SLAM systems are proposed to estimate mobile robot’ poses and reconstruct maps of surrounding environments. Most SLAM systems assume that their working environments are static. However, there are many dynamic objects in actual environments, which reduce the accuracy and robustness of the SLAM system. In this paper, a real-time robust visual SLAM (VSLAM) system working in dynamic environments named COEB-SLAM is proposed, which combines object detection, epipolar geometry constraint, and blur filtering to remove dynamic feature points. In reality, shooting fast-moving objects can cause image motion blur, which affects the dynamic SLAM’s ability to determine the object’s state. Therefore, we attempted to use blurring level as a constraint to identify dynamic objects. Moreover, the removal of a large number of feature points can lead to the failure of tracking. Thus, we use a feature points redistribution method for retaining enough high-quality static features to reduce this impact. Extensive evaluations in the public TUM RGB-D datasets and the real world dynamic environments demonstrate that COEB-SLAM has robustness in dynamic environments. Compared to ORB-SLAM2 system, COEB-SLAM reduces the average absolute trajectory error by about 90% in TUM RGB-D datasets. For the benefit of the community, we make public the source code at:https://github.com/biscuitzb/COEB-SLAM.

Recognition and pose estimation from 3D free-form objects is a key step for autonomous robotic manipulation. Recently, the point pair features (PPF) voting approach has been shown to be effective for simultaneous object recognition and pose estimation. However, the global model descriptor (e.g., PPF and its variants) that contained some unnecessary point pair features decreases the recognition performance and increases computational efficiency. To address this issue, in this paper, we introduce a novel strategy for building a global model descriptor using stably observed point pairs. The stably observed point pairs are calculated from the partial view point clouds which are rendered by the virtual camera from various viewpoints. The global model descriptor is extracted from the stably observed point pairs and then stored in a hash table. Experiments on several datasets show that our proposed method reduces redundant point pair features and achieves better compromise of speed vs accuracy.

In scenes where there are lighting changes, localization may fail for visual SLAM due to feature point tracking failure. Thus, a feature point tracking method based on multi-condition constraints is proposed for visual SLAM. The proposed method tracks the feature points of optical flow from aspects such as the overall motion position of feature points, descriptor grayscale information, and spatial geometric constraints. First, to solve the problem of feature point mismatch in complex environments, we propose a feature point mismatch removal method that combines optical flow, descriptor, and RANSAC. We eliminate incorrect feature point matches layer by layer through these constraints. The uniformity of feature point distribution in the image can then affect the accuracy of camera pose estimation, and different scenes can also affect the difficulty of feature point extraction. In order to balance the quality and uniformity of the extracted feature points, we propose an adaptive mask homogenization method that adaptively adjusts the mask radius according to the quality of feature points. Experiments conducted on the EuRoC dataset show that the proposed method which integrates the improved feature point mismatch removal method and mask homogenization method into feature point tracking, exhibits robustness and accuracy under various interferences such as lighting changes, image blurring, and unclear textures. Compared to the RANSAC method, we reduce the location error by about 85% using the EuRoC dataset.

The future smart factory necessitates the identification of robot collisions, including the materials involved, the specific robot component impacted, and whether a human is involved. This article introduces a method for classifying robot collisions based on motor current and robot link vibration. First, it analyzes vibration modal features that can reveal the collision’s location and the contacting material. Thereafter, it employs variational mode decomposition (VMD) to analyze the vibration characteristics of collisions, with optimization of the decomposition parameter. To address computational complexity, an equivalent filter bank is proposed for extracting collision features instead of VMD. These vibration features are then utilized to construct and train a neural network for collision detection and classification. Finally, the proposed method is verified on a robot platform, successfully distinguishing and detecting collisions involving four different materials: human hand, cotton hammer, rubber hammer, and steel hammer. The accuracy of detection is evaluated using support vector machine (SVM), back propagation (BP), radial basis function (RBF), and convolutional neural network (CNN) through comprehensive experiments. The primary contribution of this study lies in the accurate differentiation between human-involved and nonhuman-involved collisions, facilitating real-time robot collision monitoring amid varying loads and offering insights into the collision’s location. Experimental results demonstrate a 90% accuracy in human collision detection and a 95% accuracy in collision link detection. Experiment data and code for training and testing of neural networks are available at https://github.com/ldepn/Robot-collision-VMD.git .

For the robotic grasping of randomly stacked objects in a cluttered environment, the active multiple viewpoints method can improve grasping performance by improving the environment perception ability. However, in many scenes, it is redundant to always use multiple viewpoints for grasping detection, which will reduce the robot’s grasping efficiency. To improve the robot’s grasping performance, we present a Viewpoint Adjusting and Grasping Synergy (VAGS) strategy based on deep reinforcement learning which coordinates the viewpoint adjusting and grasping directly. For the training efficiency of VAGS, we propose a Dynamic Action Exploration Space (DAES) method based on $\epsilon$-greedy to reduce the training time. To address the sparse reward problem in reinforcement learning, a reward function is created to evaluate the impact of adjusting the camera pose on the grasping performance. According to experimental findings in simulation and the real world, the VAGS method can improve grasping success and scene clearing rate. Compared with only direct grasping, our proposed strategy increases the grasping success rate and the scene clearing rate by 10.49% and 11%.

Low-light image enhancement is one of the challenges in robotics. For low-light images, only adjusting the brightness of the image will inevitably amplify the noise and cause more severe color distortion, which is contrary to the original purpose. To balance multiple factors including brightness, color, and noise, we propose a simple and efficient lighting network of Multi-scale Residual and Attention Guidance (MRAG) to enhance low-light images.The MRAG network model is founded on Retinex theory and includes a decomposition layer that separates low-light images into reflectance and illumination maps, as well as a reflectance recovery layer and an illumination recovery layer. In the decomposition layer, an end-to-end attention mechanism based on multi-branch convolutional neural networks is adopted, and in the reflectance layer, a multi-scale residual network is used to fully learn image features. Extensive experiments on the LIME and NPE datasets show that the proposed method outperforms state-of-the-art methods in terms of image quality evaluator, with respective improvements of 3.9% and 8.9%. The experiments on the LOL dataset demonstrate that the proposed method has enhancements in the PSNR, SSIM, LOE, and NIQE evaluation metrics compared to state-of-the-art. Compared with state-of-the-art low-light enhancement methods, the proposed method has some improvements in optimizing the SLAM front end.