Abstract: In global navigation satellite system (GNSS) denied settings, such as indoor environments, autonomous mobile robots are often limited to dead-reckoning navigation techniques to determine their position, velocity, and attitude (PVA). Localization is typically accomplished by employing an inertial measurement unit (IMU), which, while precise in nature, accumulates errors rapidly and severely degrades the localization solution. Standard sensor fusion methods, such as Kalman filtering, aim to fuse precise IMU measurements with accurate aiding sensors to establish a precise and accurate solution. In indoor environments, where GNSS and no other a priori information is known about the environment, effective sensor fusion is difficult to achieve, as accurate aiding sensor choices are sparse. However, an opportunity arises by employing a depth camera in the indoor environment. A depth camera can capture point clouds of the surrounding floors and walls. Extracting attitude from these surfaces can serve as an accurate aiding source, which directly combats errors that arise due to gyroscope imperfections. This configuration for sensor fusion leads to a dramatic reduction of PVA error compared to traditional aiding sensor configurations. This paper provides the theoretical basis for the depth camera aiding sensor method, initial expectations of performance benefit via simulation, and hardware implementation thus verifying its veracity. Hardware implementation is performed on the Quanser Qbot 2™ mobile robot, with a Vector-Nav VN-200™ IMU and Kinect™ camera from Microsoft.
Abstract: Finding the optimal 3D path of an aerial vehicle under
flight mechanics constraints is a major challenge, especially when
the algorithm has to produce real time results in flight. Kinematics
models and Pythagorian Hodograph curves have been widely used
in mobile robotics to solve this problematic. The level of difficulty
is mainly driven by the number of constraints to be saturated at the
same time while minimizing the total length of the path. In this paper,
we suggest a pragmatic algorithm capable of saturating at the same
time most of dimensioning helicopter 3D trajectories’ constraints
like: curvature, curvature derivative, torsion, torsion derivative, climb
angle, climb angle derivative, positions. The trajectories generation
algorithm is able to generate versatile complex 3D motion primitives
feasible by a helicopter with parameterization of the curvature and the
climb angle. An upper ”motion primitives’ concatenation” algorithm
is presented based. In this article we introduce a new way of designing
three-dimensional trajectories based on what we call the ”Dubins
gliding symmetry conjecture”. This extremely performing algorithm
will be soon integrated to a real-time decisional system dealing with
inflight safety issues.
Abstract: Mobile robotics is gaining an increasingly important
role in modern society. Several potentially dangerous or laborious
tasks for human are assigned to mobile robots, which are increasingly
capable. Many of these tasks need to be performed within a specified
period, i.e, meet a deadline. Missing the deadline can result in
financial and/or material losses. Mechanisms for predicting the
missing of deadlines are fundamental because corrective actions can
be taken to avoid or minimize the losses resulting from missing the
deadline. In this work we propose a simple but reliable deadline
missing prediction mechanism for mobile robots through the use of
historical data and we use the Pioneer 3-DX robot for experiments
and simulations, one of the most popular robots in academia.
Abstract: This paper presents an Extended Kaman Filter
implementation of a single-camera Visual Simultaneous Localization
and Mapping algorithm, a novel algorithm for simultaneous
localization and mapping problem widely studied in mobile robotics
field. The algorithm is vision and odometry-based, The odometry
data is incremental, and therefore it will accumulate error over time,
since the robot may slip or may be lifted, consequently if the
odometry is used alone we can not accurately estimate the robot
position, in this paper we show that a combination of odometry and
visual landmark via the extended Kalman filter can improve the robot
position estimate. We use a Pioneer II robot and motorized pan tilt
camera models to implement the algorithm.
Abstract: In this work a visual and reactive contour following
behaviour is learned by reinforcement. With artificial vision the
environment is perceived in 3D, and it is possible to avoid obstacles
that are invisible to other sensors that are more common in mobile
robotics. Reinforcement learning reduces the need for intervention in
behaviour design, and simplifies its adjustment to the environment,
the robot and the task. In order to facilitate its generalisation to other
behaviours and to reduce the role of the designer, we propose a
regular image-based codification of states. Even though this is much
more difficult, our implementation converges and is robust. Results
are presented with a Pioneer 2 AT on a Gazebo 3D simulator.
Abstract: Fundamental sensor-motor couplings form the backbone
of most mobile robot control tasks, and often need to be implemented
fast, efficiently and nevertheless reliably. Machine learning
techniques are therefore often used to obtain the desired sensor-motor
competences.
In this paper we present an alternative to established machine
learning methods such as artificial neural networks, that is very fast,
easy to implement, and has the distinct advantage that it generates
transparent, analysable sensor-motor couplings: system identification
through nonlinear polynomial mapping.
This work, which is part of the RobotMODIC project at the
universities of Essex and Sheffield, aims to develop a theoretical understanding
of the interaction between the robot and its environment.
One of the purposes of this research is to enable the principled design
of robot control programs.
As a first step towards this aim we model the behaviour of the
robot, as this emerges from its interaction with the environment, with
the NARMAX modelling method (Nonlinear, Auto-Regressive, Moving
Average models with eXogenous inputs). This method produces
explicit polynomial functions that can be subsequently analysed using
established mathematical methods.
In this paper we demonstrate the fidelity of the obtained NARMAX
models in the challenging task of robot route learning; we present a
set of experiments in which a Magellan Pro mobile robot was taught
to follow four different routes, always using the same mechanism to
obtain the required control law.
Abstract: Developing techniques for mobile robot navigation constitutes one of the major trends in the current
research on mobile robotics. This paper develops a local
model network (LMN) for mobile robot navigation. The
LMN represents the mobile robot by a set of locally valid
submodels that are Multi-Layer Perceptrons (MLPs).
Training these submodels employs Back Propagation (BP) algorithm. The paper proposes the fuzzy C-means (FCM) in this scheme to divide the input space to sub regions, and then a submodel (MLP) is identified to represent a particular
region. The submodels then are combined in a unified
structure. In run time phase, Radial Basis Functions (RBFs) are employed as windows for the activated submodels. This
proposed structure overcomes the problem of changing operating regions of mobile robots. Read data are used in all experiments. Results for mobile robot navigation using the
proposed LMN reflect the soundness of the proposed
scheme.
Abstract: One of the long standing challenging aspect in mobile robotics is the ability to navigate autonomously, avoiding modeled and unmodeled obstacles especially in crowded and unpredictably changing environment. A successful way of structuring the navigation task in order to deal with the problem is within behavior based navigation approaches. In this study, Issues of individual behavior design and action coordination of the behaviors will be addressed using fuzzy logic. A layered approach is employed in this work in which a supervision layer based on the context makes a decision as to which behavior(s) to process (activate) rather than processing all behavior(s) and then blending the appropriate ones, as a result time and computational resources are saved.