Simultaneous Localization and Mapping

We describe a method of generating and utilizing visual landmarks that is well suited for SLAM applications. The landmarks created are highly distinctive and reliably detected, virtually eliminating the data association problem present in other landmark schemes. Upon subsequent detections of a landmark, a 3-D pose can be estimated. The scheme requires a single camera.

In this paper the MRL rescue robot team and its robots are explained. We have designed and built three new robots including two autonomous indoor robots and one out-door (rough terrain) robot for different situations/arenas. Our main goal of this activity is to achieve a practical rescue robot for real situation such as earthquake which is quite common in our

În zilele noastre dezvoltarea tehnologiei este mai rapidă pe zi ce trece. Din ce în ce mai multe dispozitive robotizate autonome s-au infiltrat în viețile oamenilor pentru a o face mai usoară. În foarte multe cazuri, dispozitivele robotizate autonome pot fi cea mai bună opțiune pentru rezolvarea unei probleme. Un exemplu considerat este reprezentat de metodele de explorare și mapare a unui spațiu interior greu de accesat omului. Utilizarea unui robot mobil scade riscul accidentarilor sau accidentelor umane și poate returna unui utilizator date exacte despre mediul in care acesta a fost trimis. Echipat cu dispozitive senzoriale speciale, vehiculul autonom sau comandat de un utilizator va scana mediul și va trimite informații prețioase. În șcenarii precum incendii, clădiri afectate sau zone lipsite de orice sursă de lumină  un astfel de robot este soluția optimală De asemenea, pot exista și locuri în care utilizarea unui robot este singura modalitate viabilă de acces. De exemplu, centralele nucleare, fabricile de produse chimice toxice sau orice mediu ce ar putea fi dăunător omului, pot fi explorate de vehicule comandate de la distaţă pentru realizarea diferitelor operațiuni. Realizarea unei harți ne poate fi de mare ajutor în aceste situații in care pentru un individ poate fi foarte periculos sau imposibil chiar să ajungă în zona respectivă. Problema dezvoltării unui robot mobil de acest gen implică stăpânirea conceptului de localizarea şi mapare simultană. În robotică localizarea și maparea simultană (SLAM - simultaneous localization and mapping) reprezintă problema computării, construirii şi actualizării unei hărți a unui mediu necunoscut, urmărind simultan localizarea unui agent în cadrul acesteia. Inţial această problema pare a fi o problema de tipul „oul sau găina?” intrucât pentru a ne putea localiza într-un loc avem nevoie de o hartă a acelui loc și, viceversa , pentru a putea face o hartă a unui loc trebui să știm unde ne aflăm relativ de acel loc. Există mai mulți algoritmi cunoscuți pentru rezolvarea acestei probleme, cel puțin aproximativ si în timp tractabil pentru anumite medii. Cele mai populare soluţii aproximative includ filtrul de particule, filtrul Kalman extins și Rao-Blackwellized. Popularitatea problemei SLAM este corelată cu apariția robotizării mobile interioare. Utilizarea GPS-ului nu are o soluţie de a determina eroarea de localizare pentru utilizarea în interior. În plus, SLAM oferă o alternativă atrăgătoare pentru hărțile construite de utilizatori, arătând că operarea robotului este accesibilă chiar și în absența unei infrastructuri de localizare a obiectivelor.

Obiectivele propuse si realizate ale acestui proiect sunt urmatoarele:  Dezvoltarea unui vehicul controlat de la distanţă  Implementarea si testarea modulelor hardware pentru deplasare şi interfaţarea cu mediul inconjurator
18

 Implementarea şi testarea unui algoritm care să determine harta incaperii in care se află şi poziţia relativ cu această hartă
Proiectul este structurat în 5 capitole, fiecare prezentând detalii cu privire la diferite aspect teoretice sau practice. Capitolul 1 prezintă o introducere in temă a proiectului si noțiunile teoretice necesare dezvoltării acestuia. In capitolul 2 se descrie implementarea hardware a mobilului, capitolul 3 prezintă succint câteva noțiuni legate de filtrele de particule şi cum funcționează ele pentru a calcula poziția și harta mediului în care mobilul nostru se află, capitolul 4 vobește despre implementarea software și corectarea unor erori din hardware, iar în ultimul capitol sunt discutate rezultatele obtinute în urma testării şi eventualele căi de dezvoltare ale proiectului.

In this project، we provide a new vision to benefit from the advances in machine learning algorithms that used in localization and tracking systems which depends on using wireless signals. The project includes a study to relate Received Signal Levels with the corresponding location and then getting the optimal estimate through Kalman Filter assuming that the user’s motion and the measurement process are stochastic processes. Also، a practical outdoor implementation in Mazzeh Area، Damascus collaborating with telecommunication provider Syriatel and indoor implementation in Computer Engineering Building in Electrical and Mechanical Engineering faculty by using a mobile robot.

This paper aims to present an Adaptive Decentralized Cooperative Visual Simultaneous Localization and Mapping (AD CVSLAM) solution based on a stereovision system. With the proposed solution, a group of Unmanned Aerial Vehicles (UAVs) will be able to construct a large reliable map and localize themselves in this map without any user intervention. The robust Nonlinear H∞ filter is adapted to the multiple vehicle case to ensure a robust Pose estimation; as well a new approach for feature initialization based on an Adaptive Covariance Intersection technique is used. Finally, validation and comparison using simulation data are presented and discussed.

In robotics, the term Simultaneous Localization And Mapping (SLAM)
refers to the ability of a robot to create a map of the surrounding environment, and to localize itself in that map, by using its on-board sensors. SLAM is fundamental for mobile robots that want to efficiently
solve common tasks such as search and rescue, space exploration and
self-driving. Indeed, SLAM allows robots to perform more complex actions and reasoning, and for all these reasons it is considered an enabling technology toward the
construction of truly autonomous mobile robots.
For more than three decades, the robotics research community focused on the problem of localization and mapping. This led to the development of many different SLAM
approaches, and solutions. However, while for 2D indoor applications SLAM reached
stable results, when dealing with large-scale environments, or when extending the problem to 3D data, multiple issues arise. Under such assumptions, indeed, the complexity
and the memory consumption increase significantly, in particular if the robot has to
operate for long periods of time. Despite the recent advancements in the state-of-the-art of SLAM, current 3D localization and mapping methods can robustly handle only
scenarios of moderate size (e.g. buildings). Moreover, in order for humans beings to
be safe, novel solutions must be capable to detect and recover from failures, as well as
be able to deal with noise such as dynamics aspects of the environment and perceptual
aliasing.
In the last years the demand for robots capable of helping human beings in their
daily environment, including agriculture and autonomous driving, is constantly grow-
ing. As a result of this, researchers concentrate on the development of long term SLAM
systems that are able to cope with indoor/outdoor challenging environments such as
urban scenarios, or underground tunnels. This also means that these systems must
be designed to face sensor limitations, such as data sparsity and perceptual aliasing,
and common noise factors like seasonal changes and moving objects. In this thesis we
present multiple novel solutions that improve the accuracy, efficiency and robustness of
the current state-of-the-art in localization and mapping, under such hard conditions.
Modern SLAM systems are built as the aggregation and connection of multiple independent modules like algorithms for feature/key-point detection (data enhancement),
measurement registration methods and optimization frameworks. A fundamental block
is that one related to loop closures. By detecting areas already visited in the past by
the robot, loop closing allows for a drastic reduction of the error affecting the global
map estimate. Advancing any of these parts has the effect of improving the entire
SLAM system. In this work we focused on developing innovative solutions for some of
the most important modules composing modern localization and mapping approaches.
We present a novel point cloud registration method based on an augmented error
function associated to the alignment problem. Through an extensive experimental
evaluation, our approach demonstrates the ability to outperform other state-of-the-art
algorithms. We detail a robust and effective loop closing search heuristic that, based
on the trajectory followed by the robot, significantly reduces the overall computation
time. Additionally, similarly to visual SLAM approaches, it allows for recovering from
inconsistent global map estimates due to miss detected closures. Such an algorithm is
able to correctly map hard and harsh environments like catacombs. Also, we introduce
a method that exploits 3D sensor data to build highly informative two dimensional
measurements. These measurements are then used for achieving robust indoor/outdoor
2D SLAM. Our algorithm also makes efficient use of memory, making the approach
particularly suitable for large-scale mapping. Finally, we present a very fast method
for computing robust 3D geometric features from sparse sensor data. We demonstrate
the effectiveness of that approach by performing 3D feature based SLAM with an
autonomous car at the MCity test facility.

This paper investigates 3-dimensional Simultaneous Localization and Mapping (SLAM) and the corresponding observability analysis by fusing data from landmark sensors and a strap-down Inertial Measurement Unit (IMU) in an adaptive Kalman filter (KF). In addition to the vehicle's states and landmark positions, the self-tuning filter estimates the IMU calibration parameters as well as the covariance of the measurement noise. Examining the observability of the 3D SLAM system leads to the the conclusion that the system remains observable provided that at least one of these conditions is satisfied i) two known landmarks of which the connecting line is not collinear with the vector of the acceleration are observed ii) three known landmarks which are not placed in a straight line are observed.

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.

This paper presents the Visual Simultaneous Localization and Mapping (vSLAMTM) algorithm, a novel algorithm for simultaneous localization and mapping (SLAM). The algorithm is vision-and odometry-based, and enables low-cost navigation in cluttered and populated environments. No initial map is required, and it satisfactorily handles dynamic changes in the environment, for example, lighting changes, moving objects and/or people. Typically, vSLAM recovers quickly from dramatic disturbances, such as “kidnapping”.

Reliable perception is required in order for robots to operate safely in unpredictable and complex human environments.  However, reliability of perceptual inference algorithms has been poorly studied so far.  These algorithms capture uncertain knowledge about the world in the form of probabilistic belief distributions.  A number of Monte Carlo and deterministic approaches have been developed, but their efficiency depends on the degree of smoothness of the beliefs.  In the real world, the smoothness assumption often fails, leading to unreliable perceptual inference results.

Motivated by concrete robotics problems, we propose two novel perceptual inference algorithms that explicitly consider local non-smoothness of beliefs and adapt to it.  Both of these algorithms fall into the category of iterative divide-and-conquer methods and hence scale logarithmically with desired accuracy.  The first algorithm is termed Scaling Series.  It is an iterative Monte Carlo technique coupled with annealing.  Local non-smoothness is accounted for by sampling strategy and by annealing schedule.  The second algorithm is termed GRAB, which stands for Guaranteed Recursive Adaptive Bounding.  GRAB is an iterative adaptive grid algorithm, which relies on bounds.  In this case, local non-smoothness is captured in terms of local bounds and grid resolution.  Scaling Series works well for beliefs with sharp transitions, but without many discontinuities.  GRAB is most appropriate for beliefs with many discontinuities.  Both of these algorithms far outperform the prior art in terms of reliability, efficiency, and accuracy.  GRAB is also able to guarantee that a quality approximation of the belief is produced.

The proposed algorithms are evaluated on a diverse set of real robotics problems: tactile perception, autonomous driving, and mobile manipulation.  In tactile perception, we localize objects in 3D starting with very high initial uncertainty and estimating all 6 degrees of freedom.  The localization is performed based on tactile sensory data.  Using Scaling Series, we obtain highly accurate and reliable results in under 1 second.  Improved tactile object localization contributes to manufacturing applications, where tactile perception is widely used for workpiece localization.  It also enables robotic applications in situations where vision can be obstructed, such as rescue robotics and underwater robotics.

In autonomous driving, we detect and track vehicles in the vicinity of the robot based on 2D and 3D laser range finders.  In addition to estimating position and velocity of vehicles, we also model and estimate their geometric shape.  The geometric model leads to highly accurate estimates of pose and velocity for each vehicle.  It also greatly simplifies association of data, which are often split up into separate clusters due to occlusion. The proposed Scaling Series algorithm greatly improves reliability and ensures that the problem is solved within tight real time constraints of autonomous driving. 

In mobile manipulation, we achieve highly accurate robot localization based on commonly used 2D laser range finders using the GRAB algorithm.  We show that the high accuracy allows robots to navigate in tight spaces and manipulate objects without having to sense them directly.  We demonstrate our approach on the example of simultaneous building navigation, door handle manipulation, and door opening.  We also propose hybrid environment models, which combine high resolution polygons for objects of interest with low resolution occupancy grid representations for the rest of the environment.  High accuracy indoor localization contributes directly to home/office mobile robotics as well as to future robotics applications in construction, inspection, and maintenance of buildings.

Based on the success of the proposed perceptual inference algorithms in the concrete robotics problems, it is our hope that this thesis will serve as a starting point for further development of highly reliable perceptual inference methods.

The lack of publicly accessible datasets with a reliable ground truth has prevented in the past a fair and coherent comparison of different methods proposed in the mobile robot Simultaneous Localization and Mapping (SLAM) literature. Providing such a ground truth becomes specially challenging in the case of visual SLAM, where the world model is 3-dimensional and the robot path is 6-dimensional. This work addresses both the practical and theoretical issues found while building a collection of six outdoor datasets. It is discussed how to estimate the 6-d vehicle path from readings of a set of three Real Time Kinematics (RTK) GPS receivers, as well as the associated uncertainty bounds that can be employed to evaluate the performance of SLAM methods. The vehicle was also equipped with several laser scanners, from which reference point clouds are built as a testbed for other algorithms such as segmentation or surface fitting. All the datasets, calibration information and associated software tools are available for download http://babel.isa.uma.es/mrpt/papers/dataset2009/.

This paper presents an autonomous robotic wheelchair navigation system (ARWNS). The proposed ARWNS provides the functions of indoor map construction, wheelchair localization, path planning, obstacle avoidance navigation. A front LiDar was utilized in this paper to create the indoor map, to localize the wheelchair itself, as well as to identify the front obstacles. Hector simultaneous localization and mapping (SLAM) open source codes of the Robot Operating System (ROS) were applied in this paper to realize a LiDar-based map construction and localization system. The rapidly-exploring random tree (RRT) approach was further applied in terms of the SLAM-based map information to dynamically plan a feasible path to reach the destination. Furthermore, the autonomous navigation system was proposed by utilizing artificial potential field (APF) approach so that the robotic wheelchair is capable of following the planned path to reach the destination without any collision with obstacle. Finally, a lab-made robotic wheelchair controller was implemented based on the aforementioned approaches for experimental validation in terms of different obstacle arrangements.

WLAN has lately been applied to the problem of mobility tracking and behavior analysis. To further the development of the studies in this direction the positioning system that can perform on the network side with minimal human participation is needed. One of the current limitations is the requirement on the number of reference signal available. Thus, methods that require fewer reference signals for positioning are needed. This paper provides the comparison of WLAN based positioning methods that can operate with a single AP.

This paper proposes a simultaneous localization and mapping (SLAM) scheme that is applicable to the autonomous navigation of unmanned underwater vehicles (UUVs). A SLAM scheme is an alternative navigation method for measuring the environment through which the vehicle is passing and providing the relative position of the UUV. An unscented Kalman filter (UKF) is utilized to develop a SLAM that is suitable to estimate the locations of the UUV and the surrounding objects when the motion of the UUV is highly nonlinear. A range sonar is used as a sensor to collect the data including the spatial information of the environment in which the UUV is navigating. The proposed UKF-SLAM scheme was validated through simulations and experiments that used various two and three degrees of freedom motion conditions with a real UUV in a towing tank environment. The results of these simulations and experiments showed that the proposed SLAM algorithm is capable of estimating the position of the UUV and the surrounding objects under highly nonlinear motion and demonstrated that the algorithm can perform well in various conditions. Key words-Simultaneous localization and mapping (SLAM), range sonar, unmanned underwater vehicle (UUV), unscented Kalman filter (UKF), towing tank simulation.