Search

Skin, as the crucial component of aircraft, should be milled along the edges of the frame to ensure the mounting accuracy. Traditional skin milling uses visual measurement and offline planning to generate a machining path that corresponds to the overall shape of the frame edge. To enhance adaptability and processing efficiency, an online planning method is proposed. However, it imposes greater demands on the computational efficiency, accuracy, and stability of the algorithm. In this paper, a real-time path planning method based on frame feature extraction and chronological feature optimization is proposed for the milling of attached frame-skin. The method uses a line laser to scan the surface features of both skin and frame structures, facilitating real-time extraction of feature structures. Periodic calculations of position and posture are conducted based on the time series, and a non-fitting method is used to generate a smooth and high-precision machining trajectory. The efficacy in achieving real-time path planning for the milling module is verified by experiments on prototypes of aircraft skin. The algorithm exhibits a duration of less than 8 ms in a run cycle, while maintaining accuracy within ± 0.5 mm. • Propose a real-time extraction method of the position and posture characteristics of the attached frame and skin structure. • Propose a path planning method using chronological features to improve path point continuity and establish a processing path. • The algorithm achieved a path planning period of less than 8 ms and maintained tracking path accuracy within ±0.5 mm. [ABSTRACT FROM AUTHOR]

• This paper designs a time window-based initial solution scheme to improve the quality of the initial population. • We introduce a tabu search strategy to balance the global and local search performance of the algorithm. • We adopted a fitness evaluation mechanism to provide operational plans for cross-regional agricultural machinery operations. • The paper discusses the effects of different numbers of farmlands, machinery, and operation time window lengths on the model. This article aims to solve the problems of high cost, low efficiency, and task delay caused by the lack of scientific scheduling strategies during the agricultural harvest season. Considering the constraints of operation time windows, a multimachine multiobjective cross-regional collaborative operation scheduling model is constructed with the goal of minimizing the transfer distance of agricultural machinery and nonoperational scheduling costs. A multiobjective genetic algorithm (HTSMOGA) based on a hybrid time window priority rule and a tabu search strategy is designed. The initial population is generated by utilizing time window priority rules, genes with earlier operation start times are preferentially retained, and a mixed search strategy is introduced to avoid local optimal solutions. Experiments were conducted among 24 farms in a region of Hebei Province, and the results showed that, compared with other algorithms, the HTSMOGA achieved an average reduction of 17.18 % in the transfer distance of agricultural machine operations and 19.36 % in the nonoperational waiting time. Therefore, this study provides a rational and feasible solution for the cross-regional operation scheduling of agricultural machinery cooperatives and provides theoretical support for ensuring the timely completion of harvesting tasks and the cost savings and efficiency of agricultural machinery operations. [ABSTRACT FROM AUTHOR]

The path-planning algorithms for autonomous mobile robot navigation are crucial, often relying on sampling-based methods. RRT* is a robust, sampling-based path planning algorithm. The sampling process in RRT* plays a pivotal role, where uniform sampling can lead to slow convergence, while non-uniform sampling offers faster convergence but may struggle in complex environments due to its limited exploration. Thus, achieving a balance between exploitation and exploration is essential when selecting the sampling method for the RRT* path-planning algorithm. To address this issue, this research paper introduces Hybrid-RRT*, a path planning method that utilizes hybrid sampling. This unique approach generates samples using both non-uniform and uniform samplers. Hybrid-RRT* is evaluated against three baseline path planning algorithms—RRT*-N, Informed RRT*, and RRT*—across three different 384x384 2D simulation environments. Compared to these baseline methods, Hybrid-RRT* achieves superior results across all five performance metrics: convergence rate, success rate, number of nodes visited, path length, and planning time. According to the numerical results, the proposed algorithm achieves a faster average convergence rate that is 76.14% higher than RRT*, 24% higher than Informed RRT*, and 3.33% higher than RRT*-N. Moreover, it reduces node exploration by an average of 48.53% compared to RRT* and 40.83% compared to Informed RRT*. The simulation results demonstrate that the proposed Hybrid-RRT* algorithm effectively addresses the issue of slow convergence with uniform sampling and the challenge of limited exploration with non-uniform sampling methods. • A hybrid sampling method is proposed that combines uniform and non-uniform sampling. • A novel goal-directed strategy is proposed for the selection of non-uniform samples. • A random selection method is proposed to trade off non-uniform and uniform samplers. [ABSTRACT FROM AUTHOR]

• Design conceptualization, mechanisms, modeling and control strategies for efficient glass façade cleaning robot. • Adhesion force characteristics of suction cups for dry and wet glass wall surface. • Energy efficient robot adhesion, continuous locomotion and steering control. • Hybrid PID-PSO control algorithm for path planning simulation and analysis. • Working trials of the developed glass façade cleaning robot to validate the concept. This paper presents a unique design concept, dynamic modeling, and control strategies for efficient coverage path planning of a glass façade cleaning robot (GFCR). The robot design has been conceptualized using mechanisms for adhesion, motion, steering, and cleaning. The dynamic model for robot vertical glass façade cleaning is derived using Lagrangian formulation. A modified particle swarm optimization (PSO) is used to autotune the proportional, integral, and derivative (PID) parameters for the trajectory tracking simulation and it is more efficient and robust compared to the standard PSO algorithm. The path planning algorithm using hybrid PID-PSO approach is also developed for energy-efficient coverage of the robot for glass façade cleaning. The coverage algorithm illustrates the energy-performance of the GFCR for different paths viz., horizontal line sweep (HLS), vertical line sweep (VLS), spiral line sweep (SLS), and special cell diffusion (SCD) motion. Simulation reveals the robot motion for HLS path is the most energy-efficient. The GFCR model with minimum energy consumption has been validated by working trials. The GFCR has potential applications for cleaning high-rise glass façade buildings and photovoltaic (PV) solar panels. [ABSTRACT FROM AUTHOR]