Ground-truth System Research Articles

Test Platform for Developing New Optical Position Tracking Technology towards Improved Head Motion Correction in Magnetic Resonance Imaging.

Optical tracking of head pose via fiducial markers has been proven to enable effective correction of motion artifacts in the brain during magnetic resonance imaging but remains difficult to implement in the clinic due to lengthy calibration and set up times. Advances in deep learning for markerless head pose estimation have yet to be applied to this problem because of the sub-millimetre spatial resolution required for motion correction. In the present work, two optical tracking systems are described for the development and training of a neural network: one marker-based system (a testing platform for measuring ground truth head pose) with high tracking fidelity to act as the training labels, and one markerless deep-learning-based system using images of the markerless head as input to the network. The markerless system has the potential to overcome issues of marker occlusion, insufficient rigid attachment of the marker, lengthy calibration times, and unequal performance across degrees of freedom (DOF), all of which hamper the adoption of marker-based solutions in the clinic. Detail is provided on the development of a custom moiré-enhanced fiducial marker for use as ground truth and on the calibration procedure for both optical tracking systems. Additionally, the development of a synthetic head pose dataset is described for the proof of concept and initial pre-training of a simple convolutional neural network. Results indicate that the ground truth system has been sufficiently calibrated and can track head pose with an error of <1 mm and <1°. Tracking data of a healthy, adult participant are shown. Pre-training results show that the average root-mean-squared error across the 6 DOF is 0.13 and 0.36 (mm or degrees) on a head model included and excluded from the training dataset, respectively. Overall, this work indicates excellent feasibility of the deep-learning-based approach and will enable future work in training and testing on a real dataset in the MRI environment.

Semi-supervised Learning of Dynamical Systems with Neural Ordinary Differential Equations: A Teacher-Student Model Approach

Modeling dynamical systems is crucial for a wide range of tasks, but it remains challenging due to complex nonlinear dynamics, limited observations, or lack of prior knowledge. Recently, data-driven approaches such as Neural Ordinary Differential Equations (NODE) have shown promising results by leveraging the expressive power of neural networks to model unknown dynamics. However, these approaches often suffer from limited labeled training data, leading to poor generalization and suboptimal predictions. On the other hand, semi-supervised algorithms can utilize abundant unlabeled data and have demonstrated good performance in classification and regression tasks. We propose TS-NODE, the first semi-supervised approach to modeling dynamical systems with NODE. TS-NODE explores cheaply generated synthetic pseudo rollouts to broaden exploration in the state space and to tackle the challenges brought by lack of ground-truth system data under a teacher-student model. TS-NODE employs an unified optimization framework that corrects the teacher model based on the student's feedback while mitigating the potential false system dynamics present in pseudo rollouts. TS-NODE demonstrates significant performance improvements over a baseline Neural ODE model on multiple dynamical system modeling tasks.

Theoretical and experimental investigation of variable contact forces on the rollers of a mecanum wheeled mobile robot

The modeling structures of rollers, mecanum wheels, and mecanum wheeled mobile robots presented in the literature use single contact force assumption. This assumption may give good results in a simulation environment; however, it is not strong enough to reflect reality. To make an improvement, a new aspect of mecanum wheel model is proposed in this study. The model takes the variable roller contact forces into account and investigates their effects on the performance of motion of a mecanum wheeled mobile robot. It uses all points on each roller’s curved shape so that the slippage phenomena is also taken into consideration which makes it possible to get less position estimation errors in real-time operations. The modeling structure introduced aims to reflect reality both in simulation and real applications. A simulation environment is developed for this study. To make verification, an experimental setup including a four-mecanum-wheeled mobile robot, its mechanical and electrical hardware and software infrastructures, and a ground-truth system is designed and constructed. A Robot Operating System (ROS) based control system is created and integrated into the experimental system. Different types of reference trajectories including straight-line, square-shaped, Z-shaped, and wave(S)-shaped are used to test the performance of the model proposed in both simulation and experimental studies. The tests are also conducted using the model that involves single contact force assumption to make comparisons. The details of the variable contact forces model proposed, simulation environment developed, experimental setup built, simulation and experimental studies, their results, and comparisons are given in this paper.

Data-Driven Inference on Optimal Input–Output Properties of Polynomial Systems With Focus on Nonlinearity Measures

In the context of dynamical systems, nonlinearity measures quantify the strength of nonlinearity by means of the distance of their input-output behaviour to a set of linear input-output mappings. In this paper, we establish a framework to determine nonlinearity measures and other optimal input-output properties for nonlinear polynomial systems without explicitly identifying a model but from a finite number of input-state measurements which are subject to noise. To this end, we deduce from data for the unidentified ground-truth system three possible set-membership representations, compare their accuracy, and prove that they are asymptotically consistent with respect to the amount of samples. Moreover, we leverage these representations to compute guaranteed upper bounds on nonlinearity measures and the corresponding optimal linear approximation model via semi-definite programming. Furthermore, we extend the established framework to determine optimal input-output properties described by time domain hard integral quadratic constraints.

Static and Dynamic Accuracy and Occlusion Robustness of SteamVR Tracking 2.0 in Multi-Base Station Setups.

The tracking of objects and person position, orientation, and movement is relevant for various medical use cases, e.g., practical training of medical staff or patient rehabilitation. However, these demand high tracking accuracy and occlusion robustness. Expensive professional tracking systems fulfill these demands, however, cost-efficient and potentially adequate alternatives can be found in the gaming industry, e.g., SteamVR Tracking. This work presents an evaluation of SteamVR Tracking in its latest version 2.0 in two experimental setups, involving two and four base stations. Tracking accuracy, both static and dynamic, and occlusion robustness are investigated using a VIVE Tracker (3.0). A dynamic analysis further compares three different velocities. An error evaluation is performed using a Universal Robots UR10 robotic arm as ground-truth system under nonlaboratory conditions. Results are presented using the Root Mean Square Error. For static experiments, tracking errors in the submillimeter and subdegree range are achieved by both setups. Dynamic experiments achieved errors in the submillimeter range as well, yet tracking accuracy suffers from increasing velocity. Four base stations enable generally higher accuracy and robustness, especially in the dynamic experiments. Both setups enable adequate accuracy for diverse medical use cases. However, use cases demanding very high accuracy should primarily rely on SteamVR Tracking 2.0 with four base stations.

Experimental Benchmarking of Next-Gen Indoor Positioning Technologies (Unmodulated) Visible Light Positioning and Ultra-Wideband

Within the context of the Internet of Things (IoT), many applications require high-quality positioning services. As opposed to traditional technologies, the two most recent positioning solutions: 1) ultra-wideband (UWB) and 2) (unmodulated) visible light positioning [(u)VLP] are well suited to economically supply centimeter-to-decimeter level accuracy. This manuscript benchmarks the 2-D positioning performance of an 8-anchor asymmetric double-sided two-way ranging (aSDS-TWR) UWB system and a 15-LED frequency-division multiple access (FDMA) received signal strength (RSS) (u)VLP system in terms of feasibility and accuracy. With extensive experimental data, collected at two heights in a 8 m by 6 m open zone equipped with a precise ground-truth system, it is demonstrated that both visible light positioning (VLP) and UWB already attain median and 90 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> percentile positioning errors in the order of 5 and 10 cm in line-of-sight (LOS) conditions. An approximately 20-cm median accuracy can be obtained with uVLP, whose main benefit is it being infrastructureless and thus very inexpensive. The accuracy degradation effects of non-LOS (NLOS) on UWB/(u)VLP are highlighted with four scenarios, each consisting of a different configuration of metallic closets. For the considered setup, in 2-D and with minimal tilt of the object to be tracked, VLP outscores UWB in NLOS conditions, while for LOS scenarios similar results are obtained.

Using a single inertial sensor to control exergames for children with cerebral palsy

Abstract Introduction: Children with cerebral palsy (CCP) benefit from intensive arm training. Exergames that can be played at home offer the possibility to increase the frequency of therapy but require reliable and accurate real-time motion tracking via easy-to-use sensors in unsupervised settings and magnetically disturbed environments. Method: We propose an inertial-sensor-based method with a single sensor on the wrist for real-time tracking of the inclination of the forearm. The control parameter of the game was validated with an optical marker-based ground truth system. Results: First experiments with a therapist performing training movements in a healthy and simulated spastic manner show that the forearm inclination well captures the motion dynamics. The accuracy of the inertial-sensor-based measurement is validated with respect to the reference system in three healthy subjects. Orientation offsets between the inertial sensor and the forearm marker set in the range of 2° ̶ 6° and dynamic measurement errors about 3.1° were obtained. Conclusion: This work demonstrates that the proposed method is suitable for real-time control of exergames of CCP. The validation with an optical reference system showed that the forearm inclination can be used as for feedback and therapeutic progress monitoring.

Low-Cost Hybrid Multisensor Fusion Method and Implementation for Production Intelligent Vehicles

Faced with cost limitation and 360-degree fusion and tracking requirement of production intelligent vehicles, this paper proposes a low-cost hybrid multi-sensor fusion method, which provides flexibility and scalability for multi-level information input and different sensor configurations. By combining radar, lane detection and ego vehicle information, we design a centralized fusion algorithm for 360-degree object tracking, which provides a low-computation multi-sensor association algorithm by a scheme of local sensor to global track association, and proves an optimal centralized estimator in the condition of multi-coordinate system Doppler information. To make the sensor fusion system compatible with multi-level information input and enhance the redundancy to improve security, a distributed fusion algorithm based on covariance intersection method is also designed as a part of fusion framework for the secondary fusion of high-level information. We validate the feasibility of our fusion method in an industrial-level controller with only 10% computing power consumption and build a ground truth system to quantitatively evaluate the method, which shows that our fusion method performs well on 360-degree object tracking and guarantees stable object IDs, and greatly improves the accuracy especially for sensor boundary regions. Furthermore, we also test and evaluate the performance of our fusion method in a variety of complex open road scenarios including expressway, curved road and lane fading scenarios.

Position detection for electric vehicle DWCS using VI-SLAM method

The dynamic wireless charging system (DWCS) is developed to solve the problems of large battery volume and mileage anxiety of electric vehicles. However, the accurate position detection for electric vehicle DWCS is facing challenge. The traditional communication, detection and estimation methods are difficult to accurately obtain the position. To tackle this problem, the visual inertial simultaneous localization and mapping (VI-SLAM) method is applied to the electric vehicles DWCS. Firstly, the graph optimization based tight coupling method is used to integrate the monocular visual and inertial measurement unit (IMU) measurements. Secondly, the NVIDIA TX2 and MYNT VI-sensor suite are assembled, which the MTi300-IMU is treated as the ground truth system. Finally, the mobile vehicle is controlled to race on the simulated DWCS pathway. The experimental result shows that the method achieves great performance with the accuracy of centimeter level. In particular, the root mean square error (RMSE) of pose (i.e. position and orientation) in X, Y, Z directions are 0.086 m, 0.092 m, 0.102 m and 2.423°, 1.682°, 2.501°, respectively. Compared with the smooth variable structure filter (SVSF) based SLAM method, 0.106 m, 0.098 m, 0.130 m and 3.069°, 3.261°, 2.961°, the accuracy of ours is increased by 18.87%, 6.12%, 21.54% and 21.05%, 45.42%, 15.54%, respectively.

Autonomous Fingerprinting and Large Experimental Data Set for Visible Light Positioning.

This paper presents an autonomous method of collecting data for Visible Light Positioning (VLP) and a comprehensive investigation of VLP using a large set of experimental data. Received Signal Strength (RSS) data are efficiently collected using a novel method that utilizes consumer grade Virtual Reality (VR) tracking for accurate ground truth recording. An investigation into the accuracy of the ground truth system showed median and 90th percentile errors of 4.24 and 7.35 mm, respectively. Co-locating a VR tracker with a photodiode-equipped VLP receiver on a mobile robotic platform allows fingerprinting on a scale and accuracy that has not been possible with traditional manual collection methods. RSS data at 7344 locations within a 6.3 × 6.9 m test space fitted with 11 VLP luminaires is collected and has been made available for researchers. The quality and the volume of the data allow for a robust study of Machine Learning (ML)- and channel model-based positioning utilizing visible light. Among the ML-based techniques, ridge regression is found to be the most accurate, outperforming Weighted k Nearest Neighbor, Multilayer Perceptron, and random forest, among others. Model-based positioning is more accurate than ML techniques when a small data set is available for calibration and training. However, if a large data set is available for training, ML-based positioning outperforms its model-based counterparts in terms of localization accuracy.

Position Detection for Electric Vehicle Dwcs Using Vi-Slam Method

The dynamic wireless charging system (DWCS) is developed to solve the problems of large battery volume and mileage anxiety of electric vehicles. However, the accurate position detection for electric vehicle DWCS is facing challenge. The traditional communication, detection and estimation methods are difficult to accurately obtain the position. To tackle this problem, the visual inertial simultaneous localization and mapping (VI-SLAM) method is applied to the electric vehicles DWCS. Firstly, the graph optimization based tight coupling method is used to integrate the monocular visual and IMU measurements. Secondly, the NVIDIA TX2 and MYNT VI-sensor suite are assembled, which the MTi300-IMU is treated as the ground truth system. Finally, the mobile vehicle is controlled to race on the simulated DWCS pathway. The experimental result shows that the method achieves great performance with the accuracy of centimeter level. In particular, the root mean square error (RMSE) in X, Y, Z directions are 0.086,0.092,0.102, respectively.

Dissipativity Verification With Guarantees for Polynomial Systems From Noisy Input-State Data

In this letter, we investigate the verification of dissipativity properties for polynomial systems without an explicitly identified model but directly from noise-corrupted measurements. Contrary to most data-driven approaches for nonlinear systems, we determine dissipativity properties over all finite time horizons using noisy input-state data. To this end, we propose two noise characterizations to deduce two data-based set-membership representations of the ground-truth system. Each representation then serves as a framework to derive computationally tractable conditions to verify dissipativity properties with rigorous guarantees from noise-corrupted data using sum of squares (SOS) optimization.

Efficient Upper Limb Position Estimation Based on Angular Displacement Sensors for Wearable Devices.

Motion tracking techniques have been extensively studied in recent years. However, capturing movements of the upper limbs is a challenging task. This document presents the estimation of arm orientation and elbow and wrist position using wearable flexible sensors (WFSs). A study was developed to obtain the highest range of motion (ROM) of the shoulder with as few sensors as possible, and a method for estimating arm length and a calibration procedure was proposed. Performance was verified by comparing measurement of the shoulder joint angles obtained from commercial two-axis soft angular displacement sensors (sADS) from Bend Labs and from the ground truth system (GTS) OptiTrack. The global root-mean-square error (RMSE) for the shoulder angle is 2.93 degrees and 37.5 mm for the position estimation of the wrist in cyclical movements; this measure of RMSE was improved to 13.6 mm by implementing a gesture classifier.

A system for the generation of in-car human body pose datasets

With the advent of autonomous vehicles, detection of the occupants’ posture is crucial to tackle the needs of infotainment interaction or passive safety systems. Generative approaches have been recently proposed for human body pose in-car detection, but this type of approaches requires a large training dataset for a feasible accuracy. This requirement poses a difficulty, given the substantial time required to annotate such large amount of data. In the in-car scenario, this requirement risk increases even further, since a robust human body pose ground-truth system capable of working in it is needed but inexistent. Currently, the gold standard for human body pose capture is based on optical systems, requiring up to 39 visible markers for a plug-in gait model, which in this case are not feasible given the occlusions inside the car. Other solutions, such as inertial suits, also have limitations linked to magnetic sensitivity and global positioning drift. In this paper, a system for the generation of images for human body pose detection in an in-car environment is proposed. To this end, we propose to smartly combine inertial and optical systems to suppress their individual limitations: By combining the global positioning of 3 visible head markers provided by the optical system with the inertial suit’s relative human body pose, we obtain an occlusion-ready, drift-free full-body global positioning system. This system is then spatially and temporally calibrated with a time-of-flight sensor, automatically obtaining in-car image data with (multi-person) pose annotations. Besides quantifying the inertial suit inherent sensitivity and accuracy, the feasibility of the overall system for human body pose capture in the in-car scenario was demonstrated. Our results quantify the errors associated with the inertial suit, pinpoint some sources of the system’s uncertainty and propose how to minimize some of them. Finally, we demonstrate the feasibility of using system generated data (which was made publicly available), independently or mixed with two publicly available generic datasets (not in-car), to train 2 machine learning algorithms, demonstrating the improvement in their algorithmic accuracy for the in-car scenario.

A model for vessel trajectory prediction based on long short-term memory neural network

Each vessel has its own way of sailing in the port region. Any autonomous vessel navigating such a scene should be able to predict the trajectories of surrounding ships and adjust its behaviour to avoid a collision. In this paper, combined with the sequence prediction method, a Long Short-Term Memory (LSTM) model is proposed to predict the trajectories of the vessels. The ground-truth Automatic Identification System (AIS) data in the port of Tianjin, China are used to train and test the proposed model. The experimental results prove that our model can predict ship trajectories accurately, and it is applicable to the autonomous navigation system.

Ground-truth localisation system for humanoid soccer robots using RGB-D camera

Localisation and behaviour are two vital systems for every autonomous mobile robot. As the result, establishment of a localisation ground-truth detection system is a necessity for the analysis. Depend on the requirements of the application, there are some constraints for such system like: diversity in types of providing information, accuracy of the data and ease of use - since the calibration is needed in both colour classification and geometry position - and also the cost of the system, which all should be considered. In this paper, a low cost system is presented using Asus Xtion depth camera, capable of providing ground-truth position information of the humanoid robots and the ball which is required to estimate the errors of a localisation system in RoboCup Standard Platform League (SPL) and Humanoid kid Size League (HSL) soccer robots. Furthermore, the user input requirements are reduced to very few compared to other similar systems. Source code of the system is available online at https://github.com/mrlspl/.

Interest Aware Location-Based Recommender System Using Geo-Tagged Social Media

Advances in location acquisition and mobile technologies led to the addition of the location dimension to Social Networks (SNs) and to the emergence of a newer class called Location-Based Social Networks (LBSNs). While LBSNs are richer in their model and functions than SNs, they fail so far to attract as many users as SNs. On the other hand, SNs have large amounts of geo-tagged media that are under-utilized. In this paper, we propose an Interest-Aware Location-Based Recommender system (IALBR), which combines the advantages of both LBSNs and SNs, in order to provide interest-aware location-based recommendations. This recommender system is proposed as an extension to LBSNs. It is novel in: (1) utilizing the geo-content in both LBSNs and SNs; (2) ranking the recommendations based on a novel scoring method that maps to the user interests. It also works for passive users who are not active content contributors to the LBSN. This feature is critical to increase the number of LBSN users. Moreover, it helps with reducing the cold start problem, which is a common problem facing the new users of recommender systems who get random unsatisfying recommendations. This is due to the lack of user interest awareness, which is reliant on user history in most of the recommenders. We evaluated our system with a large-scale real dataset collected from foursquare with respect to precision, recall and the f-measure. We also compared the results with a ground truth system using metrics like the normalized discounted cumulative gain and the mean absolute error. The results confirm that the proposed IALBR generates more efficient recommendations than baselines in terms of interest awareness.

Cooperative Localization with Reliable Confidence Domains Between Vehicles Sharing GNSS Pseudoranges Errors with No Base Station

Cooperation between road vehicles through information exchange is a promising way to enhance their absolute and relative positions. This paper presents an approach for generating, sharing and applying Global Navigation Satellite System (GNSS) pseudorange corrections through a V2X communication network. Conventionally, differential corrections are generated by fixed base stations with known positions and sent to mobile users. Here, the proposed cooperative method has no central server and the estimation of the raw measurements errors is done in a fully distributed way. Using a model of the correlation of the pseudorange errors and through the knowledge of the local motions of the vehicles obtained by Dead Reckoning (DR) or tracking, a non linear observability shows that the estimation problem is solvable. A cooperative and fully distributed estimation method is then presented using Set Inversion and Constraint propagation techniques. Positions, pseudorange estimated errors and DR data are shared in the network of vehicles and confidence is handled by intervals, in a bounded error context. This allows computing highly reliable confidence domains with no direct range measurements, which is crucial for applications involving close proximity navigation. Indeed, the proposed data fusion framework does not require any linearization of the equations and is insensitive to the data incest problem since the same information can be exploited several times in the computation process without making the estimation over-converge. Results using real measurements are presented to illustrate the performance of the proposed cooperative method in comparison with standalone estimation. A classical sequential Bayesian method has also been implemented on the same data set and compared in terms of accuracy and confidence with a ground truth system.

Inferring connectivity in networked dynamical systems: Challenges using Granger causality.

Determining the interactions and causal relationships between nodes in an unknown networked dynamical system from measurement data alone is a challenging, contemporary task across the physical, biological, and engineering sciences. Statistical methods, such as the increasingly popular Granger causality, are being broadly applied for data-driven discovery of connectivity in fields from economics to neuroscience. A common version of the algorithm is called pairwise-conditional Granger causality, which we systematically test on data generated from a nonlinear model with known causal network structure. Specifically, we simulate networked systems of Kuramoto oscillators and use the Multivariate Granger Causality Toolbox to discover the underlying coupling structure of the system. We compare the inferred results to the original connectivity for a wide range of parameters such as initial conditions, connection strengths, community structures, and natural frequencies. Our results show a significant systematic disparity between the original and inferred network, unless the true structure is extremely sparse or dense. Specifically, the inferred networks have significant discrepancies in the number of edges and the eigenvalues of the connectivity matrix, demonstrating that they typically generate dynamics which are inconsistent with the ground truth. We provide a detailed account of the dynamics for the Erdős-Rényi network model due to its importance in random graph theory and network science. We conclude that Granger causal methods for inferring network structure are highly suspect and should always be checked against a ground truth model. The results also advocate the need to perform such comparisons with any network inference method since the inferred connectivity results appear to have very little to do with the ground truth system.

Performance Evaluation of Automotive Radars Using Carrier-Phase Differential GPS

We describe a centimeter-level ground-truth system based on GPS relative positioning. One of the goals of the system is for performance evaluation of automotive radars by creating objective test procedures for the radars. In support of this goal, an effective scheme is used to handle signal attenuation or blockage that leads to the changes of satellite visible map and cycle slip. Multiplatform GPS measurements and low-cost vehicle dynamic sensor data are integrated for vehicle trajectory reconstruction. Experimental results are presented to illustrate the system's actual centimeter-level accuracy and superior performance over Kalman smoothing. The effectiveness of the proposed ground-truth system is demonstrated by evaluating mono-pulse short-range radars whose accuracy is about submeter level.