Pose Estimation Error Research Articles

2A1-T03 ステレオビジョンを備えたロボットヘッドのリアルタイム位置姿勢推定

We describe a location and pose estimation of a robotic head based on three dimensional coordinate of three point markers. Computation for the location and pose estimation is carried out at 30 fps, which is a frame rate of camera, and with low latency by using FPGA. In experiments for accuracy evaluation, an error in pose estimation was less than 3 degrees. This can be used for a pose estimation of humanoid head by combining complementary other methods such as one with inertial sensors.

A comparison of mobile robot pose estimation using nonlinear filters: simulation and experimental results

This paper explores and compares the nature of the nonlinear filtering techniques on mobile robot pose estimation. Three nonlinear filters are implemented including the extended Kalman filter (EKF), the unscented Kalman filter (UKF) and the particle filter (PF). The criteria of comparison is the magnitude of the error of pose estimation, the computational time, and the robustness of each filter to noise. The filters are applied to two applications including the pose estimation of a two-wheeled robot in an experimental platform and the pose estimation of a three-wheeled robot in a simulated environment. The robots both in the experimental and simulated platform move along a nonlinear trajectory like a circular arc or a spiral. The performance of their pose estimation are compared and analysed in this paper.

Robust Head Pose Estimation Using a 3D Morphable Model

Head pose estimation from single 2D images has been considered as an important and challenging research task in computer vision. This paper presents a novel head pose estimation method which utilizes the shape model of the Basel face model and five fiducial points in faces. It adjusts shape deformation according to Laplace distribution to afford the shape variation across different persons. A new matching method based on PSO (particle swarm optimization) algorithm is applied both to reduce the time cost of shape reconstruction and to achieve higher accuracy than traditional optimization methods. In order to objectively evaluate accuracy, we proposed a new way to compute the pose estimation errors. Experiments on the BFM-synthetic database, the BU-3DFE database, the CUbiC FacePix database, the CMU PIE face database, and the CAS-PEAL-R1 database show that the proposed method is robust, accurate, and computationally efficient.

Asynchronous vehicle pose correction using visual detection of ground features

The inherent noise associated with odometry manifests itself as errors in localization for autonomous vehicles. Visual odometry has been previously used in order to supplement classical vehicle odometry. However, visual odometry is limited in its ability to reduce errors in localization for large travel distances that entail the cumulative summing of individual frame-to-frame image errors. In this paper, a novel machine vision approach for tiled surfaces is proposed to address this problem. Tile edges in a laboratory environment are used to define a travel trajectory for the Quansar Qbot (autonomous vehicle) built on the iRobot iRoomba platform with a forward facing camera. Tile intersections are used to enable asynchronous error recovery for vehicle position and orientation. The proposed approach employs real-time image classification and is feasible for error mitigation for large travel distances.The average position error for an 8m travel distance using classical odometry was measured to be 0.28m. However, implementation of the proposed approach resulted in an error of 0.028m. The proposed approach therefore significantly reduces pose estimation error and could be used to supplement existing modalities such as GPS and Laser-based range sensors.

Dynamic EKF-based SLAM for autonomous mobile convergence platforms

This paper presents a new Simultaneous Localization and Mapping (SLAM) framework for solving the problem of SLAM in dynamic environments. The landmark location change causes the error of robot pose estimation and landmark mapping. In this paper, we propose the Dynamic Extended Kalman Filter (EKF) SLAM based on the independence of the dynamic landmarks. The proposed framework decomposes the SLAM problem into a traditional SLAM problem for the static landmarks and individual SLAM problems for the dynamic landmarks. Therefore, in the dynamic environments, it is able to minimize the error caused by the dynamic landmarks and reduce the uncertainty in the robot pose and the landmark locations. In order to validate the proposed approach, we implement an indoor mobile robot platform with a Red Green Blue - Depth (RGB-D) sensor and utilize Speeded Up Robust Features (SURF) algorithm to extract appearance-based features. The simulation and experimental results show the validity and robustness of the Dynamic EKF SLAM in indoor environments including the dynamic landmarks.

Simulation of Registration Accuracy of Iterative Closest Point (ICP) Method for Pose Estimation

ICP method is a powerful tool widely used in 3D point registration and pose estimation application. But there lacks an deterministic relation between the 3D point dataset and the model. Here we present some Monte Carlo simulation results for pose estimation error of one degree and the system requirement for lidar 3D point measurement techniques.

Structure and motion in urban environments using upright panoramas

Image-based modeling of urban environments is a key component of enabling outdoor, vision-based augmented reality applications. The images used for modeling may come from off-line efforts, or online user contributions. Panoramas have been used extensively in mapping cities and can be captured quickly by an end-user with a mobile phone. In this paper, we describe and evaluate a reconstruction pipeline for upright panoramas taken in an urban environment. We first describe how panoramas can be aligned to a common vertical orientation using vertical vanishing point detection, which we show to be robust for a range of inputs. The orientation sensors in modern cameras can also be used to correct the vertical orientation. Secondly, we introduce a pose estimation algorithm, which uses knowledge of a common vertical orientation as a simplifying constraint. This procedure is shown to reduce pose estimation error in comparison with the state of the art. Finally, we evaluate our reconstruction pipeline with several real-world examples.

Filtering Method for Pose Angles Based on CS-STF-IMM

In order to decrease the error of pose estimation based on image sequences, a filtering method for pose angles based on CS-STF-IMM was proposed. Firstly, current statistical (CS) model was used to establish state space model of pose transformation, which reflects the maneuver of target pose more truly. Then, strong tracking filter (STF) was introduced to overcome the shortcoming that Kalman filter can hardly track sudden maneuver. A parameters determination method based on least square fitting was proposed, which improves the performance of STF further. Based on these, CS-STF-IMM algorithm was established, which obviously improved the precision of pose estimation. Simulations showed that the proposed method has better overall performance, which is a feasible scheme for pose angles filtering.

Filtering Method for Pose Angles Based on CS-STF-IMM

In order to decrease the error of pose estimation based on image sequences, a filtering method for pose angles based on CS-STF-IMM was proposed. Firstly, current statistical (CS) model was used to establish state space model of pose transformation, which reflects the maneuver of target pose more truly. Then, strong tracking filter (STF) was introduced to overcome the shortcoming that Kalman filter can hardly track sudden maneuver. A parameters determination method based on least square fitting was proposed, which improves the performance of STF further. Based on these, CS-STF-IMM algorithm was established, which obviously improved the precision of pose estimation. Simulations showed that the proposed method has better overall performance, which is a feasible scheme for pose angles filtering.

Automatic Target Recognition of SAR Images Based on Global Scattering Center Model

This paper proposes a synthetic aperture radar (SAR) automatic target recognition approach based on a global scattering center model. The scattering center model is established offline using range profiles at multiple viewing angles, so the original data amount is much less than that required for establishing SAR image templates. Scattering center features at different target poses can be conveniently predicted by this model. Moreover, the model can be modified to predict features for various target configurations. For the SAR image to be classified, regional features in different levels are extracted by thresholding and morphological operations. The regional features will be matched to the predicted scattering center features of different targets to arrive at a decision. This region-to-point matching is much easier to implement and is less sensitive to nonideal factors such as noise and pose estimation error than point-to-point matching. A matching scheme going through from coarse to fine regional features in the inner cycle and going through different pose hypotheses in the outer cycle is designed to improve the efficiency and robustness of the classifier. Experiments using both data predicted by a high-frequency electromagnetic (EM) code and data measured in the MSTAR program verify the validity of the method.

An Inter-Landmark Approach to 4-D Shape Extraction and Interpretation: Application to Myocardial Motion Assessment in MRI

This paper presents a novel approach to shape extraction and interpretation in 4-D cardiac magnetic resonance imaging data. Statistical modeling of spatiotemporal interlandmark relationships is performed to enable the decomposition of global shape constraints and subsequently of the image analysis tasks. The introduced descriptors furthermore provide invariance to similarity transformations and thus eliminate pose estimation errors in the presence of image artifacts or geometrical inconsistencies. A set of algorithms are derived to address key technical issues related to constrained boundary tracking, dynamic model relaxation, automatic initialization, and dysfunction localization. The proposed framework is validated with a relatively large dataset of 50 subjects and compared to existing statistical shape modeling methods. The results indicate increased adaptation to spatiotemporal variations and imaging conditions.

Kinematic parameter calibration of a car-like mobile robot to improve odometry accuracy

Odometry provides fundamental pose estimates for wheeled vehicles. For accurate and reliable pose estimation, systematic and nonsystematic errors of odometry should be reduced. In this paper, we focus on systematic error sources of a car-like mobile robot (CLMR) and we suggest a novel calibration method. Kinematic parameters of the CLMR can be successfully calibrated by only a couple of test driving. After reducing deterministic errors by calibration, odometry accuracy can be further improved by redundant odometry fusion with the extended Kalman filter (EKF). Odometry fusion reduces nonsystematic or stochastic errors. Experimental verifications are carried out using a radio-controlled miniature car.

A particle filter-based approach for tracking undersea narrow telecommunication cables

The surveillance and inspection of underwater installations such as telecommunication cables are currently carried out by trained operators who, from the surface, guide a remotely operated vehicle (ROV) with cameras mounted over it. This manual visual control is, however, a very tedious job that tends to fail if the operator loses concentration. This paper describes a tracking system for underwater narrow telecommunication cables, the main objective of which is to allow an autonomous underwater vehicle to video-document the whole length of a cable. The approach is based on particle filters (PFs) because of their natural ability to model multi-dimensional multi-modal PDFs, which allows handling in a more appropriate way the ambiguities that naturally arise from undersea environments. In effect, despite the special visual features that artificial objects present, which allow distinguishing them in natural scenarios such as the seabed, distracting background such as rocks or algae growing on top and nearby cables, complicate the detection and tracking and give rise to ambiguities when rocks or marine growth form shapes and textures that resemble the cable. Apart from the different models that a PF requires, the paper also describes a set of added features, which successfully compensate some large errors in the cable pose estimation when the non-enhanced tracker is applied. Extensive experimental results over a test set of more than 10,000 frames, for which a ground truth has been manually generated, have shown the usefulness of the solution proposed. Besides, results for a set of six video sequences accounting for almost 150,000 frames and around one hour and a half of successful continuous video tracking are also discussed. All those images come from inspection runs captured by ROVs navigating over real telecommunication undersea cables.

Robustness analysis of pose estimation based on points

From the perspective of practical application,the robustness of pose estimation with coordinate detection error was discussed,given that P3P issue has exclusive solution.With the theoretical derivation on simplified condition and the simulation on normal condition,some conclusions have been obtained such as the pose estimation error is decreases with increment of camera focus ratio,is independent to the coordinate of image center point,decreases with the increment of distance between control points in object model,increases with the increment of estimation distance,etc.All these conclusions are significant to the design of pose estimation system.

Integrating perceptual level of detail with head-pose estimation and its uncertainty

Immersive virtual environments with life-like interaction capabilities can provide a high fidelity view of the virtual world and seamless interaction methods to the user. High demanding requirements, however, raise many challenges in the development of sensing technologies and display systems. The focus of this study is on improving the performance of human–computer interaction by rendering optimizations guided by head pose estimates and their uncertainties. This work is part of a larger study currently being under investigation at NASA Ames, called “Virtual GloveboX” (VGX). VGX is a virtual simulator that aims to provide advanced training and simulation capabilities for astronauts to perform precise biological experiments in a glovebox aboard the International Space Station (ISS). Our objective is to enhance the virtual experience by incorporating information about the user’s viewing direction into the rendering process. In our system, viewing direction is approximated by estimating head orientation using markers placed on a pair of polarized eye-glasses. Using eye-glasses does not pose any constraints in our operational environment since they are an integral part of a stereo display used in VGX. During rendering, perceptual level of detail methods are coupled with head-pose estimation to improve the visual experience. A key contribution of our work is incorporating head pose estimation uncertainties into the level of detail computations to account for head pose estimation errors. Subject tests designed to quantify user satisfaction under different modes of operation indicate that incorporating uncertainty information during rendering improves the visual experience of the user.

Optimization Framework for a Multiple Classifier System with Non-Registered Targets

A fundamental problem facing the designers of automatic target recognition (ATR) systems is how to deal with out-of-library or non-registered targets. This research extends a mathematical programming framework that selects the optimal classifier ensemble and fusion method across multiple decision thresholds subject to classifier performance constraints. The extended formulation includes treatment of exemplars from target classes on which the ATR system is not trained (non-registered targets). Further, a multivariate Gaussian hidden Markov model (HMM) is developed and applied using real world synthetic aperture radar (SAR) data comprised of ten registered and five non-registered target classes. The framework is exercised in an experimental design across classifier fusion methods, prior probabilities of targets and non-targets, correlation between multiple sensor looks, and levels of target pose estimation error.

Adaptive boosting for SAR automatic target recognition

The paper proposed a novel automatic target recognition (ATR) system for classification of three types of ground vehicles in the moving and stationary target acquisition and recognition (MSTAR) public release database. First MSTAR image chips are represented as fine and raw feature vectors, where raw features compensate for the target pose estimation error that corrupts fine image features. Then, the chips are classified by using the adaptive boosting (AdaBoost) algorithm with the radial basis function (RBF) network as the base learner. Since the RBF network is a binary classifier, the multiclass problem was decomposed into a set of binary ones through the error-correcting output codes (ECOC) method, specifying a dictionary of code words for the set of three possible classes. AdaBoost combines the classification results of the RBF network for each binary problem into a code word, which is then decoded as one of the code words (i.e., ground-vehicle classes) in the specified dictionary. Along with classification, within the AdaBoost framework, we also conduct efficient fusion of the fine and raw image-feature vectors. The results of large-scale experiments demonstrate that our ATR scheme outperforms the state-of-the-art systems reported in the literature

Loop closure detection in SLAM by combining visual and spatial appearance

In this paper we describe a system for use on a mobile robot that detects potential loop closures using both visual and spatial appearance of local scenes. Loop closing is the act of correctly asserting that a vehicle has returned to a previously visited location. Current approaches rely heavily on vehicle pose estimates to prompt loop closure. Paradoxically, these approaches are least reliable when the need for accurate loop closure detection is the greatest. Our underlying approach relies instead upon matching distinctive ‘signatures’ of individual local scenes to prompt loop closure. A key advantage of this method is that it is entirely independent of the navigation and or mapping process and so is entirely unaffected by gross errors in pose estimation. Another advantage, which is explored in this paper, is the possibility to enhance robustness of loop closure detection by incorporating heterogeneous sensory observations. We show how a description of local spatial appearance (using laser rangefinder data) can be combined with visual descriptions to form multi-sensory signatures of local scenes which enhance loop-closure detection.

Performance analysis of a 3-2-1 pose estimation device

This paper deals with the problem of estimating the pose of a moving rigid body by measuring the length of six wires attached to it. Since wires can be seen as extensible legs, this problem is equivalent to that of solving the forward kinematics of a six-degree-of-freedom parallel manipulator. Among all possible locations for the attachments on the moving object, the "3-2-1" configuration is shown to exhibit a large number of favorable properties. The performance analysis of this particular configuration is addressed by finding analytic expressions for the estimated pose covariance matrix and the expected value of the pose estimation error, or bias error, which has been omitted in the previous analysis of wire-based tracking devices. This analysis takes advantage of a formulation for trilateration based on Cayley-Menger determinants, which is mathematically more tractable compared to previous ones, because all terms involved are determinants with geometric meaning. This accommodates a more thorough investigation of the properties of the device.

Improved Pose Estimation for Mobile Robots by Fusion of Odometry Data and Environment Map

This paper proposes a new approach for calibration of dead reckoning process. Using the well-known UMBmark (University of Michigan Benchmark) is not sufficient for a desirable calibration of dead reckoning. Besides, existing calibration methods, usually require explicit measurement of actual motion of the robot. Some recent methods, use the smart encoder trailer or long range finder sensors such as ultrasonic or laser range finders for automatic calibration. Manual measurement is necessary in the case of the robots that are not equipped with long range detectors or such smart encoder trailer. Our proposed approach, uses an environment map that is created by fusion of proximity data, in order to calibrate the odometry error automatically. In the new approach, the systematic part of the error is adaptively estimated and compensated by an efficient and incremental maximum likelihood algorithm. Actually, environment map data are fused with the odometry and current sensory data in order to acquire the maximum likelihood estimation. The advantages of the proposed approach are demonstrated in some experiments with Khepera robot. It is shown that the amount of pose estimation error is reduced by a percentage of more than 80%.