TXTong Xu

Tong Xu

*This is an algorithmically generated author record based on publicly available information.

Tong Xu has a total of 21 co-authors. Here are the top co-authors by number of shared publications:

  • Dong Wang

    School of Instrument Science and Engineering, Southeast University, Nanjing 210096, China

    4 shared publications
  • Weigong Zhang

    School of Instrument Science and Engineering, Southeast University, Nanjing 210096, China

    3 shared publications
  • Weigong Zhang

    Southeast University,School of Instrument Science and Engineering,Nanjing,China,210096

    2 shared publications
  • Cancan Chu

    School of Instrument Science and Engineering, Southeast University, Nanjing, China

    1 shared publication
  • Dong Wang

    Southeast University,School of Instrument Science and Engineering,Nanjing,China,210096

    1 shared publication
  • Dong Wang

    Chair Materials for Electrical Engineering and Electronics Institute of Materials Science and Engineering and Institute of Micro and Nanotechnologies MacroNano TU Ilmenau 98693 Ilmenau Germany

    1 shared publication

Latest Publications

  • Open Access

    EPSO-based rigid robotic arm for obstacle avoidance object grasping

    by Tong Xu, Xiaomin Han, Honghui Liu, Yuhong Li

    The smart orchard tournament of the robot developer competition requires participants to utilise a robotic arm equipped with a monocular camera to grasp ping-pong balls bearing fruit patterns. Common issues include low object recognition accuracy, inefficient performance, and grasping failures due to collisions with obstacles. This paper proposes a novel object grasping framework, EPSO-TPPP-YOLOv8, based on a rigid robotic arm for the task of fruit picking. In the safe distance analysis section, the two obstacles, cylinder and sphere, are first modelled. The robotic arm is then considered as an object consisting of several cylinders. The formula for calculating the safe distance between the robot arm and the obstacle is provided, and the design of the obstacle avoidance path is guided according to the calculation results. In the robot arm path planning section, the hyper-parameters of the angles of each robot arm joint are first determined, and then particle swarm optimization based on a novel set of strategies and evaluation criteria is used to generate the desired paths. The findings reveal that, in the absence of human intervention, the proposed EPSO algorithm reliably identifies paths that adhere to safety boundary limits within 50 iterations. The EPSO-TPPP framework, when operated under human guidance, has been demonstrated to achieve this within 10 iterations. In the subsequent phase of vision-based object recognition phase, a distinctive dataset is constructed based on the specified task. It is an established fact that simulation and field experiments have demonstrated the efficacy of the robot arm’s obstacle-avoidance grasping functionality.


  • Open Access

    Implementation of two-roller scheduling path planning under road construction scenarios

    by Tong Xu, Junhao Liu, Zhixuan Zhou

    Previous studies have verified the feasibility of single unmanned roller tracking paths and have effectively evaluated the performance of pavement compaction. Nevertheless, the issue of scheduling faulty rollers (e.g., insufficient oil pressure) during collaborative pavement construction with multiple rollers has not been fully investigated, despite its prevalence in engineering practice. Although there are some patents that propose solutions for several cases, there is a lack of comprehensive, detailed and process-oriented scheduling strategies for specific scenarios. This paper introduces a framework for pavement construction comprising one paver and six rollers, and proposes a scheduling strategy for idle and faulty robots with the objective of addressing the problem of scheduling faulty rollers. Specifically, the traditional path planning task requires known beginning and end positions, while the scheduling position in this paper can be designed in advance. Consequently, this paper presents a methodology that leverages the ratio of discrete Bezier curve path points to discrete idle region points at distinct scheduling positions to ascertain the start and end positions of faulty and idle robots. In the scheduling process implementation phase, the paper considers the scene constraints and the model constraints of the rollers, and proposes a novel cost function that balances path length and safety distance. Furthermore, to address the issue of driving in opposite directions in the narrow passage, this paper proposes the interleaved scheduling scheme with the objective of enhancing the performance upper bound of the algorithm by significantly increasing the probability of finding a feasible solution. Moreover, the implementation of a discrete sampling of curve and position points ensures that the algorithm runs in an acceptable time. The results of comparative simulations demonstrate that the path planning algorithm proposed in this paper is more effective than alternative algorithms in addressing the specific application scenarios of this project. Furthermore, the integration of the interleaved scheduling framework has a considerable impact on the enhancement of the quality of the generated paths. The algorithm is capable of successfully completing the scheduling task of faulty and idle rollers with a road length of more than 100 m in a road width of 10.5 m, while ensuring that the other rollers are not affected in the completion of their respective tasks. The path search time is 24 s, the generated planning path length is 178.1 m and the safety distance is approximately 1.6 m.


  • Open Access

    High importance feature selection and DV‐OSR‐QSED strategy for open‐set recognition

    by Tong Xu

    A significant challenge in the domain of anti‐drone warfare is the identification of enemies or own aircraft through the analysis of data broadcast by drones (e.g. ADS‐B). This issue can be conceptualized as an open set recognition (OSR) problem. This paper proposes a DV‐OSR‐QSED framework for the purpose of data visualization‐based OSR (DV‐OSR). Phase‐based 2D high‐importance features are extracted, the DV‐OSR framework is designed and mapped to 2D, and the 5th and 95th quantile selection‐Euclidean distance (QSED) strategy is proposed. Experiments show that by using the proposed framework, the correct classification rate for known and unknown samples is 96.04% and 95.79%, the recall rate and F 1 value are 89.00% and 92.27%, and the AUC is 0.9630.

  • Top Cited Papers
    Open Access

    Recent advances in Rapidly-exploring random tree: A review

    by Tong Xu


  • A novel multi-featured decision system for multi-classification tasks

    by Tong Xu

    Feature engineering is a difficult task, and for real signal data, it is difficult to find a certain feature that can easily distinguish all classes. Multiple features can provide more information, which means the fusion of multi-feature learning strategies has potential significant advantages. Based on this premise, this paper proposes a multi-class framework based on the multi-featured decision to distinguish all the different classes, and takes automatic dependent surveillance-broadcast (ADS-B) signal data as an example, first extracts the phase features and wavelet decomposition features of the signal data, then selects the features with high discrimination between classes, then proposes a one-dimensional residual neural network based on 16 convolutional layers to learn the unique features of different features and classes separately, and finally proposes a novel multi-featured decision method based on voting method and a priori probability. Results show that the proposed one-dimensional residual neural network has better performance metrics on the test set compared to some machine learning-based and neural network-based algorithms, with classification accuracies of 86.1%, 84.6% and 83.6% on wavelet decomposition features, raw features and phase features, respectively, on ADS-B preamble signals. The proposed feature decision framework based on the voting method and a priori probability has a recall, precision and F1 value of 80.24%, 89.89% and 84.79% on ADS-B preamble signals, respectively.


  • Multi-roller Trajectory Planning under the Dynamic Obstacle in Pavement Construction

    by Tong Xu, Dong Wang, Weigong Zhang

    This paper investigates a multi-roller trajectory planning algorithm under the dynamic obstacle in pavement construction. When autonomous rollers compact the road surface, the lack of fuel in the fuel tank needs to be considered. To deal with this problem, the roller that is about to run out of oil is moved away from the working area, and at the same time, a spare roller completes the remaining work belonging to the offsite roller. The two trajectory planning methods are as follows: First, we design a cubic Bezier curve through the known state variables to find all feasible paths for the spare roller, and then, we detect the curvature of the paths and filter out paths that the roller cannot follow, and next, we filter paths that collide with the dynamic obstacle, and finally, we choose the shortest path length. The same steps are used to design the path of the roller that is about to run out of oil, except that the maximum safety distance is used in the last step instead of choosing the shortest path length. The simulation experiment shows that in the scenario of this article, the designed path has a safe distance of more than 3 meters and a length of no more than 82 meters.


  • A non-overshooting controller for vehicle path following

    by Tong Xu, Dong Wang, Weigong Zhang

    Unmanned pavement construction is of great significance in China, and one of the most important issues is how to follow the designed path near the boundary of the pavement construction area to avoid curbs or railings. In this paper, we raise a simple yet effective controller, named the proportional-integral-radius and improved particle swarm optimization (PIR-IPSO) controller, for fast non-overshooting path-following control of an unmanned articulated vehicle (UAV). Firstly, UAV kinematics model is introduced and segmented UAV steering dynamics model is built through field experiments; then, the raw data collected by differential global positioning system (DGPS) is used to build the measurement error distribution model that simulates positioning errors. Next, line of sight (LOS) guidance law is introduced and the LOS initial parameter is assigned based on human driving behavior. Besides, the initial control parameters tuned by the Ziegler-Nichols (ZN) method are used as the initial iterative parameters of the PSO controller. An improved PSO fitness function is also designed to achieve fast non-overshoot control performance. Experiments show that compared with the PSO, ZN and ZN-PSO controller, the PIR-PSO-based controller has significantly less settling time and almost no overshoot in various UAV initial states. Furthermore, compared with other controllers, the proposed PIR-IPSO-based controller achieves precise non-overshoot control, relatively less settling time and centimeter-level positioning error in various initial deviations.


  • Open Access

    A four-level test system for evaluating pavement compaction performance of autonomous articulated vehicles

    by Tong Xu, Dong Wang, Zuodong Xiao, Cancan Chu, Weigong Zhang

    This article develops a four-level test system for accurately evaluating pavement compaction performance of autonomous articulated vehicles. In the evaluation layer, various performance indicators are evaluated, including the stability, rapidity and accuracy of trajectory tracking, and the ratio of required compaction to actual compaction once and twice and compaction repeatability index when pavement compaction. The guidance and control layer can be described in terms of theory and application. At the theoretical level, the line of sight guidance algorithm and incremental proportional integral control algorithm are introduced to eliminate system control lag. Among them, the best line of sight guidance and incremental proportional integral control parameters are selected by the Elitist strategies genetic algorithm, and the initial parameters are set according to human driving experience initial control parameters. At the application level, the BECKHOFF controller, a kind of programmable logic controller, acts as the main guidance and control unit in the four-level control system, fixed speed is given to the autonomous articulated vehicle by setting the engine speed and transmission gear, and steering wheel angle is adjusted in real time by the BECKHOFF controller. In the sensor level, a simplified sensor configuration is used to reduce overall cost. The comparative simulation results of no controller, the incremental proportional integral controller, line of sight guidance-incremental proportional integral controller with human driving experience initial control parameters, line of sight guidance-incremental proportional integral controller with random initial control parameters, and elitist strategies genetic algorithm-line of sight guidance-incremental proportional integral controller with human driving experience initial control parameters manifest evidently that the proposed elitist strategies genetic algorithm-line of sight guidance-incremental proportional integral controller with human driving experience initial control parameters has almost no steady-state error, no overshoot, and short settling time. Field results show that ratio of required compaction to actual compaction once achieves 100%, ratio of required compaction to actual compaction twice achieves 94.6%, and compaction repeatability index achieves 35%.


  • Open Access

    Path Planning for Autonomous Articulated Vehicle Based on Improved Goal-Directed Rapid-Exploring Random Tree

    by Tong Xu, Yang Xu, Dong Wang, Siwei Chen, Weigong Zhang, Lihang Feng

    The special steering characteristics and task complexity of autonomous articulated vehicle (AAV) make it often require multiple forward and backward movements during autonomous driving. In this paper, we present a simple yet effective method, named head correction with fixed wheel position (HC-FWP), for the demand of multiple forward and backward movements. The goal-directed rapid-exploring random tree (GDRRT) algorithm is first used to search for a feasible path in the obstacle map, and then, the farthest node search (FNS) algorithm is applied to obtain a series of key nodes, on which HC-FWP is used to correct AAV heading angles. Simulation experiments with Dynapac CC6200 articulated road roller parameters show that the proposed improved goal-directed rapid-exploring random tree (IGDRRT), consisting of GDRRT, FNS, and HC-FWP, can search a feasible path on maps that require the AAV to move back and forth.