Search

A static world assumption is often used when considering the simultaneous localization and mapping (SLAM) problem. In reality, especially when long-term autonomy is the objective, this is not valid. This paper studies a scenario where uniquely identifiable landmarks can attend multiple discrete positions, not known a priori . Based on a feature based multi-hypothesis map representation, a multi-hypothesis SLAM algorithm is developed inspired by target tracking theory. The creation of such a map is merged into the SLAM framework allowing any available SLAM method to solve the underlying mapping and localization problem for each hypothesis. A recursively updated hypothesis score allows for hypothesis rejection and prevents exponential growth in the number of hypotheses. The developed method is evaluated in an underground mine application, where physical barriers can be moved in between multiple distinct positions. Simulations are conducted in this environment showing the benefits of the multi-hypothesis approach compared to executing a standard SLAM algorithm. Practical considerations as well as suitable approximations are elaborated upon and experiments on real data further validates the simulated results and show that the multi-hypothesis approach has similar performance in reality as in simulation. Funding: Knut and Alice Wallenberg Foundation through Wallenberg AI, Autonomous Systems and Software Program (WASP) Wallenberg AI, Autonomous Systems and Software Program (WASP)

IEEEWhile the zeitgeist of today prescribes localization as a key to eliminate the excessive transportation for manufacturing firms, firm-level practical and workable guidelines that go beyond maxims are scant. Laden with analytical methods, this paper offers a phronetic, customizable framework based on Lean Six Sigma (LSS). The framework mimics Define-Measure-Analyze-Improve-Control (DMAIC) process. The early phases of the framework, ex-ante, prescribe the use of a variety tools and techniques from Lean Six sigma (LSS) and employ fuzzy analytic hierarchy process (FAHP) to extract the vital few root-causes that may impede localization. The latter phases suggest the use of design of experiments (DOE) to evaluate the root-causes, ex-post. The paper outlines a clear and established procedure and demonstrates each analytical phase using evidence from a two-year real-world application in steel industry.

Magnetic actuation of minimally invasive medical tetherless devices holds great promise in several biomedical applications. However, there are still several challenges in noninvasive localization, both in terms of sensing detectable signals of these devices and estimating their states. In this work, a magnetic milli-roller is actuated in a viscous fluid under the influence of a rotating magnetic field. A Lyapunov-based nonlinear state observer is designed and implemented to estimate the position of the roller using a 2D array of Hall-effect sensors. We show that the local stability of the state observer yields convergence to one of the local equilibria, for pre-defined levels of sensor noise, initial conditions, and modeling errors. Performance is quantified using redundant measurements of the fields and we investigate the influence of the number of magnetic field measurements on the observability of the system. Open-loop actuation and state estimation are demonstrated and experimental results show that the localization of a 5 mm diameter roller along sinusoidal, circular and square trajectories achieve a steady-state mean absolute position error of 2.3mm, 1.67mm and 1.73mm, respectively.

Ultra-Wide Band (UWB) communication systems can be used to design low cost, power efficient and precise navigation systems for mobile robots, by measuring the Time of Flight (ToF) of messages traveling between on-board UWB transceivers to infer their locations. Theoretically, decimeter level positioning accuracy or better should be achievable, at least in benign propagation environments where Line-of-Sight (LoS) between the transceivers can be maintained. Yet, in practice, even in such favorable conditions, one often observes significant systematic errors (bias) in the ToF measurements, depending for example on the hardware configuration and relative poses between robots. This letter proposes a ToF error model that includes a standard transceiver clock offset term and an additional term that varies with the received signal power (RxP). We show experimentally that, after fine correction of the clock offset term using clock skew measurements available on modern UWB hardware, much of the remaining pose dependent error in LoS measurements can be captured by the (appropriately defined) RxP-dependent term. This leads us to propose a simple bias compensation scheme that only requires on-board measurements (clock skew and RxP) to remove most of the observed bias in LoS ToF measurements and reliably achieve cm-level ranging accuracy. Because the calibrated ToF bias model does not depend on any extrinsic information such as receiver distances or poses, it can be applied before any additional error correction scheme that requires more information about the robots and their environment.

Autonomous exploration is a fundamental task for mobile robots. However, a major limitation is that in practical applications, robots usually face dynamic obstacles such as pedestrian crowds. These dynamic objects create significant challenges for both collision-free navigation and accurate localization/mapping, which can compromise the safety and exploration performance of the robot. In this work, a hierarchical approach is proposed for both effective exploration and collision-free navigation in crowded environments. The central idea of our approach is to combine local and global information to ensure the safety and efficiency of the exploration planner. Besides, our planning method utilizes a reinforcement learning (RL)-based obstacle avoidance algorithm that allows the robot to safely follow the exploration planner's path through the pedestrian crowd. The proposed system is thoroughly evaluated in simulation environments, and the results show that it outperforms existing methods in terms of not only the exploration efficiency but also the localization and mapping accuracy.

Scene text detection has attracted increasing concerns with the rapid development of deep neural networks in recent years. However, existing scene text detectors may overfit on the public datasets due to the limited training data, or generate inaccurate localization for arbitrary-shape scene texts. This paper presents an arbitrary-shape scene text detection method that can achieve better generalization ability and more accurate localization. We first propose a Scale-Aware Data Augmentation (SADA) technique to increase the diversity of training samples. SADA considers the scale variations and local visual variations of scene texts, which can effectively relieve the dilemma of limited training data. At the same time, SADA can enrich the training minibatch, which contributes to accelerating the training process. Furthermore, a Shape Similarity Constraint (SSC) technique is exploited to model the global shape structure of arbitrary-shape scene texts and backgrounds from the perspective of the loss function. SSC encourages the segmentation of text or non-text in the candidate boxes to be similar to the corresponding ground truth, which is helpful to localize more accurate boundaries for arbitrary-shape scene texts. Extensive experiments have demonstrated the effectiveness of the proposed techniques, and state-of-the-art performances are achieved over public arbitrary-shape scene text benchmarks (e.g., CTW1500, Total-Text, and ArT).

Ultra-wideband (UWB) wireless indoor positioning systems rely on time of flight (TOF) to estimate distances but can be biased and miscalculated due to non-line-of-sight (NLOS) transmission channels in complex environments. Therefore, to remove errors, several machine learning techniques have been proposed for identifying NLOS signals from Channel Impulse Responses (CIRs). However, as CIR signals could be heavily influenced by various environments, current NLOS classifiers are not universal to provide satisfactory accuracy for new scenarios and require detailed measurements on a large number of CIRs for training. Hence, we propose a generalization method based on data augmentation via noise injection and transfer learning to allow the deep neural network (DNN) trained under a lab condition to be applied to various and even harsh practical scenarios with the need to measure massive training data minimized. This paper presents the first demonstration that it is effective to utilize a lab-based pre-trained DNN for real-world transfer and white Gaussian noise data augmentation for ML-based NLOS identification on UWB CIRs to address the problem when it is not feasible to measure sufficient training data. Our testing results show that in two scenarios, corridor and parking lot, with only 50 CIR signals as the training set, the accuracy of the NLOS identification model after applying the proposed method is increased from 84.4% to 98.8% and from 81.1% to 97.1%, respectively.

Both weakly supervised single object localization and semantic segmentation techniques learn an object's location using only image-level labels. However, these techniques are limited to cover only the most discriminative part of the object and not the entire object. To address this problem, we propose an attention-based dropout layer, which utilizes the attention mechanism to locate the entire object efficiently. To achieve this, we devise two key components, 1) hiding the most discriminative part from the model to capture the entire object, and 2) highlighting the informative region to improve the classification power of the model. These allow the classifier to be maintained with a reasonable accuracy while the entire object is covered. Through extensive experiments, we demonstrate that the proposed method effectively improves the weakly supervised single object localization accuracy, thereby achieving a new state-of-the-art localization accuracy on the CUB-200-2011 and a comparable accuracy existing state-of-the-arts on the ImageNet-1k. The proposed method is also effective in improving the weakly supervised semantic segmentation performance on the Pascal VOC and MS COCO. Furthermore, the proposed method is more efficient than existing techniques in terms of parameter and computation overheads. Additionally, the proposed method can be easily applied in various backbone networks.

Visible light positioning (VLP) is a promising technology since it can provide high accuracy indoor localization based on the existing lighting infrastructure. However, existing approaches often require dense LED distributions and persistent line-of-sight (LOS) between transmitter and receiver. What's more, sensors are imperfect, and their measurements are prone to errors. Through multi sensors fusion, we can compensate the deficiencies of stand-alone sensors and provide more reliable pose estimations. In this work, we propose a loosely-coupled multi-sensor fusion method based on VLP and Simultaneous Localization and Mapping (SLAM), using light detection and ranging (LiDAR), odometry, and rolling shutter camera. Our multi-sensor localizer can provide accurate and robust robot localization and navigation in LED shortage/outage situations. The experimental results show that our proposed scheme can provide an average accuracy of 2.5 cm with around 42 ms average positioning latency.

Indoor localization of human objects has many important applications nowadays. Here, we propose a new multiresolution device-free indoor localization scheme using wireless radio frequency (RF) fingerprints. In our proposed new device-free approach, all transceiver devices are fixed in an indoor environment, so human targets do not need to carry any transceiver device with them. First, the indoor geometry is divided into several zones, and then, the received signal strength indicators measured by the receiving antennas are input features to our designed innovative machine learning model to identify within which zone the target is. Our proposed machine learning model, i.e., a multiresolution random forest classifier, is composed of a cascade architecture, which integrates and distills learned results over various zoning resolutions. The proposed new multiresolution approach greatly outperforms the existing random forest classifier. The average Euclidean distance error resulting from our proposed new technique is 1.0376 m.

Indoor location-based services (LBS) have exhibited large commercial and social values in smart cities, and urgent demands of which have spurred many localization techniques. Existing indoor localization approaches mostly rely on fingerprint techniques, leveraging either spatially discrete fingerprints or temporally consecutive ones for localization. However, these approaches often suffer from large errors or high time overhead in practice due to signal ambiguities or long input sequences. To overcome these drawbacks, this paper proposes a framework utilizing multiple adaptive representations of signal sequences for localization, where each representation indicates a corresponding signal structure with underlying location clues. As an example, the proposed approach takes geomagnetic signal sequences as input and infers location features from two intuitive representations, e.g., spatial and temporal ones. With adaptive signal representations, the proposed approach takes specifically optimized neural networks to extract corresponding location clues respectively and fuses them to generate more distinguishing features for more accurate localization. Furthermore, the ensemble learning mechanism is adopted in the approach and a weighted k-NN based location estimation algorithm is devised to enhance the robustness. Extensive experiments in three different trial sites demonstrate that the proposed approach outperforms state-of-the-art competing schemes by a wide margin, reducing mean localization error by more than 46%.

Due to its high-dimensional data characteristics, the channel state information (CSI) of Wi-Fi signals has become a strong candidate for use in indoor localization. In addition, machine learning techniques can improve the accuracy of indoor localization systems using multiview CSI data received at multiple access points (APs). However, in complex environments, most CSI views collected at APs in non-line-of-sight (NLoS) configurations relative to a transmitter may lose so much useful data information as to become nonsalient. In this paper, we propose a practical machine learning approach named unsupervised view-selective deep learning (UVSDL), in which only the most salient CSI view is selected in an unsupervised manner to be applied in regression for localization. In an experiment in a complex building, our variational deep learning (VDL)-based regression method with the most salient CSI view achieves a localization accuracy of 1.36 m, significantly outperforming the best-known system BiLoc by 25 %.

Accurate positioning is an essential requirement of autonomous vehicular navigation system (AVNS) for safe driving. Although the vehicle position can be obtained in global position system friendly environments, in GPS denied environments (such as suburb, tunnel, forest, or underground scenarios) the positioning accuracy of AVNS is easily reduced by the trajectory error of the vehicle. In order to solve this problem, the plane, sphere, cylinder and cone are often selected as the ground control targets to eliminate the trajectory error for AVNS. However, these targets usually suffer from the limitations of incidence angle, measuring range, scanning resolution, and point cloud density, etc. To bridge this research gap, an adaptive continuum shape constraint analysis (ACSCA) method is presented in this article to design a new target with optimized identifiable specific shape to eliminate the trajectory error for AVNS. First of all, according to the proposed ACSCA method, we conduct extensive numerical simulations to explore the optimal ranges of the vertexes and the faces for target shape design, and based on these trials, the optimal target shape is found as icosahedron, which composes of ten vertexes, 20 faces and combines the properties of plane and volume target. Moreover, the algorithm of automatic detection and coordinate calculation is developed to recognize the icosahedron target and calculate its coordinates information for AVNS. Finally, a series of experimental investigation were performed to evaluate the effectiveness of the designed icosahedron target in GPS denied environments. The experimental results demonstrate that compared with the plane, sphere, cylinder and cone targets, the developed icosahedron target can produce better performances than the above targets in terms of the clustered minimum registration error, ambiguity and range of field-of-view; also can significantly improve the positioning accuracy of AVNS in GPS denied environments.

The lighthouse localization system has recently been developed and used for localization in virtual reality (VR). Not only for VR but also for a general indoor positioning systems (IPSs) it has several advantages over existing methods, including low cost, wide detection area, and easy setup. Here, we adopt the stereo configuration of a lighthouse for improved sensing performance and propose a novel calibration method for stereo configuration. For the stereo calibration, the exact positions of sufficient corresponding points in two sensor coordinates need to be determined. A printed checkerboard is widely used for stereo camera systems because it is easy to construct and its accuracy is guaranteed owing to its printing accuracy. However, in the case of the lighthouse system, it is very difficult or impossible to construct a highly accurate calibration board similar to the checkerboard mainly because of manufacturing errors. In this study, we use a receiver sensor and a two-axis linear stage equipped with micrometers. By moving predetermined distances along the x and y directions on the linear stage, we can obtain multiple-point information with high accuracy, which can then be used for the stereo calibration of two lighthouses. In this paper, the calibration and pose estimation procedures are described in detail, and the pose estimation result of the perspective-n-points method is compared with that of the triangulation method. Finally, the pose estimation accuracy of the proposed system is compared with that of a commercial system that is widely used for highly accurate medical applications.

This article proposes a method to detect vocal signals and the corresponding locations of multiple humans using a multiple-input and multiple-output (MIMO) frequency-modulated continuous-wave (FMCW) radar system. The signals of multiple humans are extracted with digital beamforming by the virtual array generated by MIMO radar processing. To remove body motion artifacts that distort the speech information detected by the radar, a body motion effect cancellation method is proposed. The technique allows recovering vocal signals from detected distorted signals by estimating phase variations due to body motions. A combination of the proposed vocal signal detection algorithm with a human localization one enables to detect speaking humans by means of the MIMO radar. Experiments verify that the proposed method can accurately locate and detect speaking humans based on their vocal vibration signals.

We consider a distributed cooperative localization (CL) algorithm for a group of mobile robots in GPS-denied environments. Each robot is equipped with IMU and optical cameras. Using the CL method, mobile robots in a group achieve stable self-localization by communicating and processing interrobot relative measurements with sparse absolute positioning information such as a few identifiable landmarks. This paper provides a fully-distributed algorithm for bearing-only cooperative localization in a form of stepwise update, in which each agent plays as both the observer and the observed to update its states in two steps. It’s an iterative algorithm that requires less memory, and does not need all-to-all communication. In the proposed method, the covariance intersection (CI) is generalized to a bearing measurement model and it can be applied in switching topology. The simulation result shows the feasibility of the present method.

Nowadays, the autonomous vehicle has undergone an impressive development, in which one of the most challenging tasks is to pursue high-precision and continuous navigation in complex scenarios. However, GNSS signals are very likely to be blocked and the inertial system suffers from error drift over time. To achieve centimeter-level positioning accuracy, a tightly-coupled model is built via fusing visual, GNSS, and INS measurements, based on a priori visual feature map. The visual map is first generated by a visual mobile surveying system, and the quality of the map is improved by outlier elimination and road lane tracking. This visual feature map consists of two layers that involve 3D coordinates of robust visual features and road lane markings, then two GNSS/INS/Vision fusion models are developed. The first one fuses visual 2D-3D feature matching, GNSS, and INS measurement. The proposed system is effective even when satellites are all out of lock. The second one fuses lane line constraints with GNSS, INS, and distance measurement indicator (DMI). In the experiment part, we show how a visual feature map is generated and refined, then, the performance of the positioning algorithm is evaluated. For the first fusion model, we achieve 10 cm RMS accuracy when only one feature point was matched and all GNSS signals are blocked for 20 minutes. For the other fusion model, the system maintained 10 cm RMS accuracy or even better when a few GNSS satellites are out of lock for 300 seconds.

Thanks to their high carrying capacity and strong maneuverability, carlike robots which move with non-holonomic constraints, are frequently utilized in numerous coverage operation fields. In such fields, the robots need to complete the coverage task via autonomous path planning and tracking, which is referred to as a “Coverage Navigation for Carlike Robots” (CNCR) problem. When it comes to CNCR, two folds of challenges are put forward as follows. On one hand, the local path planner should not only consider the non-holonomic constraints of carlike robots but also should meet the high coverage rate required by the coverage task. On the other hand, during CNCR, there must be many U-turns in narrow areas. As a result, Reeds-Shepp-type paths will be planned inevitably. Tracking this sort of multi-segment paths stably is also an urgent problem to be solved in CNCR. To this end, this letter introduces a reference line constraint and a heuristic initialization strategy along with a cubic parametric path smoother to the time elastic band planner and designs a multi-segment-supported path tracker. Experimental results show that the proposed path planner largely reduces the average deviation off the reference line and obtains a high success rate even in complex environments. Besides, the designed path tracker achieves stable and high-precision tracking for multi-segment paths. Real-world experiments reveal that the proposed path planner and tracker are fully competent for CNCR, rendering our sweeping robot to attain an actual average coverage rate of 92.21%. To make our results fully reproducible, relevant codes and data are made available online at https://github.com/csLinZhang/cncr.git .

We consider the problem of abnormality localization for clinical applications. While deep learning has driven much recent progress in medical imaging, many clinical challenges are not fully addressed, limiting its broader usage. While recent methods report high diagnostic accuracies, physicians have concerns trusting these algorithm results for diagnostic decision-making purposes because of a general lack of algorithm decision reasoning and interpretability. One potential way to address this problem is to further train these models to localize abnormalities in addition to just classifying them. However, doing this accurately will require a large amount of disease localization annotations by clinical experts, a task that is prohibitively expensive to accomplish for most applications. In this work, we take a step towards addressing these issues by means of a new attention-driven weakly supervised algorithm comprising a hierarchical attention mining framework that unifies activation- and gradient-based visual attention in a holistic manner. Our key algorithmic innovations include the design of explicit ordinal attention constraints, enabling principled model training in a weakly-supervised fashion, while also facilitating the generation of visual-attention-driven model explanations by means of localization cues. On two large-scale chest X-ray datasets (NIH ChestX-ray14 and CheXpert), we demonstrate significant localization performance improvements over the current state of the art while also achieving competitive classification performance. Our code is available on https://github.com/oyxhust/HAM.

This paper studies the multi-robot cooperative localization (CL) problem, a challenging scenario in which robots may receive spurious sensor data, potentially causing inconsistent state estimates. To address this problem, this paper presents a fully decentralized CL algorithm based on covariance union (CU), referred to as DCL-CU. The proposed approach is fault-tolerant and supports generic measurement models. Extensive Monte Carlo simulations and a group of real-world experiments were conducted to verify the performance of the proposed DCL-CU approach. The results show that the DCL-CU approach can efficiently deal with spurious sensor data.

This letter presents a decentralized algorithm that allows a swarm of identically programmed agents to cooperatively estimate their global poses using local range and bearing measurements. The design of our algorithm explicitly considers the phase asynchrony of each agent's local clock, moreover, the execution of our algorithm does not require each agent to actively keep the same neighbors over time. A theoretical analysis about the effect of each agent's sensing noise and communication loss is given, in addition, we validate the presented algorithm via experiments running on a swarm of up to 256 simulated robots and a swarm of 100 physical robots. The results from the experiments show that the presented algorithm allows each agent to estimate its global pose quickly and reliably. Video of 100 robots executing the presented algorithm as well as supplementary material can be found in [1] and [2] .

The core problem of visual multi-robot simultaneous localization and mapping (MR-SLAM) is how to efficiently and accurately perform multi-robot global localization (MR-GL). The difficulties are two-fold. The first is the difficulty of global localization for significant viewpoint difference. Appearance-based localization methods tend to fail under large viewpoint changes. Recently, semantic graphs have been utilized to overcome the viewpoint variation problem. However, the methods are highly time-consuming, especially in large-scale environments. This leads to the second difficulty, which is how to perform real-time global localization. In this letter, we propose a semantic histogram based graph matching method that is robust to viewpoint variation and can achieve real-time global localization. Based on that, we develop a system that can accurately and efficiently perform MR-GL for both homogeneous and heterogeneous robots. The experimental results show that our approach is about 30 times faster than Random Walk based semantic descriptors. Moreover, it achieves an accuracy of 95% for global localization, while the accuracy of the state-of-the-art method is 85%.

A new extreme learning machine (ELM) localization technique that uses received signal strength indicator fingerprints only is proposed for multifloor environments. This structured scheme forms multiple individual ELMs for the floors as well as for the geographically formed data clusters of each floor. Multifloor environments often have huge amount of training and online measurement data. To maximize efficiency, we develop a data preprocessing algorithm, aiming to: 1) efficiently extract out only the essential information from the vast amount of data sets and reduce the data dimension and 2) transform the floor-level data sets and positioning data sets of each floor into a proper structure that is suitable for the proposed ensemble ELM technique. The proposed solution is unique in that its offline phase exploits multiple individual ELMs for all floors to generate a set of floor-level classification functions with the preprocessed training data sets, and for each floor, it exploits multiple ELMs for the data clusters to generate a set of position regression functions. The online phase executes a coarse localization step to estimate the floor by using the floor-level classification functions and a refined step to estimate the position on the floor by using the position regression functions. The proposed algorithm and several existing algorithms are implemented to perform localization using the same measured datasets in a multistory building. For both floor estimation and localization on the floor, it outperforms existing schemes. For most cases, the performance gap is substantial.

Finding a small object (e.g., earbuds, keys or a wallet) in an indoor environment (e.g., in a house or an office) can be frustrating. In this paper, we propose an innovative system, called HyperEar , to localize such an object using only a single smartphone, based on enhanced time-difference-of-arrival (TDoA) measurements over acoustic signals issued from the object. One major challenge is the hardware limitations of a Commercial-Off-The-Shelf (COTS) phone with a short separation between the two microphones and the low sampling rate of such microphones. HyperEar enhances the accuracy of TDoA measurements by virtually increasing distances between microphones through sliding the phone in the air. HyperEar requires no communication for synchronization between the phone and the object and is a low-cost and easy-to-use system. We evaluate the performance of HyperEar via extensive experiments in various indoor conditions and the results demonstrate that, for an object of 7 m away, HyperEar can achieve a mean localization accuracy of about 15 cm when the object in normal indoor environments.

Developments in sensing technology have shown that UWB radars are becoming increasingly valuable sensing devices that can be used for monitoring of humans in military/police and civilian areas. It is known that the applicability of particular methods of human localization depends on the character of persons’ motion. With respect to this finding, the researchers’ attention has been aimed at two fundamental directions. While the former is focused on the localization of moving persons (MP), the latter approach is intended to localize static persons (SP). Then, a proper fusion of the methods developed for MP and SP localization allows monitoring of persons moving with an unknown time-variable character of motion (MP-SP). The analyses of the currently known methods of MP and SP localization in terms of their use for MP-SP localization have shown that while MP localization methods are in principle well developed, SP localization methods are not sufficiently adapted for their use in MP-SP monitoring. Motivated by these findings, we would like to introduce a new radar signal processing procedure for SP localization (SPL) that could be an efficient component of algorithms to be applied for MP-SP monitoring. The novel features of the proposed SPL consist especially in a new approach to SP detection and inclusion of SP tracking in SPL. Moreover, SPL is characterized by relatively low computational complexity and is, therefore, suitable for real-time implementation. Experimental results have shown that SPL introduced in this paper provides very good performance for multiple person localization for line-of-sight and through-the-wall scenarios.

Existing interacting multiple models (IMMs) are limited by the time delay in responding to system model jumps due to the nature of the soft hand-off algorithm that interacts among subfilters. To address this issue, a novel method for deep-learning-aided localization of a multimodel system is proposed in this paper. The main contribution of the proposed algorithm is that a mode estimation network based on a bidirectional long short-term memory network (BiLSTM) is newly proposed to quickly and accurately estimate the multimodal system mode, which minimizes the delay. In addition, a federated Kalman filter with a selective reinitialization algorithm from the proposed BiLSTM is proposed for better estimation of multimodal systems. Simulation and flight test results of a UAV demonstrate that the proposed algorithm yields better localization performance than the conventional IMM algorithm because the proposed mode estimation network has fast and accurate mode detection.

Smartphone-based indoor localization has attracted considerable attentions in both research and industrial areas. However, the localization accuracy and robustness are still challenging problems due to low-cost noisy devices, especially in those complicated localization environments. Considering that pedestrian dead-reckoning (PDR) devices are widely equipped in recent smartphones, we propose a novel indoor localization fusing algorithm that integrates both wireless fidelity (WiFi) features and PDR features. By formulating the fusing indoor localization as a recursive function approximation problem, a sliding-window-based displacement scheme is designed to generate a time-series-based feature data set. We further apply the long short-term memory (LSTM) network for data fusion and localization on this data set by taking advantage of its benefits in time-series prediction and characterization. To evaluate the performance of the proposed algorithm, we compare it with state-of-the-art filter-based localization algorithms in three typical movements and three postures of holding smartphones. Extensive experiment results demonstrate the accuracy and robustness of the proposed algorithm in indoor localization, even in some extreme environments.

Localization without Global Navigation Satellite Systems (GNSS) is a critical functionality in autonomous operations of unmanned aerial vehicles (UAVs). Vision-based localization on a known map can be an effective solution, but it is burdened by two main problems: places have different appearance depending on weather and season, and the perspective discrepancy between the UAV camera image and the map make matching hard. In this work, we propose a localization solution relying on matching of UAV camera images to georeferenced orthophotos with a trained convolutional neural network model that is invariant to significant seasonal appearance difference (winter-summer) between the camera image and map. We compare the convergence speed and localization accuracy of our solution to six reference methods. The results show major improvements with respect to reference methods, especially under high seasonal variation. We finally demonstrate the ability of the method to successfully localize a real UAV, showing that the proposed method is robust to perspective changes., Published in IEEE Robotics and Automation Letters (Volume: 7, Issue: 4, October 2022)

Mobile opportunistic networks (MON) has been used for provisioning delay-tolerant applications. In MON the device communicates with each other with no assured end-to-end paths from source and destination because of frequent topology changes, node mobility, low density, and intermittent connectivity. In MON the device battery drains very fast for performing activities such as scanning, transceiver, and other computational processes, impacting the overall performance thus, designing energy-efficient routing is a challenging task. The routing employs a store-carry-and-forward mechanism for packet communication, where the packet is composed of time-to-live (TTL) and is kept in buffer till the opportunity arises. In improving delivery ratio message replication has been adopted; however, induces high network congestion. Here we present a location-aware hybrid microscopic routing (LAHMR) scheme for MON. The LAHMR provides an effective packet transmission scheme with location awareness and high reliability by limiting unnecessary packets being circulated in the network. Experiment outcome shows the LAHMR scheme achieves a much better delivery ratio with less delay, and also reduces the number of a forwarder for transmitting a packet, aiding in the reduction of network overhead concerning recent routing method namely the social-aware reliable forwarding (SCARF) technique.

There are about 253 million people with visual impairment worldwide. Many of them use a white cane and/or a guide dog as the mobility tool for daily travel. Despite decades of efforts, electronic navigation aid that can replace white cane is still research in progress. In this paper, we propose an RGB-D camera based visual positioning system (VPS) for real-time localization of a robotic navigation aid (RNA) in an architectural floor plan for assistive navigation. The core of the system is the combination of a new 6-DOF depth-enhanced visual-inertial odometry (DVIO) method and a particle filter localization (PFL) method. DVIO estimates RNA's pose by using the data from an RGB-D camera and an inertial measurement unit (IMU). It extracts the floor plane from the camera's depth data and tightly couples the floor plane, the visual features (with and without depth data), and the IMU's inertial data in a graph optimization framework to estimate the device's 6-DOF pose. Due to the use of the floor plane and depth data from the RGB-D camera, DVIO has a better pose estimation accuracy than the conventional VIO method. To reduce the accumulated pose error of DVIO for navigation in a large indoor space, we developed the PFL method to locate RNA in the floor plan. PFL leverages geometric information of the architectural CAD drawing of an indoor space to further reduce the error of the DVIO-estimated pose. Based on VPS, an assistive navigation system is developed for the RNA prototype to assist a visually impaired person in navigating a large indoor space. Experimental results demonstrate that: 1) DVIO method achieves better pose estimation accuracy than the state-of-the-art VIO method and performs real-time pose estimation (18 Hz pose update rate) on a UP Board computer; 2) PFL reduces the DVIO-accrued pose error by 82.5% on average and allows for accurate wayfinding (endpoint position error ≤ 45 cm) in large indoor spaces.

Proprioceptive localization refers to a new class of robot egocentric localization methods that do not rely on the perception and recognition of external landmarks. These methods are naturally immune to bad weather, poor lighting conditions, or other extreme environmental conditions that may hinder exteroceptive sensors such as a camera or a laser ranger finder. These methods depend on proprioceptive sensors such as inertial measurement units (IMUs) and/or wheel encoders. Assisted by magnetoreception, the sensors can provide a rudimentary estimation of vehicle trajectory which is used to query a prior known map to obtain location. Named as graph-based proprioceptive localization (GBPL), we provide a low cost fallback solution for localization under challenging environmental conditions. As a robot/vehicle travels, we extract a sequence of heading-length values for straight segments from the trajectory and match the sequence with a pre-processed heading-length graph (HLG) abstracted from the prior known map to localize the robot under a graph-matching approach. Using the information from HLG, our location alignment and verification module compensates for trajectory drift, wheel slip, or tire inflation level. We have implemented our algorithm and tested it in both simulated and physical experiments. The algorithm runs successfully in finding robot location continuously and achieves localization accurate at the level that the prior map allows (less than 10m)., 13 pages, 32 figures

Radio channel state information (CSI) measured with many receivers is a good resource for localizing a transmitter device with machine learning with a discriminative model. However, CSI localization is nontrivial when the radio map is complicated, such as in building corridors. This article introduces a view-selective deep learning (VSDL) system for indoor localization using CSI of WiFi. The multiview training with CSI obtained from multiple groups of access points (APs) generates latent features on a supervised variational deep network. This information is then applied to an additional network for dominant view classification to minimize the regression loss of localization. As noninformative latent features from multiple views are rejected, we can achieve a localization accuracy of 1.28 m, which outperforms by 30% the best known accuracy in practical applications in a complex building environment. To the best of our knowledge, this is the first approach to apply variational inference and to construct a practical system for radio localization. Furthermore, our work investigates a methodology for supervised learning with multiview data where informative and noninformative views coexist.

The localization accuracy is critical for the development of future autonomous systems and location-based services. The accuracy level for localization is difficult to achieve in the case of urban and GPS denied environments due to high scattering. Fingerprint-based localization techniques promise to address these challenges. However, this technique demands to build a radio map before localization, which is a time-consuming and labor-intensive task. This article designs a crowd-sourced based localization system to address the radio map building problem in fingerprinting localization system. In this method, the first initial radio map is constructed from the path-loss RSS model, followed by the update of the fingerprints with crowd-sourcing. Finally, the vehicle location is estimated from the RSS sample by matching it with an updated radio map with a deep learning algorithm. The main advantage of the proposed approach is the calibration-free crowd-sourced fingerprint generation and its applicability in various location-based services in urban infrastructure.

In the context of positioning an agent with a single-anchor, this contribution focuses on the Fisher information about the position, orientation and clock offset of the agent provided by single-bounce reflections. The availability of prior knowledge of the agent’s environment is taken into account via a prior distribution of the position of virtual anchors, and the rank, intensity and direction of provided information is studied. We show that when no prior knowledge is available, single-bounce reflections offer position information in the direction parallel to the reflecting surface, irrespective of the agent’s and anchor’s locations. We provide a geometrically intuitive explanation of the results and present numerical examples demonstrating their potential implications.

For indoor localization system, the integration of various fingerprints could intuitively improve localization accuracy since different fingerprints complement each other. In spite of the potential benefits, however, the limitation of applying various fingerprints in improving localization accuracy still remains unknown. Moreover, how to design efficient indoor localization methods through fully exploiting the features of different fingerprints is to be explored as well. In this work, we investigate the location error of a fingerprint-based indoor system with the application of hybrid fingerprints. It manifests that the location error is dependent on correlation coefficient of different types of fingerprints. For instance, the exploiting correlation between different types of fingerprints is shown to effectively reduce the location error, which gradually decreases with the number of adopted fingerprints. On this basis, we propose a hybrid received signal strength (RSS) and channel state information (CSI) localization algorithm (HRC), which is designed based on deep learning. The HRC fully exploits quick construction of fingerprint database with coarse-grained RSS and rich multipath information of fine-grained CSI. The RSS and CSI with high correlation are selected to construct fingerprint database, aiming to improve localization accuracy. Moreover, a deep auto-encoder is used to reduce the computation complexity and the deep neural network is trained for location estimation. Experimental results, which are obtained by designed indoor localization system, validate that the location error of HRC can be reduced by 77.3% and 20.3%, compared with the existing localization methods and HRC without RSS/CSI selection by correlation coefficient, respectively.

Wireless fidelity received signal strength (RSS) fingerprints are widely used for localization. However, device heterogeneity and temporal variation in the RSS values are the bottlenecks for accurate localization in the Internet of Things (IoT) networks. This letter addresses these challenges by proposing a novel differential $r$ -vectors which are the device-invariant signature of a particular location and are invariant, even with temporal RSS variations. The proposed network addresses the vanishing gradient problem which occurs in existing deep learning based localization methods, and hence, the localization accuracy improves. We evaluate the proposed method on two real-world datasets comprising device heterogeneity and temporal RSS variations. Additionally, the proposed method outperforms state-of-the-art fingerprinting-based methods that address these issues.

Flooding-based Denial-of-service (DoS) attacks have been prevalent in Network-on-Chip (NoC) architectures, due to its shared nature and open access to all the on-chip modules. A Malicious Intellectual Property (MIP) within a System-on-Chip (SoC) creates such an attack by flooding the NoC with useless packets resulting in significant bandwidth reduction. Finding the location of an MIP is crucial to restore regular network operations and curtail system performance degradation. In this work, we propose Sniffer, an efficient MIP localization framework which employs a low-overhead machine learning approach to accurately trace the attack path and take a collective decision to locate the MIPs. Experimental results show that Sniffer is able to provide high accuracy for MIP localization without incurring significant overheads.

Dynamic environments such as urban areas are still challenging for popular visual-inertial odometry (VIO) algorithms. Existing datasets typically fail to capture the dynamic nature of these environments, therefore making it difficult to quantitatively evaluate the robustness of existing VIO methods. To address this issue, we propose three contributions: firstly, we provide the VIODE benchmark, a novel dataset recorded from a simulated UAV that navigates in challenging dynamic environments. The unique feature of the VIODE dataset is the systematic introduction of moving objects into the scenes. It includes three environments, each of which is available in four dynamic levels that progressively add moving objects. The dataset contains synchronized stereo images and IMU data, as well as ground-truth trajectories and instance segmentation masks. Secondly, we compare state-of-the-art VIO algorithms on the VIODE dataset and show that they display substantial performance degradation in highly dynamic scenes. Thirdly, we propose a simple extension for visual localization algorithms that relies on semantic information. Our results show that scene semantics are an effective way to mitigate the adverse effects of dynamic objects on VIO algorithms. Finally, we make the VIODE dataset publicly available at https://github.com/kminoda/VIODE., Comment: 8 pages, 8 figures, accepted for publication in the IEEE Robotics and Automation Letters (RA-L) January 31, 2021. IEEE Xplore link: https://ieeexplore.ieee.org/document/9351597/

International audience; We present a novel FastSLAM approach for a robotic system inspecting structures made of large metal plates. By taking advantage of the reflections of ultrasonic guided waves on the plate boundaries, it is possible to recover, with enough precision, both the plate shape and the robot trajectory. Contrary to our previous work, this approach takes into account the dispersive nature of guided waves in metal plates. This is leveraged to construct beamforming maps from which we solve the mapping problem through plate edges estimation for every particle, in a FastSLAM fashion. It will be demonstrated, with real acoustic measurements obtained on different metal plates, that such a framework achieves better results in terms of convergence and accuracy, while the complexity of the algorithm is sensibly reduced.

Simultaneous localization and object tracking (SLOT) is essentially important for autonomous systems. Tightly-coupled and loosely-coupled methods are two commonly used back-end frameworks for the state-of-the-art solutions of SLOT problem. However, some inherent limitations exist in these two frameworks. In particular, the tightly-coupled method is usually disturbed by the poor observations of some dynamic objects, and the performance of a loosely-coupled one completely depends on that of classical static simultaneous localization and mapping (SLAM) process. Motivated by these observations, we propose a novel switching-coupled back-end solution and theoretically derive its concrete form using probability representation. Based on the switching strategy and the proposed objects classification criteria where the object uncertainty, observation quality and prior information are jointly considered, the dynamic objects’ states are flexibly coupled with camera's state and static landmarks’ states. For implementation, the measurement constraints of “good” dynamic objects and static landmarks are simultaneously leveraged to perform SLAM and good object tracking (SLAMGOT) process, and those of “bad” objects are used for bad object tracking (BOT) process based on the obtained camera state. Extensive evaluations on synthetic scenes, KITTI datasets and real-world experiments demonstrate the performance of the proposed method.

Proprioceptive localization (PL) refers to robot or vehicle egocentric localization methods that do not rely on the perception and recognition of external landmarks. These methods depend on a prior map and proprioceptive sensors such as inertial measurement units and/or wheel encoders. PL is intended to be a low-cost and fallback solution when everything else fails due to bad weather or poor environmental conditions. With the development of communication technology, vehicle-to-vehicle (V2V) communication enables information exchange between vehicles. It becomes possible to leverage V2V communication to develop a multiple vehicle/robot collaborative localization scheme. Named as collaborative graph-based proprioceptive localization (C-GBPL), we extract heading-length sequence from the trajectory as features. When rendezvousing with other vehicles, the ego vehicle aggregates the features from others and forms a merged query graph. We match the query graph with a pre-processed heading-length graph (HLG) abstracted from a prior map to localize the vehicle under a graph-to-graph matching approach. We have implemented our algorithm and tested it in both simulated and physical experiments. The C-GBPL algorithm significantly outperforms its single-vehicle counterpart in localization speed and robustness to trajectory and map degeneracy.

Localization is a key task for autonomous vehicles. It is often solved with GNSS but due to multipath the performance is often not sufficient. Feature localization systems using LiDAR can deliver an accurate localization but the creation of the necessary feature maps is an effortful task. With digitization of urban planning processes a lot of street level data is being generated and increasingly becomes openly available. We propose a novel feature localization system which utilizes geometric fingerprinting to robustly associate features to a feature map generated from this open data from the city of Berlin. With this association, we perform a precise localization of a vehicle in areas spanning over several square kilometers using an optional IMU, the vehicle's CAN-odometry and an initial pose estimate. We evaluated our system with our autonomous car in real world scenarios and achieved a centimeter precision localization accuracy outperforming a high-cost GNSS. The source code will be published at https://github.com/dcmlr/fingerprint-localization .