Search

Over the past two decades, technological developments in the robotics field have made it possible to gather data from areas that are hostile and not-reachable environments for humans. In line with these advancements in data collecting tools and procedures, the need and demand for computationally efficient methods for processing the gathered data have been increased from different science and engineering fields. Among the others, optical data is one of the main data sources that low-cost robotic vehicles can obtain easily nowadays. Due to the different limitations, obtained optical data usually cannot cover a large area in a single image. Therefore, optical mapping methods (image mosaicing) are needed to create higher resolution maps by combining comparatively smaller resolution images. These methods rely on pairwise image registration and one of the main bottlenecks in the case of image-only information available is that the quadratic growth of image matching attempts with respect to the total number of images. In this paper, we propose an algorithmically improved end-to-end framework for creating 2D optical maps from a set of randomly ordered images with the aim of reducing computational efforts needed via lowering the total number of image matching attempts. We present extensive and comparative experimental results with its counterpart approach using four real datasets obtained from the underwater environment using Unmanned Underwater Vehicles (UUVs).

Mapping is an essential part for the adoption of robots in agricultural and forestry environments. Providing the robot with the ability to map its surroundings, facilitates its navigation and is necessary for implementing obstacle avoidance, without human interference. Herein we present the challenges of outdoor environments, present an overview of existing mapping frameworks and then evaluate their suitability for field applications. Two widely used mapping frameworks, OctoMap and RTAB-Map are analyzed within the Robot Operating System (ROS) ecosystem and a parametric study is carried out in order to assess their performance, under both simulated and real-world constraints. Finally this work aims to be utilized as a deployment reference guide for mobile robotic applications in outdoor environments.

Three-dimensional (3D) maps are indispensable for autonomous robot navigation in outdoor environments. Loop closure, which is a key technology in robotic mapping, is essential for creating consiste...

Robotic mapping is the problem of inferring a representation of a robot’s surroundings using noisy measurements as it navigates through an environment. As robotic systems move toward more challenging behaviors in more complex scenarios, such systems require richer maps so that the robot understands the significance of the scene and objects within. This dissertation focuses on semantic-aware robotic mapping in unknown, loosely structured environments. The first main contribution is a Bayesian kernel inference semantic mapping framework that formulates a unified probabilistic model for occupancy and semantics, and provides a closed-form solution for scalable dense semantic mapping. This framework significantly reduces the computational complexity of learning-based continuous semantic mapping and achieves high accuracy in the meantime. Next, a novel and flexible multi-task multi-layer Bayesian mapping framework is proposed to provide even richer environmental information. A two-layer robotic map of semantics and traversability is built as a strong example. Moreover, it is readily extendable to include more layers according to needs. Both mapping algorithms were verified using publicly available datasets or through experimental results on a Cassie-series bipedal robot. Finally, instead of modeling the terrain traversability using metrics defined by domain knowledge, an energy-based deep inverse reinforcement learning method is developed to learn the traversability from demonstrations. The proposed method considers robot proprioception and can learn reward maps that lead to more energy-efficient future trajectories. Experiments are conducted using a dataset collected by a Mini-Cheetah robot in a campus environment with various scenes.

We present a novel 3D mapping method leveraging the recent progress in neural implicit representation for 3D reconstruction. Most existing state-of-the-art neural implicit representation methods are limited to object-level reconstructions and can not incrementally perform updates given new data. In this work, we propose a fusion strategy and training pipeline to incrementally build and update neural implicit representations that enable the reconstruction of large scenes from sequential partial observations. By representing an arbitrarily sized scene as a grid of latent codes and performing updates directly in latent space, we show that incrementally built occupancy maps can be obtained in real-time even on a CPU. Compared to traditional approaches such as Truncated Signed Distance Fields (TSDFs), our map representation is significantly more robust in yielding a better scene completeness given noisy inputs. We demonstrate the performance of our approach in thorough experimental validation on real-world datasets with varying degrees of added pose noise., Comment: 3DV 2021. Equal contribution between the first two authors. Code: https://github.com/ethz-asl/neuralblox

For robots to navigate and interact more richly with the world around them, they will likely require a deeper understanding of the world in which they operate. In robotics and related research fields, the study of understanding is often referred to as semantics, which dictates what does the world "mean" to a robot, and is strongly tied to the question of how to represent that meaning. With humans and robots increasingly operating in the same world, the prospects of human-robot interaction also bring semantics and ontology of natural language into the picture. Driven by need, as well as by enablers like increasing availability of training data and computational resources, semantics is a rapidly growing research area in robotics. The field has received significant attention in the research literature to date, but most reviews and surveys have focused on particular aspects of the topic: the technical research issues regarding its use in specific robotic topics like mapping or segmentation, or its relevance to one particular application domain like autonomous driving. A new treatment is therefore required, and is also timely because so much relevant research has occurred since many of the key surveys were published. This survey therefore provides an overarching snapshot of where semantics in robotics stands today. We establish a taxonomy for semantics research in or relevant to robotics, split into four broad categories of activity, in which semantics are extracted, used, or both. Within these broad categories we survey dozens of major topics including fundamentals from the computer vision field and key robotics research areas utilizing semantics, including mapping, navigation and interaction with the world. The survey also covers key practical considerations, including enablers like increased data availability and improved computational hardware, and major application areas where..., 81 pages, 1 figure, published in Foundations and Trends in Robotics, 2020

In recent years, place recognition techniques have been extensively studied in the domain of robotic mapping, referred to as appearance-based mapping. Nonetheless, the majority of these methods focus the challenges of place recognition in offline or supervised scenarios, which in certain conditions, e.g., unknown environments, is infeasible. In this paper, we address the challenges of online place recognition and demonstrate the general applicability of our approach in versatile environments. To this end, a modified growing self-organizing network of neurons is proposed, which incrementally adapts itself to learn the topology of the perceptual space formed by gist features. Given a query image and the network state at any time instant, the expected activity of the network is estimated using a proposed Bayesian framework, and the current place is categorized as familiar or novel. Exhaustive experiments on 11 challenging sequences signify the strength of our algorithm for a reliable and real-time place recognition on routes as large as 18 km. Compared to many state-of-the-art approaches, our method does not need offline training or environment-specific parameter tuning.

In this paper, a novel 2-D laser-based autonomous exploration approach for mobile robots is proposed, which is based on a novel polygon map construction approach and graph-based simultaneous localization and mapping (SLAM) with directional endpoint features. This approach is composed of three modules: graph-based SLAM using directional endpoint features, polygon map construction, and exploration. Different from existing approaches in the field of 2-D SLAM, the newly proposed 2-D graph-SLAM is based on 3-D “directional endpoint” features; on this basis, a well-known data structure “circular-doubly linked list” is applied to construct a novel polygon map for navigation. Note that it is efficient for circular-doubly linked list to initialize and update the polygon map. In addition, we propose a new information entropy calculation approach to quantify the entropy of the polygon map. Then for each candidate goal, we could obtain corresponding information gain and make next decision through collision detection. Comparative experimental results with respect to the well-known Gmapping and Karto SLAM are presented to show superior performance of the proposed graph-based SLAM. The autonomous exploration experiments in the office and hallway environments show the effectiveness of the proposed approach for robotic mapping and exploration tasks. Note to Practitioners —This paper is motivated by the challenges of autonomous exploration for mobile robots. We suggest a novel autonomous exploration approach through directional endpoint feature-based graph simultaneous localization and mapping (SLAM) and polygon map construction. This newly proposed approach does not require any artificial landmark or underlying occupancy grid map, which is easy to implement. The experiments are carried out in the office and hallway environments, including graph-based SLAM and autonomous exploration for mobile robots. The experimental results show the effectiveness of the proposed autonomous exploration framework. In the future research, we will address autonomous exploration in more challenging dynamic environments.

Layout of the interior works during building construction is time consuming and error prone. Given the cost involved, both for initial layout and for later rework where errors occur, researchers have sought to automate the layout task. Some have adopted marker-less augmented reality methods using heads-up displays or cameras, and others have proposed robots capable of marking out the works. The former encumber the workers, the latter are expensive to set up and are sensitive to site conditions, and neither has yet achieved the required accuracy. In this paper, we propose a more efficient approach to project relevant information from a building information model (BIM) onto the construction surface, directly augmenting the construction site with the design information. This is done using a portable system consisting of a 2-D laser scanner, an angled adjustable projector, and a camera. The system localizes itself within the already built outer walls using the laser scanner and the BIM model using a method derived from robotic mapping; it calibrates the projection correction parameters (keystone correction) using image analysis; and it projects the information with the angled projector. Testing results showed that the localization was accurate within a few millimeters and less than 3°, and the final projected image’s error was approximately 1 cm. Initial calibration requires less than 1 min and does not require specialist skills. The system automates the layout task, preserves accuracy, and can provide rich model information on any interior surface. Note to Practitioners —Measuring and marking up of the interior works during building construction is time consuming and error prone. Layout must be repeated for each trade: partitions, false ceilings, mechanical, electrical, and plumbing systems, and flooring and furnishings all require accurate marking of locations on floors, walls, and ceiling surfaces. Current automation is limited to the use of robotic total stations, but these only locate specific predetermined points. We propose a simple solution for automated layout, in which images from a building information model are projected directly onto the work surface. The system projects any desired information—drawings and images—onto the work surface (floor, walls, or ceiling) in the correct location, scale, and orientation. The prototype apparatus consists of a laser range scanner, a projector, and a camera. Projection of the work instructions directly onto the work surface is accurate and immediate. It saves the time required for workers to interpret and then mark up the dimensional information and it avoids the human error involved. The prototype is limited to environments in which computer projection is practical and currently requires planar surfaces.

Scale-invariant feature transform (SIFT) is one of the widely used interest point features. It has been successfully applied in various computer vision algorithms like object detection, object tracking, robotic mapping, and large-scale image retrieval. Although SIFT descriptor is highly robust towards scale and rotation variations, the high computational complexity of the SIFT algorithm inhibits its use in applications demanding real-time response and in algorithms dealing with very large-scale databases. In order to be effective for image matching process in near real-time, the Compute Unified Device Architecture (CUDA) application programming interface of a graphics processing unit (GPU) is incorporated to speed up or improve the SIFT method. Experimental results show that the proposed GPU-based SIFT framework is suitable for image application in real time. It can improve the image matching process both in time and accuracy compared with conventional SIFT method.

Mobile robot navigation is a well researched discipline, looking back to a relatively long history however it is still a rich, active area for research and development. The ultimate goal for robots and intelligent vehicles seems to be autonomous navigation in complex real life scenarios. In order to achieve higher levels of autonomy sophisticated sensors and a sound understanding of the robot and its interaction with the environment is needed. The tasks involved can be divided to two basic categories; internal tasks involve keeping track of internal dynamic parameters, like speed accelerations, internal states etc. On the other hand the vehicle needs to be aware of external factors like obstacles, points of interest, possible routes from a to b and the respective costs. This is generally called robotic mapping. Intelligent Vehicle stands for a vehicle that senses the environment and provides information or control to assist the driver in optimal vehicle operation. Intelligent Vehicle systems operate at the tactical level of driving (throttle, brakes, steering) as contrasted with strategic decisions such as route choice, which might be supported by an on-board navigation system. (Bishop, 2005) Optical sensors supply by far the most information and as greater and greater processing capabilities become readily available their use becomes more widespread. Many researchers and companies have made more or less successful attempts at creating optical sensors for speed measurement, however to the knowledge of the authors no accurate high speed solution exist at the low price range. The aim of this article is to introduce a novel method for optical speed measurement and put it into perspective by summarizing other navigation methods and reviewing recent related work and possible applications. Also an introduction to optical flow calculation is given and practical considerations on texture analysis and sensor parameters are discussed backed up with simulation results.

In this chapter, we have presented, in short, the major algorithms in the field of mapping dynamic environments. We categorize the objects in the robot environments into three categories (static, low dynamic and high dynamic objects) and for each category, we present a number of solutions proposed so far in the literature. Special attention is paid on the occupancy grid structure and its variations that have been applied on dynamic environments giving promising results. Of course, it was not possible to cover all the available techniques or examine relative to mapping issues, such as localization and navigation. Summarizing, we can state that there exist a number of promising algorithms for mapping in dynamic environments. These algorithms are able to create valid maps of the static areas of the world and detect and model (low or high) dynamic objects in the robot environment. Nevertheless, a large number of challenging issues still remains to be solved. First of all, many algorithms that are presented in this work are not real-time. They make use of iterative techniques (e.g. EM, clustering) with execution times that depend on the size of the data. The necessity for real-time algorithms is obvious. A robot must be able to identify its environment rapidly in order to react on time. This need is more urgent while navigating outdoors. Outdoor mapping is an issue that remains open. Streets, parks and garages are some examples of places where a mobile robot can be of extreme usefulness. However, these environments pose new challenging issues. Their rate of change, complexity and size make many existing algorithms inapplicable and impose the need for new algorithms suitable for the special characteristics of such environments (some examples can be found in (Nuchter et al., 2007) and (Yoon et al., 2007)). In order to deal with these problems, an interesting direction that has to be investigated is the combination of computer vision with laser data. So far, only a few algorithms have applied techniques borrowed from the computer vision field e.g. (Anguelov et al., 2004), (Yoon et al., 2007). Vision, however, can provide a wealth of information about the state of the environment (such as humans walking by the robot or chairs in the middle of a room) and it can be really helpful in the task of object identification. The combination of vision and laser data can provide important information about the robot environment. Another way to gain more information about the objects that surround the robot is to acquire 3D laser range data of them e.g. (Hahnel et al., 2003b), (Ryde & Hu, 2006), (Harati & Siegwart, 2007). Humans possess a complete prior knowledge of the model of the objects in the environment and can easily distinguish two different objects even when they are partially observed in contrast to modern robots. To bridge the gap between human perception and robot perception of the environment, we can utilize 3D laser data that provide more information including information about the objects' shapes. To conclude, the problem of mapping in dynamic environments is a really challenging and extremely active research field in robotics. In the future, we will be able to design robots that will be capable of moving among humans in their own environments without interrupting human activities. In order to do that, a milestone that has to be reached is a complete solution to the problem of mapping dynamic environment.

In digital agriculture, large-scale data acquisition and analysis can improve farm management by allowing growers to constantly monitor the state of a field. Deploying large autonomous robot teams to navigate and monitor cluttered environments, however, is difficult and costly. Here, we present methods that would allow us to leverage managed colonies of honey bees equipped with miniature flight recorders to monitor orchard pollination activity. Tracking honey bee flights can inform estimates of crop pollination, allowing growers to improve yield and resource allocation. Honey bees are adept at maneuvering complex environments and collectively pool information about nectar and pollen sources through thousands of daily flights. Additionally, colonies are present in orchards before and during bloom for many crops, as growers often rent hives to ensure successful pollination. We characterize existing Angle-Sensitive Pixels (ASPs) for use in flight recorders and calculate memory and resolution trade-offs. We further integrate ASP data into a colony foraging simulator and show how large numbers of flights refine system accuracy, using methods from robotic mapping literature. Our results indicate promising potential for such agricultural monitoring, where we leverage the superiority of social insects to sense the physical world, while providing data acquisition on par with explicitly engineered systems.

Asset inspection of large structures such as storage tanks in the oil, gas and petrochemical industry is challenging, either requiring labour-intensive manual measurements or using robotic deployment to make the measurements. Current robotic systems employ point-by-point scanning, which is time-consuming. Using guided waves for such inspections is attractive as they provide a mechanism for monitoring the inaccessible areas and simultaneously providing structural location data to speed up the inspection process. In this research, shear horizontal (SH) guided waves generated by electromagnetic acoustic transducers (EMATs) are used to screen a large area using a crawler. EMATs with 22 mm wavelength are used to generate the first two SH modes: a non-dispersive SH0 and highly dispersive SH1 on a 10 mm thick steel sample. Previously, we have demonstrated the feasibility of guided wave-based occupancy grid mapping (GW-OGM) for mapping a structure’s edges. In this work, the GW-OGM technique is generalised to identify and estimate the location of a flat bottom hole in a pitch-catch mode. The simulation and empirical data demonstrate that the location of damage can be identified as the robot navigates on the component, with full coverage. Moreover, the simulated data are in good agreement with the experimental results on the generation of SH wave modes.

[EN] Autonomous navigation is a complex problem that involves different tasks, such as location of the mobile robot in the scenario, robotic mapping, generating the trajectory, navigating from the initial point to the target point, detecting objects it may encounter in its path, etc. This paper presents a new optimal trajectory planning algorithm that allows the assessment of the energy efficiency of autonomous light vehicles. To the best of our knowledge, this is the first time in the literature that this is carried out by minimizing the travel time while considering the vehicle's dynamic behavior, its limitations, and with the capability of avoiding obstacles and constraining energy consumption. This enables the automotive industry to design environmentally sustainable strategies towards compliance with governmental greenhouse gas (GHG) emission regulations and for climate change mitigation and adaptation policies. The reduction in energy consumption also allows companies to stay competitive in the marketplace. The vehicle navigation control is efficiently implemented through a middleware of component-based software development (CBSD) based on a Robot Operating System (ROS) package. It boosts the reuse of software components and the development of systems from other existing systems. Therefore, it allows the avoidance of complex control software architectures to integrate the different hardware and software components. The global maps are created by scanning the environment with FARO 3D and 2D SICK laser sensors. The proposed algorithm presents a low computational cost and has been implemented as a new module of distributed architecture. It has been integrated into the ROS package to achieve real time autonomous navigation of the vehicle. The methodology has been successfully validated in real indoor experiments using a light vehicle under different scenarios entailing several obstacle locations and dynamic parameters., This work has been partially funded by FEDER-CICYT project with reference DPI2017-84201-R financed by Ministerio de Economia, Industria e Innovacion (Spain).

Depth estimation plays a significant role in industrial applications, e.g. augmented reality, robotic mapping and autonomous driving. Traditional approaches for capturing depth, such as laser or depth sensor based methods, are difficult to use in most scenarios due to the limitations of high system cost and limited operational conditions. As an inexpensive and convenient approach, using the computational models to estimate depth from a single RGB image offers a preferable way for the depth prediction. Although the design of computational models to estimate the depth map has been widely investigated, the majority of models suffers from low prediction accuracy due to the sole utilization of a one-stage regression strategy. Inspired by both theoretical and practical success of two-stage regression, we propose a two-stage deep regression model, which is composed of two state-of-the-art network architectures, i.e. the fully convolutional residual network (FCRN) and the conditional generation adversarial network (cGAN). FCRN has been proved to possess a strong prediction ability for depth prediction, but fine details in the depth map are still incomplete. Accordingly, we have improved the existing cGAN model to refine the FCRN-based depth prediction. The experimental results show that the proposed two-stage deep regression model outperforms existing state-of-the-art methods.

Methods for autonomous navigation systems using sonars in air traditionally use the time-of-flight technique for obstacle detection and environment mapping. However, this technique suffers from constructive and destructive interference of ultrasonic reflections from multiple obstacles in the environment, requiring several acquisitions for proper mapping. This paper presents a novel approach for obstacle detection and localisation using inverse problems and compressed sensing concepts. Experiments were conducted with multiple obstacles present in a controlled environment using a hardware platform with four transducers, which was specially designed for sending, receiving and acquiring raw ultrasonic signals. A comparison between the performance of compressed sensing using Orthogonal Matching Pursuit and two traditional image reconstruction methods was conducted. The reconstructed 2D images representing the cross-section of the sensed environment were quantitatively assessed, showing promising results for robotic mapping tasks using compressed sensing.

Este proyecto evalúa los parámetros influyentes ante variaciones de iluminación en el funcionamiento de los algoritmos Filtro de Kalman y Filtro de partículas aplicados a SLAM (Mapeo y Localización Simultanea) llevando a cabo experimentos en entornos cerrados que permiten realizar pruebas con distintas intensidades de luz. Para esto se utiliza el robot móvil diferencial Turtlebot3 Burger acompañado del sensor Kinect V1, ya que posee un sensor de profundidad y una cámara RGB, esto permite llevar a cabo construcción de mapas 3D y evidenciar variaciones en los procesos de mapeo dependiendo de la iluminación, lo cual es un requerimiento de este trabajo. El filtro de partículas (PF) es uno de los algoritmos de estimación más adaptados para SLAM, al igual que el filtro de Kalman que interactúa con estados pasados, presentes y estimaciones futuras del sistema, siendo empleado para seguir sistemas estocásticos dinámicos mediante sensores ruidosos. Los algoritmos emplean un método de predicción y corrección, es decir, pronostica un nuevo estado a partir de su estimación previa agregando un factor de corrección de la misma proporción al error de predicción, de tal manera que este error se minimiza con cada iteración, lo cual lo hace adecuado para trabajar ante variaciones de iluminación ya que esta técnica demanda que se pueda reconstruir y localizar en el menor tiempo posible cualquier sistema autónomo bajo condiciones naturales. This project evaluates the parameters influencing light variations in the operation of the Kalman Filter and Particle Filter algorithms applied to SLAM (Simultaneous Mapping and Localization), carrying out experiments in closed environments that allow tests with different light intensities. For this, the Turtlebot3 Burger differential mobile robot is used accompanied by the Kinect V1 sensor, since it has a depth sensor and an RGB camera, this allows to carry out 3D map construction and show variations in the mapping processes depending on the lighting, which is a requirement of this job. The particle filter (PF) is one of the most adapted estimation algorithms for SLAM, as is the Kalman filter that interacts with past, present and future system estimates, being used to follow dynamic stochastic systems using noisy sensors. The algorithms use a prediction and correction method, that is, they forecast a new state from its previous estimate by adding a correction factor of the same proportion to the prediction error, in such a way that this error is minimized with each iteration, which which makes it suitable for working with lighting variations since this technique demands that any autonomous system can be rebuilt and located in the shortest possible time under natural conditions. Ingeniero Electronico http://unidadinvestigacion.usta.edu.co Pregrado

Traditional dense volumetric representations for robotic mapping make simplifying assumptions about sensor noise characteristics due to computational constraints. We present a framework that, unlike conventional occupancy grid maps, explicitly models the sensor ray formation for a depth sensor via a Markov Random Field and performs loopy belief propagation to infer the marginal probability of occupancy at each voxel in a map. By explicitly reasoning about occlusions our approach models the correlations between adjacent voxels in the map. Further, by incorporating learnt sensor noise characteristics we perform accurate inference even with noisy sensor data without ad-hoc definitions of sensor uncertainty. We propose a new metric for evaluating probabilistic volumetric maps and demonstrate the higher fidelity of our approach on simulated as well as real-world datasets., Paper to be published at RSS 2020

Robotic mapping and odometry are the primary competencies of a navigation system for an autonomous mobile robot. However, the state estimation of the robot typically mixes with a drift over time, and its accuracy is degraded critically when using only proprioceptive sensors in indoor environments. Besides, the accuracy of an ego-motion estimated state is severely diminished in dynamic environments because of the influences of both the dynamic objects and light reflection. To this end, the multi-sensor fusion technique is employed to bound the navigation error by adopting the complementary nature of the Inertial Measurement Unit (IMU) and the bearing information of the camera. In this paper, we propose a robust tightly-coupled Visual-Inertial Navigation System (VINS) based on multi-stage outlier removal using the Multi-State Constraint Kalman Filter (MSCKF) framework. First, an efficient and lightweight VINS algorithm is developed for the robust state estimation of a mobile robot by practicing a stereo camera and an IMU towards dynamic indoor environments. Furthermore, we propose strategies to deal with the impacts of dynamic objects by using multi-stage outlier removal based on the feedback information of estimated states. The proposed VINS is implemented and validated through public datasets. In addition, we develop a sensor system and evaluate the VINS algorithm in the dynamic indoor environment with different scenarios. The experimental results show better performance in terms of robustness and accuracy with low computation complexity as compared to state-of-the-art approaches.

Mapping the interior of the building is the first step in order to obtain the description of that building. In this work, an autonomous robot is designed and built. The robot is able to navigate through an unknown environment, avoiding obstacles and simultaneously build a 3D map on its own. This autonomous feature is implemented using minimal and economically viable hardware such as Kinect v2 sensor and Arduino. Also, an open-source platform Robotic Operating System is used, on top of which the mapping and navigation algorithms are implemented. The 3D map built can be used with a lot of existing devices which involves localization. Post-processing procedures are used to obtain point clouds for building the 3D map. Emphasis is put on the efficient use of the available resources to generate the 3D maps of unknown indoor environment.

This paper presents an approach and introduces new open-source tools that can be used to evaluate robotic mapping algorithms. Also described is an extensive subterranean mine rescue dataset based upon the DARPA Subterranean (SubT) challenge including professionally surveyed ground truth. Finally, some commonly available approaches are evaluated using this metric.

This work details the development of an indoor navigation and mapping system using a non-central catadioptric omnidirectional camera and its implementation for mobile applications. Omnidirectional catadioptric cameras find their use in navigation and mapping of robotic platforms, owing to their wide field of view. Having a wider field of view, or rather a potential 360° field of view, allows the system to “see and move” more freely in the navigation space. A catadioptric camera system is a low cost system which consists of a mirror and a camera. Any perspective camera can be used. A platform was constructed in order to combine the mirror and a camera to build a catadioptric system. A calibration method was developed in order to obtain the relative position and orientation between the two components so that they can be considered as one monolithic system. The mathematical model for localizing the system was determined using conditions based on the reflective properties of the mirror. The obtained platform positions were then used to map the environment using epipolar geometry. Experiments were performed to test the mathematical models and the achieved location and mapping accuracies of the system. An iterative process of positioning and mapping was applied to determine object coordinates of an indoor environment while navigating the mobile platform. Camera localization and 3D coordinates of object points obtained decimetre level accuracies.

With the rapid advancement of laser scanning and photogrammetry technologies, frequent geometric data collection at construction sites by contractors has been increased for the purpose of improving constructability, productivity, and onsite safety. However, the conventional static laser scanning method suffers from operational limitations due to the presence of many occlusions commonly found in a typical construction site. Obtaining a complete scan of a construction site without information loss requires that laser scans are obtained from multiple scanning locations around the site, which also necessitates extra work for registering each scanned point cloud. As an alternate solution to this problem, this paper introduces an autonomous mobile robot which navigates a scan site based on a continuously updated point cloud map. This mobile robot system utilizes the 2D Hector Simultaneous Localization and Mapping (SLAM) technique to estimate real-time positions and orientations of the robot in the x-y plane. Then, the 2D localization information is used to create 3D point clouds of unknown environments in real time to determine its navigation paths as a pre-scanning process. The advantage of this framework is the ability to determine the optimal scan position and scan angle to reduce the scanning time and effort for gathering high resolution point cloud data in real-time. The mobile robot system is able to capture survey-quality RGB-mapped point cloud data, and automatically register the scans for geometric reconstruction of the site. The performance of the overall system was tested in an indoor environment and validated with promising results.

International audience; This paper proposes a new approach, interval Simultaneous Localization and Mapping (i-SLAM), which addresses the robotic mapping problem in the context of interval methods, where the robot sensor noise is assumed bounded. With no prior knowledge about the noise distribution or its probability density function, we derive and present necessary conditions to guarantee the map convergence even in the presence of nonlinear observation and motion models. These conditions may require the presence of some anchoring landmarks with known locations. The performance of i-SLAM is compared with the probabilistic counterparts in terms of accuracy and efficiency.

With the recent discovery of water-ice and lava tubes on the Moon and Mars along with the development of in-situ resource utilization (ISRU) technology, the recent planetary exploration has focused on rover (or lander)-based surface missions toward the base construction for long-term human exploration and habitation. However, a 3D terrain map, mostly based on orbiters’ terrain images, has insufficient resolutions for construction purposes. In this regard, this paper introduces the visual simultaneous localization and mapping (SLAM)-based robotic mapping method employing a stereo camera system on a rover. In the method, S-PTAM is utilized as a base framework, with which the disparity map from the self-supervised deep learning is combined to enhance the mapping capabilities under homogeneous and unstructured environments of planetary terrains. The overall performance of the proposed method was evaluated in the emulated planetary terrain and validated with potential results.

Simultaneous localization and environment mapping (SLAM) is the core to robotic mapping and navigation as it constructs simultaneously the unknown environment and localizes the agent within. However, in millimeter wave (mmWave) research, SLAM is still at its infancy. This paper consists a first of its kind in mapping an indoor environment based on the RSS, Time-Difference-of-Arrival, and Angle-of-Arrival measurements. We introduce MOSAIC as a new approach for SLAM in indoor environment by exploiting the map-based channel model. More precisely, we perform localization and environment inference through obstacle detection and dimensioning. The concept of virtual anchor nodes (VANs), known in literature as the mirrors of the real anchors with respect to the obstacles in the environment, is explored. Then, based on these VANs, the obstacles positions and dimensions are estimated by detecting the zone of paths obstruction, points of reflection, and obstacle vertices. Then, extended Kalman filter is adapted to the studied environment to improve the estimation of the points of reflection hence the mapping accuracy. Cramer-Rao lower bounds are also derived to find the optimal number of anchor nodes. Simulation results have shown high localization accuracy and obstacle detection using mmWave technology.