Fengshan Zou

Harbin Institute of Technology

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Jilai Song

Ocean University of China

Kai Jia

Guangzhou Institute of Geography

Pei Di

Beijing Jiaotong University

Daokui Qu

Chinese Academy of Sciences

Daokui Qu

Shenyang Institute of Automation

Fang Xu

Sichuan University

Fang Xu

Zhejiang University of Technology

Fang Xu

Yantai University

Daji Tian

Shenyang Institute of Automation

Hongyan Liu

Shenyang Institute of Automation

Cooperative Institutions

Chinese Academy of Sciences

Shenyang Institute of Automation

University of Chinese Academy of Sciences

Shanghai Jiao Tong University

Jilin University

Harbin Institute of Technology

Ministry of Education of the People's Republic of China

Cytoskeleton (United States)

Shandong Academy of Sciences

Qilu University of Technology

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Multi Mobile Robot Path Planning Based on Rough-Fine Search Strategy

2022 IEEE International Conference on Real-time Computing and Robotics (RCAR) (2019)

Xianqun Huang Fengshan Zou Xiaoxiao Zhu Mingjing Sun Qixin Cao

Multi-mobile robot path planning technology is the key technology in the field of intelligent warehousing and intelligent logistics. Solving the optimal solution for multi-mobile robot path planning will cause "dimension explosion". Existing multi-mobile robot path planning algorithm could not get the solution with high quality fast, causes the low efficiency of multi-mobile robot systems. This paper proposes a multi-mobile robot path planning method based on rough-fine search strategy. Firstly, the environment map is discretized to two kinds of grid maps with different grid sizes. And then, the M* algorithm is used to obtain the rough optimal solution in the grid map with large size, in the next, the rough optimal solution was mapped to the fine grid map as the teaching path, Finally, in the fine grid map, select the Gaussian distribution band with teaching path as the mean value. In the Gaussian distribution band, the RRT* algorithm is used to optimize the teaching path. Through experimental verification, our algorithm can obtain a higher quality path and has a faster solution speed.

10.1109/rcar47638.2019.9044093

Cite

Citations (2)

Improved Algorithm for YOLOX-S Object Detection Based on Diverse Branch Block (DBB)

Proceedings of the 2021 5th International Conference on Electronic Information Technology and Computer Engineering (2022)

Li Zhang Fengshan Zou Xiaofeng Wang Zizhong Wei Yang Li

The YOLO series of algorithms are commonly used in the field of object detection, and YOLOX, as one of them, is an improvement on YOLOv5, which improves the detection accuracy but decreases the inference speed. In order to further improve the detection accuracy of YOLOX and solve the problem of slow inference speed, the YOLOX-DBB algorithm is proposed on the basis of YOLOX-S with reference to the idea of structural re-parameterization. Experimental results on the VOC dataset showed that YOLOX-DBB achieved 62.88% mAP0.5:0.95, a 3.47% improvement over YOLOX-S; meanwhile, YOLOX-DBB achieved 82.46% mAP0.5, an improvement of 1.02%. In inference speed, YOLOX-DBB takes only 10.36(ms) to infer a picture, while YOLOX-S takes 12.55(ms) in the same environment. Therefore, the new network has improved detection accuracy, and the inference speed is also faster than the original network. Finally, YOLOX-DBB was applied to the art craft surface defect detection task, and the results showed that it still had better indicators than YOLOX-S.

Speedup

10.1145/3573428.3573715

Cite

Citations (2)

Robust Robot Grasp Detection in Multimodal Fusion

MATEC Web of Conferences (2017)

Qiang Zhang Daokui Qu Fang Xu Fengshan Zou

Accurate robot grasp detection for model free objects plays an important role in robotics. With the development of RGB-D sensors, object perception technology has made great progress. Reach feature expression by the colour and the depth data is a critical problem that needs to be addressed in order to accomplish the grasping task. To solve the problem of data fusion, this paper proposes a convolutional neural networks (CNN) based approach combined with regression and classification. In the CNN model, the colour and the depth modal data are deeply fused together to achieve accurate feature expression. Additionally, Welsch function is introduced into the approach to enhance robustness of the training process. Experiment results demonstrates the superiority of the proposed method.

Robustness

Sensor Fusion

Feature (linguistics)

10.1051/matecconf/201713900060

Cite

Citations (27)

Approach for accurate calibration of RGB-D cameras using spheres

Optics Express (2020)

Hongyan Liu Daokui Qu Fang Xu Fengshan Zou Jilai Song

RGB-D cameras (or color-depth cameras) play key roles in many vision applications. A typical RGB-D camera has only rough intrinsic and extrinsic calibrations that cannot provide the accuracy required in many vision applications. In this paper, we propose a novel and accurate sphere-based calibration framework for estimating the intrinsic and extrinsic parameters of color-depth sensor pair. Additionally, a method of depth error correction is suggested, and the principle of error correction is analyzed in detail. In our method, the feature extraction module can automatically and reliably detect the center and edges of the sphere projection, while excluding noise data and outliers, and the projection of the sphere center on RGB and depth images is used to obtain a closed solution of the initial parameters. Finally, all the parameters are accurately estimated within the framework of nonlinear global minimization. Compared to other state-of-the-art methods, our calibration method is easy to use and provides higher calibration accuracy. Detailed experimental analysis is performed to support our conclusions.

RGB color model

Feature (linguistics)

10.1364/oe.392414

Cite

Citations (17)

Real-time position path smoothing for robotic manipulators by constructing Composite Trajectory Spline

Mechanism and Machine Theory (2024)

Weiguang Yu Daokui Qu Fang Xu Fengshan Zou Mingmin Liu

Position (finance)

Smoothing

Spline (mechanical)

10.1016/j.mechmachtheory.2024.105855

Cite

Citations (0)

Foot terrain impact modeling and motion performance research of hexapod robot based on passive elastic leg

Journal of Physics Conference Series (2020)

Y F Liu Bo Su Lelai Zhou Haoyang Gao Liang Ding

Abstract In order to improve the performance of the multi-legged robot, the impact of foot-terrain is an important target to measure the performance of the robot. In order to ensure the compliance contact of foot-terrain for the legged robot, it is essential to add passive spring between the calf and foot. The mechanical structure makes the leg of hexapod robot retractable passive elastic structure. For multi-legged robots with passive elastic legs, the foot-terrain impact model is presented to measure the dynamic performance and provide theoretical support for the low impact force motion planning. The traditional planning method will cause additional impact for the switching case between the swing and stance phase. This paper presents a motion planning method that considers the deformable leg structure and impact performance. The experiments have been carried out, and the results have demonstrated the effectiveness of the foot terrain impact model and motion planning method in comparison with a conventional method.

Hexapod

Legged robot

Foot (prosody)

10.1088/1742-6596/1507/5/052016

Cite

Citations (1)

3D Laser Odometry for a Mobile Robot Platform

Daji Tian Fengshan Zou Fang Xu Pei Di

We propose a method for odometry in real-time using a 2D lidar with a pitching motor. The self-position estimation problem is difficult due to the measurement datasets of point clouds collected at different time stamps, and errors in motion estimation leading to the registration problem for distortion removed point cloud. To validate the proposed method, the corresponding 3D maps can be built without the need for inertial measurements. However, using an Inertial Measurement Unit (IMU) may help the motion estimation in the Levenberg-Marquardt iterative process. This paper aim to decompose difficult problem in mobile robot localization and navigation in two parts: One process calculates the robot's motion estimation in a rough way; another method matches the neighbor frame of point clouds obtained in the global reference for the accurate transform matrix. Thus we obtain the odometry of the robot platform, by implementing the precise motion estimation, we can then construct the environment map consequently. The algorithm is tested by dataset from experiments in the factory laboratory, and the reconstruction map show that our implementation is efficient and accurate.

Odometry

Iterative closest point

Distortion (music)

Visual Odometry

10.1109/cyber.2017.8446268

Cite

Citations (1)

The Research of Depth Perception Method Based on Sparse Random Grid

Dongxue Li Fang Xu Fengshan Zou Pei Di Hongyu Wang

In this paper, we propose a high-resolution depth sensing method based on structured light. In 3D contour scanning, passive binocular stereo vision is difficult to obtain enough 3D information for objects with inconspicuous surface features. To solve this problem, based on the binocular stereo vision principle and the structured light projection method in active vision, a method of obtaining sparse depth based on random mesh is proposed. Four templates are projected onto the surface of the object, which are random meshes and three templates with phase difference. In addition, the relative phase image is calculated according to the three-step phase-shifting mode. Finally, the depth map calculated in the conventional structured light method. In this paper, we demonstrate the effectiveness algorithm to show that our depth sensing are more accurate and resolution than the existing methods in the experiments.

Structured Light

Template

Depth map

10.1109/hfr.2018.8633523

Cite

Citations (0)

Error distribution estimation based weighted least square estimation for service robot hand-eye calibration

2017 IEEE 3rd Information Technology and Mechatronics Engineering Conference (ITOEC) (2017)

Fang Xu Qiang Zhang Fengshan Zou Kai Jia Yanan Zhang

Accurate hand-eye calibration is a very significant task in robotics. Many algorithms focus on L ₂ optimization, and Least Square Estimation (LSE) is usually employed to deal with this problem. For the reason that LSE can only be applied to occasions that errors obey normal distribution, LSE can lead to a lower precision of calibration result when gross errors occur. Different from the existing approaches, this paper introduces an error distribution estimation based weighted LSE for hand-eye calibration task. Firstly, transformation matrix is computed by traditional LSE. Error distribution is estimated on the basis of Gaussian kernel density estimation. Each data is weighted according to the density estimation and the fine result can be conducted by the weighted data. To evaluate the proposed method, an experiment was designed and the test result demonstrates the robustness of the proposed approach.

Robustness

Kernel density estimation

10.1109/itoec.2017.8122371

Cite

Citations (1)

Gait planning of quadruped robot based on divergence component of motion

Journal of ZheJiang University (Engineering Science) (2021)

Ming-min Liu Daokui Qu Fang Xu Fengshan Zou Kai Jia

Component (thermodynamics)

Divergence (linguistics)

10.3785/j.issn.1008-973x.2021.02.004

Cite

Citations (1)