Loop Closure Error Research Articles

Keypoint-based registration of TLS point clouds using a statistical matching approach

Abstract Laser scanning is a wide-spread practice to capture the environment. Besides the fields of robotics and self-driving cars, it has been applied in the field of engineering geodesy for documentation and monitoring purposes for many years. The registration of scans is still one of the main sources of uncertainty in the final point cloud. This paper presents a new keypoint-based method for terrestrial laser scan (TLS) registration for high-accuracy applications. Based on detected 2D-keypoints, we introduce a new statistical matching approach that tests wheter keypoints, scanned from two scan stations, can be assumed to be identical. This approach avoids the use of keypoint descriptors for matching and also handles wide distances between different scanner stations. The presented approach requires a good coarse registration as initial input, which can be achieved for example by artificial laser scanning targets. By means of two evaluation data sets, we show that our keypoint-based registration leads to the smallest loop closure error when traversing several stations compared to target-based and ICP registrations. Due to the high number of observations compared to the target-based registration, the reliability of the our keypoint-based registration can be increased significantly and the precision of the registration can be increased by about 25 % on average.

Zero Velocity Detection Without Motion Pre-Classification: Uniform AI Model for All Pedestrian Motions (UMAM)

Foot-mounted positioning devices are becoming more and more popular in the different application field. For example, inertial sensors are now embedded in safety shoes to monitor security. They allow positioning with zero velocity update to bound the error growth of foot-mounted inertial sensors. High positioning accuracy depends on robust zero velocity detector (ZVD). Existing Artificial Intelligent (AI)-based methods classify the pedestrian dynamics to adjust ZVD at the cost of high computation costs and error propagation from miss-classification. We propose a machine learning model to detect zero velocity moments without any pre-classification step, named Uniform AI Model for All pedestrian Motions (UMAM). Performance is evaluated by benchmarking on two new subjects of opposite gender and different size, not included in the training data set, over complex indoor/outdoor paths of 2 km for subject 1 and 2.1 km for subject 2. We obtain an average 2D loop closure error of less than 0.37%.

Mismatched image identification using histogram of loop closure error for feature-based optical mapping

Image registration is one of the most fundamental steps in optical mapping from mobile platforms. Lately, image registration is performed by detecting salient points in two images and matching their descriptors. Robust methods [such as Random Sample Consensus (RANSAC)] are employed to eliminate outliers and compute the geometric transformation between the coordinate frames of images, typically a homography when the images contain views of a flat area. However, the image registration pipeline can sometimes provide a sufficient number of wrong inliers within the error bounds even when images do not overlap at all. Such mismatches occur especially when the scene has repetitive texture and shows structural similarity. Such pairs prevent the trajectory (thus, a mosaic) from being estimated accurately. In this paper, we propose to utilize closed-loop constraints for identifying mismatches. Cycles appear when the camera revisits an area that was imaged before, which is a common practice especially for mapping purposes. The proposed method exploits the fact that images forming a cycle should have an identity mapping when all the homographies between images in the cycle are multiplied. Our proposal obtains error statistics for each matched image pair extracting several cycle bases. Then, by using a previously trained classifier, it identifies image pairs by comparing error histograms. We present experimental results with different image sequences.

Structure from motion for ordered and unordered image sets based on random k-d forests and global pose estimation

In this paper, we present a new fast and robust method for structure from motion (SfM) for data sets potentially comprising thousands of ordered or unordered images. Our work focuses on the two most time-consuming procedures: (a) image matching and (b) pose estimation. For image matching, a new method employing a random k-d forest is proposed to quickly obtain pairs of overlapping images from an unordered set. After that, image matching and the estimation of relative orientation parameters are performed only for pairs found to be very likely to overlap. For pose estimation, we use a two-stage global approach, separating the determination of rotation matrices and translation parameters; the latter are computed simultaneously using a new method. In order to cope with outliers in the relative orientations, which global approaches are particularly sensitive to, we present a new constraint based on triplet loop closure errors of rotation and translation. Finally, a robust bundle adjustment is carried out to refine the image orientation parameters.We demonstrate the potential and limitations of our pipeline using various real-world datasets including ordered image data acquired from UAV (unmanned aerial vehicle) and other platforms as well as unordered data from the internet. The experiments show that our work performs better than comparable state-of-the-art SfM systems in terms of run time, while we achieve a similar accuracy and robustness.

A two-level optimized graph-based simultaneous localization and mapping algorithm

PurposeBecause submaps including a subset of the global map contain more environmental information, submap-based graph simultaneous localization and mapping (SLAM) has been studied by many researchers. In most of those studies, helpful environmental information was not taken into consideration when designed the termination criterion of the submap construction process. After optimizing the graph, cumulative error within the submaps was also ignored. To address those problems, this paper aims to propose a two-level optimized graph-based SLAM algorithm.Design/methodology/approachSubmaps are updated by extended Kalman filter SLAM while no geometric-shaped landmark models are needed; raw laser scans are treated as landmarks. A more reasonable criterion called the uncertainty index is proposed to combine with the size of the submap to terminate the submap construction process. After a submap is completed and a loop closure is found, a two-level optimization process is performed to minimize the loop closure error and the accumulated error within the submaps.FindingsSimulation and experimental results indicate that the estimated error of the proposed algorithm is small, and the maps generated are consistent whether in global or local.Practical implicationsThe proposed method is robust to sparse pedestrians and can be adapted to most indoor environments.Originality/valueIn this paper, a two-level optimized graph-based SLAM algorithm is proposed.

Efficient Velodyne SLAM with point and plane features

This paper develops and tests a plane based simultaneous localization and mapping algorithm capable of processing the uneven sampling density of Velodyne-style scanning LiDAR sensors in real-time. The algorithm uses an efficient plane detector to rapidly provide stable features, both for localization and as landmarks in a graph-based SLAM. When planes cannot be detected or when they provide insufficient support for localization, a novel constraint tracking algorithm selects a minimal set of supplemental point features to be provided to the localization solver. Several difficult indoor and outdoor datasets, totaling 6981 scans, each with $$\sim $$ 70,000 points, are used to analyze the performance of the algorithm without the aid of any additional sensors. The results are compared to two competing state-of-the-art algorithms, GICP and LOAM, showing up to an order of magnitude faster runtime and superior accuracy on all datasets, with loop closure errors of 0.14–0.95 m, compared to 0.44–66.11 m.

A brain-inspired SLAM system based on ORB features

This paper describes a brain-inspired simultaneous localization and mapping (SLAM) system using oriented features from accelerated segment test and rotated binary robust independent elementary (ORB) features of RGB (red, green, blue) sensor for a mobile robot. The core SLAM system, dubbed RatSLAM, can construct a cognitive map using information of raw odometry and visual scenes in the path traveled. Different from existing RatSLAM system which only uses a simple vector to represent features of visual image, in this paper, we employ an efficient and very fast descriptor method, called ORB, to extract features from RGB images. Experiments show that these features are suitable to recognize the sequences of familiar visual scenes. Thus, while loop closure errors are detected, the descriptive features will help to modify the pose estimation by driving loop closure and localization in a map correction algorithm. Efficiency and robustness of our method are also demonstrated by comparing with different visual processing algorithms.

Polar Metric-Weighted Norm-Based Scan Matching for Robot Pose Estimation

A novel point-to-point scan matching approach is proposed to address pose estimation and map building issues of mobile robots. Polar Scan Matching (PSM) and Metric-Based Iterative Closest Point (Mb-ICP) are usually employed for point-to-point scan matching tasks. However, due to the facts that PSM considers the distribution similarity of polar radii in irrelevant region of reference and current scans and Mb-ICP assumes a constant weight in the norm about rotation angle, they may lead to a mismatching of the reference and current scan in real-world scenarios. In order to obtain better match results and accurate estimation of the robot pose, we introduce a new metric rule, Polar Metric-Weighted Norm (PMWN), which takes both rotation and translation into account to match the reference and current scan. For robot pose estimation, the heading rotation angle is estimated by correspondences establishing results and further corrected by an absolute-value function, and then the geometric property of PMWN called projected circle is used to estimate the robot translation. The extensive experiments are conducted to evaluate the performance of PMWN-based approach. The results show that the proposed approach outperforms PSM and Mb-ICP in terms of accuracy, efficiency, and loop closure error of mapping.

Minimization of self-potential survey mis-ties acquired with multiple reference locations

Self-potential (SP) surveys often involve many interconnected lines of data along available roads or trails, with the ultimate goal of producing a unique map of electric potentials at each station relative to a single reference point. Multiple survey lines can be tied together by collecting data along intersecting transects and enforcing Kirchhoff’s voltage law, which requires that the total potential drop around any closed loop equals zero. In practice, however, there is often a nonzero loop-closure error caused by noisy data; traditional SP processing methods redistribute this error evenly over the measurements that form each loop. The task of distributing errors and tying lines together becomes nontrivial when many lines of data form multiple interconnected loops because the loop-closure errors are not independent, and a unique potential field cannot be determined by processing lines sequentially. We present a survey-consistent processing method that produces a unique potential field by minimizing the loop-closure errors over all lines of data simultaneously. When there are no interconnected survey loops, the method is equivalent to traditional processing schemes. The task of computing the potential field is posed as a linear inverse problem, which easily incorporates prior information about measurement errors and model constraints. We investigate the use of both [Formula: see text] and [Formula: see text] measures of data misfit, the latter requiring an iterative-solution method with increased computational cost. The [Formula: see text] method produces more reliable results when outliers are present in the data, and is similar to the [Formula: see text] result when only Gaussian noise is present. Two synthetic examples are used to illustrate this methodology, which is subsequently applied to a field data set collected as part of a geothermal exploration campaign in Nevis, West Indies.

Optimal Stochastic Design of Four-Bar Mechanisms for Tolerance and Clearance

A formulation of optimal stochastic design of mechanisms considering tolerances on the link lengths and clearances in the joints is presented. The mechanical error is analyzed for the 3σ band of confidence level through a new stochastic model of a four-bar linkage. A multiobjective optimization problem with two objective functions, mechanical error and manufacturing cost, is considered by a weighting method. Tolerance widths and nominal clearance sizes are taken as design variables. The conflicting nature between the two objectives is represented in the form of tradeoff curves. A concept of loop closure error is also introduced and applied to optimum design of the nominal linkage system and the stochastic linkage system. Numerical results are given for a four-bar function generator and a four-bar path generator.