ELY-SLAM: An efficient SLAM algorithm with lightweight YOLO-based collaborative perception

In this paper, an improved algorithm for 3D SLAM is presented. A data dimensional reduction algorithm—Principal Component Analysis (PCA) is adopted, so as to speed up the rate of feature extraction from RGB-D images(640×480). To overcome the poor efficiency performance on the original RGB-D SLAM, a novel memory management algorithm and an incremental appearance-based loop closure detector are used to realize the 3D mapping and robot's trajectory estimation. In the back-end, the robot's trajectory and global map is optimized by g2o. The experiments demonstrate that the proposed method is faster and more effective than the original RGB-D SLAM.

An efficient and accurate visual SLAM loop closure detection algorithm

Simultaneous localization and map building (SLAM) is a fundamental and complex problem in mobile robot research. In SLAM, Kalman-filter-like implementations are widely adopted to localize a mobile robot and build a map simultaneously and incrementally. However, this approach requires extensive computations of order O(N/sup 2/), where N is the total number of landmarks. To make the computations more manageable, we propose a logarithmic map partitioning algorithm that partitions the global map into one local region and several sub-maps. The size of each sub-map is based on its distance from the mobile robot, and in each sub-map, a centroid landmark is selected to represent all the landmarks in the sub-map for SLAM computations. With this logarithmic-map partitioning, it maintains correlation updates with each sub-map and provides an efficient suboptimal solution to the SLAM problem. The number of landmarks reduces from N to a logarithm-based function of N, and the computational requirement reduces from O(N/sup 3/) to O(N/sup 2/), where N/sub L/ is the number of local landmarks. Furthermore, utilizing the compressed extended Kalman filter, the real-time computational complexity reduces to O(N/sub L//sup 2/). Computer simulation results showed that the proposed algorithm is consistent and efficient for a large number of landmarks.

In this paper, we propose a scheme of generating a hybrid visual map for SLAM. The hybrid visual map has two levels of map representations: 1) the absolute map representation of highly distinctive visual planes via EKFSLAM and 2) the relative map representation of dense visual features for each visual plane via sparse information filter update. The absolute map can maintain its global consistency by matching the visual plane, the group of visual features. It improves data association and reduces the number of landmarks against individual visual features. Moreover, the relative map can reconstruct a 3-D map of visual features efficiently without loosing dense visual information. The performance of the proposed method was verified by the experimental results of consistent hybrid visual maps in an indoor environment.

This article presents a very efficient SLAM algorithm that works by hierarchically dividing a map into local regions and subregions. At each level of the hierarchy each region stores a matrix representing some of the landmarks contained in this region. To keep those matrices small, only those landmarks are represented that are observable from outside the region. A measurement is integrated into a local subregion using O(k2) computation time for k landmarks in a subregion. When the robot moves to a different subregion a full least-square estimate for that region is computed in only O(k3 log n) computation time for n landmarks. A global least square estimate needs O(kn) computation time with a very small constant (12.37 ms for n = 11300). The algorithm is evaluated for map quality, storage space and computation time using simulated and real experiments in an office environment.

The data loss caused by unreliable network seriously impacts the results of remote visual SLAM systems. From our experiment, a loss of less than 1 second of data can cause a visual SLAM algorithm to lose tracking. We present a novel buffering method, ORBBuf, to reduce the impact of data loss on remote visual SLAM systems. We model the buffering problem as an optimization problem by introducing a similarity metric between frames. To solve the buffering problem, we present an efficient greedy-like algorithm to discard the frames that have the least impact on the quality of SLAM results. We implement our ORBBuf method on ROS, a widely used middleware framework. Through an extensive evaluation on real-world scenarios and tens of gigabytes of datasets, we demonstrate that our ORBBuf method can be applied to different state-estimation algorithms (DSO and VINS-Fusion), different sensor data (both monocular images and stereo images), different scenes (both indoor and outdoor), and different network environments (both WiFi networks and 4G networks). Our experimental results indicate that the network losses indeed affect the SLAM results, and our ORBBuf method can reduce the RMSE up to 50 times comparing with the Drop-Oldest and Random buffering methods.

During the continual transitions from lab research to real-world applications of vision-based algorithms, there are significant challenges, e.g. the robustness to adapt to complex environments and the high demands of multi-task, multi-modal learning models. More specifically on visual simultaneous localization and mapping (visual SLAM) for mobile robots, two notable limitations are: (i) the drift issue during pose estimation, especially in dynamic environments, makes the positioning system unstable; (ii) the additional loop detection takes many computational resources for image retrieval and global feature geometry checking. This thesis explores the approaches of visual localization on unmanned ground vehicles by solving those limitations. The conventional feature extraction and matching pipeline for vision-based tasks involve three steps: (i) local feature detection and description, (ii) feature matching, and (iii) outlier rejection. The first contribution of the thesis is adding a feature selection and anticipation stage to reduce tracking drift. We explore the noise model of image features to select a subset of all the observed image features with the best ”contribution” during data association and pose estimation across multiple frames. Conventional SLAM algorithms take a strong assumption of scene rigidity, which limits the application under challenging environments. The second part of the thesis addresses the tough issue in dynamic environments with moving objects. We presented GMC, namely the motion clustering approach, a lightweight dynamic object filtering method. It can distinguish moving objects from static landmarks. Based on the theory of motion coherence within a particular image area, GMC could segment dynamic objects in 3D space. We can provide an efficient and robust correspondence algorithm that can extract dynamic objects from a static background with the method. In this way, we propose a dynamic SLAM system that is real-time and free from expensive GPU processors. In contrast to GMC, the thesis’s third part turns to an object-aware learning-based model for more general dynamic scenarios. We use object detection and tracking as points, lines, planes, etc. We utilize semantic information and extract sparse image features simultaneously to keep track of dynamic objects. The static background and different dynamic objects are jointly optimized in a newly developed bundle adjustment sliding window. The estimated 3D bounding boxes can provide more robust camera tracking and better scene understanding, and better map merging. The fourth part of the thesis leverages the emerging feature learning framework. It proposes a unified self-supervised model called LGDNet to generate both global and local image feature descriptors end-to-end. Global feature descriptors embed the whole image into a compact representation, leading to easier scene comparison. On the other hand, local features focus more on the local region similarities of some salient parts for structure from motion. Our proposed method can directly extract features together with descriptors that encode both local maximum responses and global context information, avoiding duplicate calculations based on different feature extraction criteria.

We introduce a method for efficiently rasterizing large occupancy grids. Efficient Maximum Likelihood Estimation (MLE) of robot trajectories has been shown to be highly scalable using sparse SLAM algorithms such as SqrtSAM, but unfortunately such approaches don't directly provide a rasterized grid map. We harness these existing SLAM methods to compute maximum likelihood (ML) robot trajectories and introduce a new efficient algorithm to rasterize a dynamic occupancy grid. We propose a spatially-aware data structure that enables the cost of a map update to be proportional to the impact of any loop closures, resulting in better average case performance than naive methods. Furthermore, we show how redundant sensor data can be exploited to improve map quality and speed up rasterization. We evaluate our method using several data sets collected using a team of 14 autonomous robots and show success in mixed indoor-outdoor urban environments as large as 220m × 170m, with 0.1m resolution.