This paper focuses on the mobile robot path planning problem of optimizing the performance metrics of bidirectional A* algorithm in randomized two-dimensional map environments. An algorithm called ...direction constraints adaptive extended bidirectional A* (DCAE-BA*), which is an improvement of the traditional target dynamic bidirectional A* algorithm (TTD-BA*), is proposed to improve the performance metrics of the algorithm. Regarding the improvement, we propose the adaptive extension method and the direction-constrained optimal node extension method (DCONE). Simulation experiments were conducted for DCAE-BA*, TTD-BA* and traditional A* algorithm (A*) in a large number of random two-dimensional map environments. The simulation experimental scenarios consider four types of start and end point relative directions and three obstacle proportions to objectively and comprehensively evaluate the performance of the proposed algorithms. The results show that different scenarios have a significant impact on the algorithm performance metrics. Finally, the overall performance of the proposed algorithm is evaluated with a large number of experiments in “random” scenarios, and the results show that DCAE-BA* obtains significantly better search time for all three obstacle proportions, and better path length and number of expanded nodes for 10% and 25% obstacle proportions. The effectiveness of the proposed DCAE-BA* algorithm is demonstrated, which provides an essential reference for the path planning of mobile robots in a random 2D map environment.
•DCAE-BA* is improved by adaptive extension and direction-constrained optimal node extension (DCONE).•The Start-End Dir and obstacle proportions have a significant impact on the algorithm metrics.•Overall, the DCAE-BA* has the highest computational efficiency for all three types of obstacle proportion.•DCAE-BA* optimizes the path length in the “diagonal” scenario most significantly.•As the proportion of obstacles increases, the metric optimization effect decreases.
The extraction of navigation paths is essential for the autonomous navigation of an agricultural robot. As most of the current navigation path extraction algorithms are time-consuming, difficult and ...susceptible to interference, the median point Hough transformation algorithm was proposed in this paper based on an optical system to fit the navigation paths of a greenhouse tomato-cucumber spraying robot quickly and accurately. The algorithm consists of four parts: interception of the target area, image segmentation, navigation points extraction and navigation path fitting. Image segmentation was based on the improved grayscale factor and Otsu method to achieve the segmentation of soil and plants. The navigation points then were extracted based on the relative coordinate center of the white pixel points of the binary image. Finally, the navigation path was fitted by the median point Hough transformation method. The experimental results show that the maximum error of the navigation path fitted by the improved Hough transformation was 0.5° and the time taken was 7.13 ms, which is 45. 24 ms faster than the traditional Hough transform and on average 5.1° more accurate than the least squares method.
Collision avoidance is one of the most important requirements for autonomous vehicles, particularly in complex and congested traffic scenarios where trajectories have little safety redundancy. ...However, simultaneously reaching the required accuracy and universal feasibility for different collision-avoidance behaviours is difficult due to the multi-state coupled motion of vehicles. To achieve the maximum traversability and ensure the safety of autonomous vehicles in any complex scenarios, we propose a quasi-critical collision-avoidance strategy based on a newly developed algorithm: the exclusive area-of-relative-velocity vector. This strategy first involves the construction of an exclusive area-of-velocity vector for each object vehicle to extract its position relative to the subject vehicle. In this procedure, to establish a subject-motion-decoupled scenario, projective transformation is applied to regularise the moving elliptical contour of the subject vehicle as a settled circle while retaining all positional relationships between the subject and object vehicles using the invariants. Subsequently, a group of escaping conditions for this exclusive area are established to express this quasi-critical collision-avoidance strategy explicitly and mathematically. The ultimate ability to escape from such an area is determined through theoretical derivations and experiments according to vehicle dynamics. In terms of real scenario data, a set of escaping equations is established to calculate the escaping conditions subject to the current state and the ultimate motion ability. Via scenario verifications, this strategy is shown to represent the safety boundary accurately and ensure quasi-critical collision-avoidance conditions under complex scenarios.
•The multi-state coupled motion of self-driving vehicles can be effectively decoupled while retaining their positional relationships using the motion-decoupled extraction method proposed in this study.•The exclusive area of relative velocity vector algorithm proposed in this study produces trajectories that are free of potential collisions and available for tracking.•We propose a collision avoidance strategy with equivalence, apriority, behaviour universality, and spatio-temporal uniformity for autonomous vehicles in complex traffic condition.
This paper addresses the tracking control problem of TCP/AWM network systems in presence of nonresponsive data flows of category user datagram protocol (UDP) flows. Firstly, a modified network system ...model is established by a certain suitable variable transformation, and then a fuzzy logic system (FLS) emulator is used to approximate the nonlinear terms in the network dynamics representation system. Secondly, inspired by the idea of the prescribed performance control (PPC), a novel finite-time performance function (NFTPF) is proposed. In turn, an adaptive finite-time congestion control strategy is designed by compatible usage as appropriate of a barrier Lyapunov function (BLF), the backstepping control synthesis, and an event-triggered mechanism. The proposed control strategy can not only make the tracking error to satisfy the pre-assigned transient and steady-state performance, but also ensure that all the closed-loop signals remain semi-globally uniformly ultimately bounded (SGUUB). In addition, the designed congestion control strategy eliminates potential occurrence of Zeno behavior. A set of simulation results are presented to clarify the feasibility and effectiveness of proposed methodological approach and the designed congestion controller.
•YOLOv8-GP based on key point detection is proposed.•C2f in the backbone network is replaced by C2f-Faster-EMA.•Multi-scale feature fusion is achieved by using BiFPN in Neck network.•While reducing ...the calculation, the AP of grape is improved by 3.3 %.•The positioning error of the picking point is kept within 30 pixels.
Precise positioning of fruit and picking point is crucial for harvesting table grapes using automated picking robots in an unstructured agricultural environment. Most studies employ multi-step methods for locating picking points based on fruit detection, leading to slow detection speed, cumbersome models, and algorithmic fragmentation. This study proposes an improved YOLOv8-GP (YOLOv8-Grape and picking point) model based on YOLOv8n-Pose to solve the problem of simultaneous detection of grape clusters and picking points. YOLOv8-GP is an end-to-end network that integrates object detection and key point detection. Specifically, the Bottleneck in C2f is replaced with FasterNet Block that incorporates EMA (Efficient Multi-Scale Attention), resulting in C2f-Faster-EMA. BiFPN is applied to substitute the original PAN as Neck network. The FasterNet Block, designed based on partial convolution (PConv), reduces redundant computation and memory access, thereby extracting spatial features more efficiently. The EMA attention mechanism achieves performance gains with lower computational overhead. Furthermore, BiFPN is employed to enhance the effect of feature fusion. Experimental results demonstrate that YOLOv8-GP achieves AP of 89.7 % for grape cluster detection and a Euclidean distance error of less than 30 pixels for picking point detection. Additionally, the number of Params is reduced by 47.73 %, and the model complexity GFlops is 6.1G. In summary, YOLOv8-GP offers excellent detection performance, while the reduced number of parameters and model complexity contribute to lower deployment costs and easier implementation on mobile robots.
In view of the disadvantages of the existing pose estimation algorithm, which has low real-time performance and the positioning accuracy will be greatly reduced in dynamic scene, a compound deep ...learning and parallel computing algorithm (DP-PE) is proposed. The detection algorithm based on deep learning is used to detect dynamic objects in the environment, and the dynamic feature points are removed before the matching of feature points to reduce the impact of dynamic objects on the positioning accuracy; A method for distinguishing “pseudo-dynamic objects” is proposed to solve the problem that the stationary vehicles and pedestrians in the environment are regarded as dynamic objects. The parallel computing framework for feature point extraction and matching is established on CPU-GPU heterogeneous platform to speed up DP-PE; In the localization part of DP-PE, we propose a 3D interior point detection strategy to achieve parallel search of map points, and the saturated linear kernel function is used to act on reprojection error to realize the parallelization of pose optimization. We verify the algorithm on KITTI dataset, the experimental results show that average speedup ratio of feature point extraction and matching is 6.5 times, and the overall computational efficiency of DP-PE is about 7 times higher than that before acceleration, which can realize high precision and efficient pose estimation in dynamic scene.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, UILJ, UKNU, UL, UM, UPUK
•An algorithm for detecting ripe fruit under natural illumination and occlusion conditions is proposed.•A graying factor K is proposed to remove the complex background of colorful ripe ...fruits.•Improving the Hough Transform by selecting equidistant points and increasing judgement threshold.•A real contour Hough Transform was proposed to detect ripe fruits or occluded circles.
For an accurate detection of ripe fruits under uneven illumination and ubiquitous occlusion conditions, this paper proposes a method to detect and locate ripe fruits based on machine vision. There are four key steps in this method including image graying and background removal, binary image optimization, true contour fragments extraction, and fruit fitting. For testing the proposed method, field experiments were conducted with tomato and citrus, and the ripe fruits in complex environments were successfully detected and located. From the detection experiments, it showed that the recognition rate for ripe fruits in the near zone of the proposed method was higher than 97.44%, the average time consumption was 0.2966 s, and the positioning error was less than 4.41%. In addition, it can be concluded from the comparative experiment that the proposed method is superior to conventional Hough transform, random Hough transform, and other methods based on deep learning in terms of detection rate, time performance and positioning accuracy. Therefore, it can be applied to picking robots for real-time detecting and locating ripe fruits.
This paper studies the congestion control problem of multi-bottleneck TCP/AWM networks. Firstly, considering the interference of network UDP flow and the mutual influence between nodes, a ...multi-bottle-neck network node switching model is established. Then, in order to relax the requirement of the prescribed performance technique for the initial value, a new prescribed performance function is designed, and the designed function can be well adapted to the congestion control of the multi-bottleneck network. Finally, for the newly established multi-bottleneck network node switching model and adopting a new prescribed performance function, a novel prescribed performance event-triggered controller is developed. The designed congestion controller enables different nodes of a multi-bottleneck network to achieve consistency and good stability across nodes with time-varying number of sessions, and is validated by simulation.
Abstract This article is concerned with the trajectory tracking control problem for the nonlinear systems in the sense of the predefined settling time and accuracy. In contrast with the existing ...works, we focus on the cases where the system dynamics, its bounding functions, the unmatched disturbances, and the time‐varying parameters are totally unknown; the derivatives of the desired trajectory are not required to be available. They significantly challenge the identification and/or approximation‐based control solutions. To overcome this obstacle, a novel robust prescribed performance control approach via state feedback is put forward in this article. It not only ensures the natural satisfaction of the specific initial condition but also realizes a full‐time performance specification for trajectory tracking. Furthermore, for the case of unmeasured state variables, an output‐feedback control approach is further derived by adopting an input‐driven filter and conducting trivial changes on the design procedure. Moreover, both approaches exhibit significant simplicity, without the needs for parameter identification, function approximation, disturbance estimation, derivative calculation, or command filtering. Three simulation studies are conducted to clarify and verify the above theoretical findings.
This paper presents a sliding mode control based on particle swarm optimization neural network and adaptive reaching law, and the proposed control method solves the problem of chattering and tracking ...performance degradation of a multi-joint manipulator caused by uncertainties such as external disturbances and modeling error. First, to address the problem that the precise dynamic system of the manipulator is difficult to establish, the radial basis function neural network (RBFNN) is used to approximate the uncertainty of the manipulator model, and the parameters of the neural network are optimized through the adaptive natural selection particle swarm optimization algorithm (ASelPSO) to improve the approximation ability and reduce the approximation error. Second, to eliminate chattering, adaptive reaching law is selected to improve the dynamic quality of approaching motion. Finally, a comparative simulation experiment is carried out with a 3-DOF manipulator as the research object. The results show that the control method has obvious improvements in eliminating chattering, improving tracking accuracy, and increasing convergence speed, which verifies the feasibility and superiority of the control scheme.