检索结果-内蒙古大学图书馆

BEV2PR: BEV-Enhanced Visual Place Recognition with Structural Cues

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Ge, Fudong Zhang, Yiwei Shen, Shuhan Hu, Weiming Wang, Yue Gao, Jin CASIA China School of Artificial Intelligence University of Chinese Academy of Sciences China School of Information Science and Technology ShanghaiTech University China State Key Laboratory of Industrial Control Technology Institute of Cyber-Systems and Control Zhejiang University China

In this paper, we propose a new image-based visual place recognition (VPR) framework by exploiting the structural cues in bird’s-eye view (BEV) from a single monocular camera. The motivation arises from two key observations about place recognition methods based on both appearance and structure: 1) For the methods relying on LiDAR sensors, the integration of LiDAR in robotic systems has led to increased expenses, while the alignment of data between different sensors is also a major challenge. 2) Other image-/camera-based methods, involving integrating RGB images and their derived variants (e.g., pseudo depth images, pseudo 3D point clouds), exhibit several limitations, such as the failure to effectively exploit the explicit spatial relationships between different objects. To tackle the above issues, we design a new BEV-enhanced VPR framework, namely BEV2PR, generating a composite descriptor with both visual cues and spatial awareness based on a single camera. The key points lie in: 1) We use BEV features as an explicit source of structural knowledge in constructing global features. 2) The lower layers of the pre-trained backbone from BEV generation are shared for visual and structural streams in VPR, facilitating the learning of fine-grained local features in the visual stream. 3) The complementary visual and structural features can jointly enhance VPR performance. Our BEV2PR framework enables consistent performance improvements over several popular aggregation modules for RGB global features. The experiments on our collected VPR-NuScenes dataset demonstrate an absolute gain of 2.47% on Recall@1 for the strong Conv-AP baseline to achieve the best performance in our setting, and notably, a 18.06% gain on the hard set. Copyright © 2024, The Authors. All rights reserved.

关键词：

Model predictive control with fractional-order delay compensation for fast sampling systems

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2021年第7期64卷 165-180页

作者： Ze ZHOU Zhitao LIU Hongye SU Liyan ZHANG State Key Laboratory of Industrial Control Technology Institute of Cyber-Systems and ControlZhejiang University School of Automation Wuhan University of Technology

Model predictive control(MPC) is widely used in fast sampling systems owing to its fast regulating ability. However, the sampling delay is a key issue and tends to be a fractional multiple of the sampling period. If the fractional-order delay is not accurately offset, the controller output will exhibit errors, thus resulting in oscillations in controlled system. Moreover, the MPC delay compensation algorithm is limited to the computation time. To address the problems of fractional delay and computational burden in fast sampling systems, we propose a new method to compensate for the fractional-order sampling ***, we use a finite-impulse-response fractional delay filter based on a Lagrange interpolation polynomial to approximate the fractional portion. Moreover, we prove that high accuracy and simplicity can be ensured when the polynomial order is one. Then, we estimate the current state variable using the delayed sampling signal and control signals of past moments. Further, we obtain the current control signal according to the estimated state variable. By considering the simultaneous existence of computational and sampling delays,a full compensation strategy is proposed. Computational simulation results validate the proposed MPC algorithm with fractional-order delay compensation and demonstrate its advantages.

关键词： model predictive control fast sampling systems sampling delay fractional-order Lagrange interpolation polynomial

A Decoupled and Linear Framework for Global Outlier Rejection over Planar Pose Graph

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wu, Tianyue Gao, Fei The State Key Laboratory of Industrial Control Technology Institute of Cyber-Systems and Control Zhejiang University Hangzhou310027 China

We propose a robust framework for planar pose graph optimization contaminated by loop closure outliers. Our framework rejects outliers by first decoupling the robust PGO problem wrapped by a Truncated Least Squares kernel into two subproblems. Then, the framework introduces a linear angle representation to rewrite the first subproblem that is originally formulated in rotation matrices. The framework is configured with the Graduated Non-Convexity (GNC) algorithm to solve the two non-convex subproblems in succession without initial guesses. Thanks to the linearity property of the angle representation, our framework requires only a linear solver to optimally solve the optimization problems encountered in GNC. We extensively validate the proposed framework, named DEGNC-LAF (DEcoupled Graduated Non-Convexity with Linear Angle Formulation) in planar PGO benchmarks. It turns out that it runs significantly (sometimes up to over 30 times) faster than the standard and general-purpose GNC while resulting in high-quality estimates. Copyright © 2022, The Authors. All rights reserved.

关键词： Statistics

Safe Dynamic Motion Generation in Configuration Space Using Differentiable Distance Fields

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chi, Xuemin Li, Yiming Huang, Jihao Dai, Bolun Liu, Zhitao Calinon, Sylvain State Key Laboratory of Industrial Control Technology Institute of Cyber-Systems and Control Zhejiang University Hangzhou China Idiap Research Institute Martigny1920 Switzerland Lausanne1015 Switzerland Fauna Robotics New YorkNY United States

Generating collision-free motions in dynamic environments is a challenging problem for high-dimensional robotics, particularly under real-time constraints. control Barrier Functions (CBFs), widely utilized in safety-critical control, have shown significant potential for motion generation. However, for high-dimensional robot manipulators, existing QP formulations and CBF-based methods rely on positional information, overlooking higher-order derivatives such as velocities. This limitation may lead to reduced success rates, decreased performance, and inadequate safety constraints. To address this, we construct time-varying CBFs (TVCBFs) that consider velocity conditions for obstacles. Our approach leverages recent developments on distance fields for articulated manipulators, a differentiable representation that enables the mapping of objects’ position and velocity into the robot’s joint space, offering a comprehensive understanding of the system’s interactions. This allows the manipulator to be treated as a point-mass system thus simplifying motion generation tasks. Additionally, we introduce a time-varying control Lyapunov function (TVCLF) to enable whole-body contact motions. Our approach integrates the TVCBF, TVCLF, and manipulator physical constraints within a unified QP framework. We validate our method through simulations and comparisons with state-of-the-art approaches, demonstrating its effectiveness on a 7-axis Franka robot in real-world experiments. © 2024, CC BY.

关键词： Lyapunov functions

Tactile-Based Object Pose Estimation Employing Extended Kalman Filter 8

学校读者我要写书评

暂无评论

Tactile-Based Object Pose Estimation Employing Extended Kalm...

8th IEEE International Conference on Advanced Robotics and Mechatronics, ICARM 2023

作者： Lin, Qiguang Yan, Chaojie Li, Qiang Ling, Yonggen Zheng, Yu Lee, Wangwei Wan, Zhaoliang Huang, Bidan Liu, Xiaofeng College of IoT Engineering Hohai University Jiangsu Key Laboratory of Special Robotic Technology Jiangsu Changzhou213022 China Zhejiang University Institute of Cyber-System and Control State Key Laboratory of Industrial Control and Technology Hangzhou China Bielefeld33619 Germany Tencent Robotics X Shenzhen China School of Computer Science and Engineering Sun Yat-sen University Guangzhou China

ISBN: (纸本)9798350300178

In this paper, we present a new approach to estimate the pose of an object being manipulated by a multi-fingered robotic hand. The method utilizes advanced tactile sensors with high spatial resolution to optimize the estimation of the object's pose using an Extended Kalman Filter (EKF) based approach. We defined and derived the state and measurement equations, as well as evaluated the estimation accuracy in grasping tasks. The approach is able to effectively account for the pose transition caused by tactile pushing, and the mapping from the object's pose to the contact position and normal direction as measured by the tactile sensor. The method was evaluated in multiple grasping experiments in simulation scenarios. Results show that the estimation can converge towards the ground truth in a relatively short period of time, with displacement and rotation errors remaining within acceptable levels. This new method has the potential to improve the accuracy and reliability of robotic grasping and manipulation tasks. © 2023 IEEE.

关键词： Extended Kalman filters

Knowledge Distillation on Driving Intention Generator: Learn Human-like Semantic Reasoning

学校读者我要写书评

暂无评论

Knowledge Distillation on Driving Intention Generator: Learn...

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

作者： Hejun Wang Jiajun Dong Yunkai Wang Yanmei Jiao Rong Xiong the State Key Laboratory of Industrial Control Technology and Institute of Cyber-Systems and Control Zhejiang University Hangzhou China the School of Information Science and Engineering Hangzhou Normal University Hangzhou China

Semantic information plays a significant role in autonomous vehicles. However, it has remained open how to economically compute and efficiently incorporate semantic information into other frequently applied processes, such as construction of local navigation map, driving decision making and trajectories planning. To address this issue, we introduce Knowledge Distillation to convey semantic reasoning capability. Following the existing intention-guided framework, a semantically incorporated module associates given navigation instruction with acquired visual perception to synthesize driving intention. Our major contribution is to teach the driving intention generator how to propagate considering semantics implicitly. Specifically, we construct a semantic supervisor identical to the driving intention generator. The supervisor transfers semantic reasoning ability to the generator during training process, with which the generator is expected to perform stably and reliably even in extremely sophisticated surroundings or under frequently inaccurate navigation instruction. Through exhaustive experimental validation, our model demonstrates a superior robustness and adaptability compared to the state-of-the-art. Finally, we explain the intrinsic mechanism of our model with visualization.

关键词：

Driving Performance Analysis for Connected Vehicle Platoons: Velocity Perturbation and Fuel Consumption

学校读者我要写书评

暂无评论

Driving Performance Analysis for Connected Vehicle Platoons:...

2023 IEEE International Conference on Unmanned systems, ICUS 2023

作者： Zhao, Yihan Duan, Xuting Zhou, Jianshan Zhang, Long School of Transportation Science and Engineering Beihang University Beijing China State Key Lab of Intelligent Transportation System Beijing Key Laboratory for Cooperative Vehicle Infrastructure Systems and Safety Control Beijing China State Key Lab of Intelligent Transportation System Beijing Key Laboratory for Cooperative Vehicle Infrastructure Systems and Safety Control Zhongguancun Laboratory Beijing China National Key Laboratory of Science and Technology on Information System Security Institute of Systems Engineering Beijing China

ISBN: (纸本)9798350316308

Vehicle platoon is one of the concrete manifestations of multi-vehicle cooperative control and intelligent transportation. However, the velocity perturbation in the driving process will bring about stability and safety problems. The connected autonomous vehicles (CAVs) can obtain the motion status of the lead vehicle in real time and respond to it in time, which is of far-reaching significance to platoon control. Based on this, this paper focuses on the longitudinal driving strategy of connected autonomous vehicle platoons. Specifically, we provide a discrete double-integral model controller for each vehicle in a generic scenario, and the controller performs acceleration and deceleration control at each timeslot according to the state of the vehicle in front and its own state. Simulation experiments are designed by varying the size of the platoon and the number of connected autonomous vehicles. The experimental results show that the introduction of connected autonomous vehicles can effectively suppress the platoon velocity perturbation, increase the stability of the platoon driving, and significantly reduce the fuel consumption at the same time. © 2023 IEEE.

关键词： Autonomous vehicles

L2COcc: Lightweight Camera-Centric Semantic Scene Completion via Distillation of LiDAR Model

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wang, Ruoyu Ma, Yukai Yao, Yi Tao, Sheng Li, Haoang Zhu, Zongzhi Liu, Yong Zuo, Xingxing Institute of Cyber-Systems and Control Zhejiang University Hangzhou China School of Computation Information and Technology Technical University of Munich Munich Germany Zhejiang Guoli Security Technology Co. Ltd. Ningbo China State Key Laboratory of Industrial Control Technology China

Semantic Scene Completion (SSC) constitutes a pivotal element in autonomous driving perception systems, tasked with inferring the 3D semantic occupancy of a scene from sensory data. To improve accuracy, prior research has implemented various computationally demanding and memory-intensive 3D operations, imposing significant computational requirements on the platform during training and testing. This paper proposes L2COcc, a lightweight camera-centric SSC framework that also accommodates LiDAR inputs. With our proposed efficient voxel transformer (EVT) and cross-modal knowledge modules, including feature similarity distillation (FSD), TPV distillation (TPVD) and prediction alignment distillation (PAD), our method substantially reduce computational burden while maintaining high accuracy. The experimental evaluations demonstrate that our proposed method surpasses the current state-of-the-art vision-based SSC methods regarding accuracy on both the SemanticKITTI and SSCBench-KITTI-360 benchmarks, respectively. Additionally, our method is more lightweight, exhibiting a reduction in both memory consumption and inference time by over 23% compared to the current state-of-the-arts method. Code is available at our project page: https://***/L2COcc/. © 2025, CC BY.

关键词： Benchmarking

CT-NeRF: Incremental Optimization of Neural Radiance Field and Camera Poses with Complex Trajectory

学校读者我要写书评

暂无评论

IEEE Transactions on Circuits and systems for Video technology 2025年

作者： Ran, Yunlong Li, Yanxu Ye, Qi Huo, Yuchi Cui, Zhaopeng Bai, Zechun Sun, Jiahao Chen, Jiming Zhejiang University College of Control Science and Engineering Hangzhou China Zhejiang University College of Control Science and Engineering State Key Laboratory of Industrial Control Technology China Key Laboratory of Collaborative Sensing and Autonomous Unmanned Systems of Zhejiang Province China Zhejiang University State Key Lab of CAD and CG China

Neural radiance field (NeRF) has achieved impressive results in high-quality 3D scene reconstruction. However, NeRF heavily relies on precise camera poses. While recent works like BARF have introduced camera pose optimization within NeRF, their applicability is limited to simple trajectory scenes. Existing methods struggle while tackling complex trajectories involving large rotations. To address this limitation, we propose CT-NeRF, an incremental reconstruction and optimization pipeline using only RGB images without pose and depth input. In this pipeline, we first propose a local-global bundle adjustment under a pose graph connecting neighboring frames to enforce the consistency between poses to escape the local minima caused by only pose consistency with the scene structure. Further, we instantiate the consistency between poses as a reprojection error constraint resulting from pixel-level correspondences between input image pairs. Through the incremental reconstruction, CT-NeRF enables the recovery of both camera poses and scene structure and is capable of handling scenes with complex trajectories. We evaluate the performance of CT-NeRF on two real-world datasets, NeRF-Buster and Free-Dataset, which feature complex trajectories. Results show CT-NeRF outperforms existing methods in novel view synthesis and pose estimation accuracy. © 1991-2012 IEEE.

关键词： 3D Reconstruction bundle adjustment NeRF novel view synthesis structure from motion