moreover, applied path planning simulation showcase practical application potential. A comparison with shows generates shorter faster paths, effectively addressing real-world

Язык: Английский

---

NLOS Identification for UWB Positioning Based on IDBO and Convolutional Neural Networks

IEEE Access, Год журнала: 2023, Номер 11, С. 144705 - 144721

Опубликована: Янв. 1, 2023

Ultra-wideband (UWB) is regarded as the technology with most potential for precise indoor location due to its centimeter-level ranging capabilities, good time resolution, and low power consumption. However, Because of presence non-line-of-sight (NLOS) error, accuracy UWB localization deteriorates significantly in harsh volatile conditions. Therefore, identifying NLOS conditions crucial enhancing location. This paper proposes a convolutional neural network (CNN) classification method based on an improved Dung Beetle Optimizer (DBO). Firstly, standard DBO, Circle chaotic mapping, non-uniform Gaussian variational strategy, multi-stage perturbation strategy are used optimize exploration capability enhance performance original DBO method, superiority-seeking ability IDBO demonstrated by testing 23 benchmark functions. In addition, algorithm, we propose IDBO-CNN model, help IDBO, identification adjusting hyperparameters CNN be closer optimal solution. Experiments conducted open-source dataset demonstrate that capable achieve desired effect. comparison conventional approach, F1-score achieved enhanced 3.31%, which demonstrates has superior accuracy.

Язык: Английский

---