Lei Wang

Lei Wang

Auto-generated profile, based on publicly available information

Lei Wang

Faculty of Robot Science and Engineering, Northeastern University, Shenyang, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 May 2024 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
LZLe Zhou
Le Zhou
YZYunzhou Zhang
Yunzhou Zhang
LWLei Wang
Lei Wang
JZJinpeng Zhang
Jinpeng Zhang
For those systems consisted of camera and LiDAR, extrinsic calibration is unavoidable. Traditional methods usually start from a known initial pose and alternately optimize the extrinsic parameters of these sensors. Their main limitations are that a good initial guess is required and matching features are inevitably corrupted by outliers, which affect the accuracy of calibration significantly. In this paper, we present a novel approach to estimate the extrinsic parameters between the LiDAR and the camera. We first obtain the initial transformations by leveraging the structural regularity such as the 2D-3D dominant directions and 2D-3D parallel line pairs of the structured scenes, then achieve 2D-3D line registration using the nearest neighbor approach. Furthermore, to handle outliers, we propose an effective line feature matching approach, termed as local triangular preserving matching (LTP), the major objective of which is to maintain the triangular topological consistency among neighboring vector arcs generated by inliers. We formulate it into a mathematical model, and derive a simple closed-form solution. Then, the calibration parameters are estimated by using correspondences from LTP and further refined by the Levenberg-Marquardt algorithm (LM) with outlier removal strategy. Experimental results show that our method can achieve better performance over state-of-the-art methods.

15 June 2023 in IEEE Transactions on Cognitive and Developmental Systems
Verified
Journal Article
Closed access
LWLei Wang
Lei Wang
YZYunzhou Zhang
Yunzhou Zhang
DZDelong Zhu
Delong Zhu
SCSonya Coleman
Sonya Coleman
DKDermot Kerr
Dermot Kerr
Learning from small amounts of samples with reinforcement learning (RL) is challenging in many tasks, especially in real-world applications, such as robotics. Meta-Reinforcement Learning (meta-RL) has been proposed as an approach to address this problem by generalizing to new tasks through experience from previous similar tasks. However, these approaches generally perform meta-optimization by focusing direct policy search methods on validation samples from adapted policies, thus requiring large amounts of on-policy samples during meta-training. To this end, we propose a novel algorithm called Supervised Meta-Reinforcement Learning with Trajectory Optimization (SMRL-TO) by integrating Model-Agnostic Meta-Learning (MAML) and iLQR-based trajectory optimization. Our approach is designed to provide online supervision for validation samples through iLQR-based trajectory optimization and embed simple imitation learning into the meta-optimization rather than policy gradient steps. This is actually a bi-level optimization that needs to calculate several gradient updates in each meta-iteration, consisting of off-policy reinforcement learning in the inner loop and online imitation learning in the outer loop. SMRL-TO can achieve significant improvements in sample efficiency without human-provided demonstrations, due to the effective supervision from iLQR-based trajectory optimization. In this paper, we describe how to use iLQR-based trajectory optimization to obtain labeled data and then how leverage them to assist the training of meta-learner. Through a series of robotic manipulation tasks, we further show that compared with the previous methods, the proposed approach can substantially improve sample efficiency and achieve better asymptotic performance.

13 March 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
LWLei Wang
Lei Wang
YZYunzhou Zhang
Yunzhou Zhang
SZShangdong Zhu
Shangdong Zhu
HCHe Cao
He Cao
LZLe Zhou
Le Zhou
YLYixiu Liu
Yixiu Liu
6-DoF grasp pose detection in clutter is an essential but challenging task in the vision measurement. However, most current detection methods often fail to consider the geometric information and partial occlusions of objects in clutter, resulting in inaccurate robotic grasping and potential collisions. In this work, we propose an end-to-end grasping system to directly generate 6-DoF grasp poses from point clouds in clutter. Our system first selects the most feasible positions as the grasp centers, and then determines their grasp parameters, such as approaching vectors and in-plane rotations. Specifically, we design an optimality criterion based on the geometric information of objects to extract rich object features from the scene and identify high-graspablity areas. To alleviate the collision problem caused by occlusions, we focus on the spatial relations among objects and explicitly train a collision predictor to directly select collision-free points as grasp centers from high-graspablity areas. Based on these grasp centers, our model can easily generate diverse sets of collision- free and robust grasp poses for successful grasping. To efficiently determine grasp parameters, we exploit neighbor search to capture the local context around each grasp center and then implicitly learn relations among objects using the self-attention mechanism. Compared with previous methods, our grasping system improves performance by a large margin (>10%), and achieves state-of-the-art results on GraspNet-1Billion, a largescale dataset. Besides, our method has a high success rate for grasping in simulation and real-world experiments, verifying the validity in clutter.

23 October 2024 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
HCHe Cao
He Cao
YZYunzhou Zhang
Yunzhou Zhang
SZShangdong Zhu
Shangdong Zhu
LWLei Wang
Lei Wang
In unstructured environments, robots are likely to encounter desktop objects that they have never seen before. Classifying these objects precisely is a prerequisite for accomplishing object-specific manipulation tasks. However, it is time-consuming to collect large-scale object classification datasets. Inspired by the prompt tuning methods, we propose the PRO-CLIP network, which is a category measurement method for desktop objects. Specifically, PRO-CLIP performs few-shot classification based on the knowledge from pretrained vision-language model. It utilizes token-level and prompt-level optimal transportations to jointly fine-tune the learnable vision-language prompts. For token-level stage, we propose the image patch reweighting mechanism to make alignments focus on the image patches that are close to the patch prototypes. This allows the patch embeddings have converging category representations, which reduces intra-class differences of visual features. For prompt-level stage, we propose a cascading optimal transportation module to simultaneously consider task-irrelevant knowledge in zero-shot features and task-specific knowledge in few-shot features. Due to the generalization performance of task-irrelevant knowledge, the proposed module achieves feature regularization during optimal transportation. Finally, we propose the UP loss to supervise the whole network. It contains unbalanced logit-level consistency losses and visual prototype loss. The logit-level consistency losses are used to make learnable features close to zero-shot features. The prototype loss makes the visual features approach to the corresponding prototypes in distance. We demonstrate the effectiveness of our method by performing few-shot classification experiments on different datasets including desktop objects. The relevant code will be available at https://github.com/NeuCV-IRMI/proclip.

15 April 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
LRLei Rong
Lei Rong
YZYunzhou Zhang
Yunzhou Zhang
JZJinpeng Zhang
Jinpeng Zhang
LWLei Wang
Lei Wang
Illumination variations and rapid motion significantly impair the performance of visual inertial navigation systems (VINS). Existing VINS algorithms struggle to achieve efficient and precise localization under conditions where both illumination variations and rapid motion occur simultaneously. To address this challenge, we propose an illumination-robust image interval-guided adaptive multifarious hybrid visual inertial odometry (IRIH-VIO). Our approach begins with a metric-guided, multi-stage adaptive iterative image enhancement algorithm that processes environmental images with fluctuating lighting into a sequence of images with consistent and normalized lighting in real-time, using only the CPU. The enhanced image sequence is then segmented into various intervals based on the contrast distribution of the original images. We subsequently perform adaptive switching and fusion of filtering and optimization in the outcomes of each interval. We introduce covariance inflation in the filtering stage to enhance sensitivity and resilience to rapid movements. Additionally, we developed a visual information weighting technique for feature extraction in the optimization process and incorporated it into the hybrid marginalization process. Experimental results from public datasets and real-world scenarios demonstrate that IRIH-VIO achieves superior performance in terms of accuracy and computational efficiency compared to state-of-the-art methods.