Current robot manipulation and control paradigms have largely been developed for static or highly structured environments such as those common in factories. For most techniques in robot trajectory generation, such as heuristic-based geometric planning, this has led to putting a high cost on contact with the world. This approach and methodology can be prohibitive to robots operating in many unmodeled and dynamic environments. This dissertation presents work on using haptic based feedback (torque and tactile sensing) to formulate a controller for robot manipulation in clutter. We define “clutter” as any environment in which we expect the robot to make both incidental and purposeful contact while maneuvering and manipulating. The controllers developed in this dissertation take the form of single or multi-time step Model Predictive Control (a form of optimal control which incorporates feedback) which attempts to regulate contact forces at multiple locations on a robot arm while reaching to a goal. The results and conclusions in this dissertation are based on extensive testing in simulation (tens of thousands of trials) and testing in realistic scenarios with real robots incorporating tactile sensing. The approach is novel in the sense that it allows contact and explicitly incorporate the contact and predictive model of the robot arm in calculating control effort at every time step. The expected broader impact of this research is progress towards a new foundation of reactive feedback controllers that will include a higher likelihood of success in many constrained and dynamic scenarios such as reaching into containers without line of sight, maneuvering in cluttered search and rescue situations or working with unpredictable human co-workers.

Starting a few decades ago, the unmanned wheeled vehicle research has drawn lately more attention, especially for off-road environment. As the demand to use electric vehicles increased, the need to conceptualize the use of electrically driven vehicles in autonomous operations became a target. That is because in addition to the fact that they are more environmentally friendly, they are also easier to control. This also gives another reason to enhance further the energy economy of those unmanned electric vehicles. Off-road vehicles research was always challenging, but in the present work the nature of the off-road land is utilized to benefit from in order to enhance the energy consumption of those vehicles. An algorithm for energy consumption optimization for electrically driven unmanned wheeled vehicles is presented. The algorithm idea is based on the fact that in off-road conditions, when the vehicle passes a ditch or a hole, the kinetic energy gained while moving downhill could be utilized to reduce the energy consumption for moving uphill if the dimensions of the ditch/hole were known a distance ahead. Two manipulated variables are evaluated: the wheels DC motors supply voltage and the DC armature current. The developed algorithm is analysed and compared to the PID speed iii controller and to the open-loop control of DC motors. The developed predictive controller achieved encouraging results compared to the PID speed control and also compared to the open-loop control. Also, the use of the DC armature current as a manipulated variable showed more noticeable improvement over using the DC input voltage. Experimental work was carried out to validate the predictive control algorithm. A mobile robot with two DC motor driven wheels was deployed to overcome a ditch-like hindrance. The experimental results verified the simulation results. A parametric study for the predictive control is conducted. The effect of changing the downhill angle and the uphill angle as well as the size of the prediction horizon on the consumed electric energy by the DC motors is addressed. The simulation results showed that, when using the proposed approach, the larger the prediction horizon, the lower the energy consumption is.

Wireless sensor networks (WSN) increasingly permeate modern societies nowadays. But in spite of their plethora of successful applications, WSN are often unable to surmount many operational challenges that unexpectedly arise during their lifetime. Fortunately, robotic agents can now assist a WSN in various ways. This thesis illustrates how mobile robots which are able to carry a limited number of sensors can help the network react to sensor faults, either during or after its deployment in the monitoring region. Two scenarios are envisioned. In the first one, carrier robots surround a point of interest with multiple sensor layers (focused coverage formation). We put forward the first known algorithm of its kind in literature. It is energy-efficient, fault-reactive and aware of the bounded robot cargo capacity. The second one is that of replacing damaged sensing units with spare, functional ones (coverage repair), which gives rise to the formulation of two novel combinatorial optimization problems. Three nature-inspired metaheuristic approaches that run at a centralized location are proposed. They are able to find good-quality solutions in a short time. Two frameworks for the identification of the damaged nodes are considered. The first one leans upon diagnosable systems, i.e. existing distributed detection models in which individual units perform tests upon each other. Two swarm intelligence algorithms are designed to quickly and reliably spot faulty sensors in this context. The second one is an evolving risk management framework for WSNs that is entirely formulated in this thesis.

In this thesis work, a mobile robot is designed to be used in search and rescue operations to help the human rescue workers. The difficult physical conditions in the ruins obstruct the movement. Therefore, it is aimed to design a search and rescue robot which can move easily on rough terrain and climb over the obstacles. The designed robot is made up of three modules. A connecting unit is designed that is situated between each module. This connecting unit which is composed of two universal and one revolute joint gives 5 DOF relative motions to the modules. On the other hand, the wheel&rsquo / s continuous contact with the ground is important while moving on rough terrain. In order to increase the adaptation of the robot to the rough terrain the rear axle is connected to the body with a revolute joint. Besides, skid steering system is used in the design of the robot to attain a compact and light solution which requires few parts. In the study, kinematic equations and dynamic equations of the robot are obtained to be used by the control program. The dynamic equations are obtained by using the Newton &ndash / Euler formulation. The forces, which are transmitted by the connecting unit to the modules, and the reaction forces formed between the wheels and the ground are derived by using these equations. &ldquo / Follow-the-Leader approach&rdquo / is used as a control strategy to make the modules move in formation and to reduce the tracking problem. In this approach, the first module is the leader and the second and third modules follow it. A Matlab program is written to control the robot by using the constructed mathematical model of the robot. The reaction forces between the wheels and the ground are calculated through using the Matlab program written. Moreover to make the simulations of the robot for some cases, a model is constructed in ADAMS program.

Topological maps are light-weight, graphical representations of environments that are scalable and amenable to symbolic manipulation. Thus, they are well- suited for basic robot navigation applications, and also provide a representational basis for the procedural and semantic information needed for higher-level robotic tasks. However, their widespread use has been impeded in part by the lack of reliable, general purpose algorithms for their construction. In this dissertation, I present a probabilistic framework for the construction of topological maps that addresses topological ambiguity, is failure-aware, computa- tionally efficient, and can incorporate information from various sensing modalities. The framework addresses the two major problems of topological mapping, namely topological ambiguity and landmark detection. The underlying idea behind overcoming topological ambiguity is that the com- putation of the Bayesian posterior distribution over the space of topologies is an effective means of quantifying this ambiguity, caused due to perceptual aliasing and environment variability. Since the space of topologies is combinatorial, the posterior on it cannot be computed exactly. Instead, I introduce the concept of Probabilistic Topological Maps (PTMs), a sample-based representation that ap- proximates the posterior distribution over topologies given the available sensor measurements. Sampling algorithms for the efficient computation of PTMs are described. The PTM framework can be used with a wide variety of landmark detection schemes under mild assumptions. As part of the evaluation, I describe a novel landmark detection technique that makes use of the notion of "surprise" in mea- surements that the robot obtains, the underlying assumption being that landmarks are places in the environment that generate surprising measurements. The com- putation of surprise in a Bayesian framework is described and applied to various sensing modalities for the computation of PTMs. The PTM framework is the first instance of a probabilistic technique for topo- logical mapping that is systematic and comprehensive. It is especially relevant for future robotic applications which will need a sparse representation capable of accomodating higher level semantic knowledge. Results from experiments in real environments demonstrate that the framework can accomodate diverse sensors such as camera rigs and laser scanners in addition to odometry. Finally, results are pre- sented using various landmark detection schemes besides the surprise-based one.

The field of autonomous walking robots has been dominated by the trajectory-control approach, which rigidly dictates joint angle trajectories at the expense of both energy efficiency and stability, and the passive dynamics approach, which uses no actuators, relying instead on natural mechanical dynamics as the sole source of control. Although the passive dynamics approach is energy efficient, it lacks the ability to modify gait or adapt to disturbances. Recently, minimally actuated walkers, or dynamic walkers, have been developed that use hip or ankle actuators---knees are always passive---to regulate mechanical energy variations through the timely application of joint torque pulses. Despite the improvement minimal actuation has provided, energy efficiency remains below target values and perturbation rejection capability (i.e., stability) remains poor. In this dissertation, we develop and analyze a simplified robotic system to assess biologically inspired methods of improving energy efficiency and stability in dynamic walkers. Our system consists of a planar, dynamically swinging leg with hip and knee actuation. Neurally inspired, nonlinear oscillators provide closed-loop control without overriding the leg's natural dynamics. We first model the passive stiffness of muscles by applying stiffness components to the joints of a hip-actuated swinging leg. We then assess the effect active knee control has on unperturbed and perturbed leg swinging. Our results indicate that passive joint stiffness improves energy efficiency by reducing the actuator work required to counter gravitational torque and by promoting kinetic energy transfer between the shank and thigh. We also found that active knee control 1) is detrimental to unperturbed leg swinging because it negatively affects energy efficiency while producing minimal performance improvement and 2) is beneficial during perturbed swinging because the perturbation rejection improvement outweighs the reduction in energy efficiency. By analyzing the effects of applying passive joint stiffness and active knee control to dynamic walkers, this work helps to bridge the gap between the performance capability of trajectory-control robots and the energy-efficiency of passive dynamic robots.

The ability of robotic units to navigate various terrains is critical to the advancement of robotic operation in real world environments. Next generation robots will need to adapt to their environment in order to accomplish tasks that are either too hazardous, too time consuming, or physically impossible for human-beings. Such tasks may include accurate and rapid explorations of various planets or potentially dangerous areas on planet Earth. This research investigates a navigation control methodology for a wheel-legged robot based on active vision. The method presented is designed to control the reconfigurability of the robot (i.e. control the usage of the wheels and legs), depending upon the obstacle/terrain, based on perception. Surface estimation for robot reconfigurability is implemented using a region growing method and a characterization and traversability assessment generated from camera data. As a result, a mathematical approach that directs necessary navigation behavior is implemented to control robot mobility. The hybrid wheeled-legged rover possesses a four-legged or six-legged walking system as well as a four-wheeled mobility system.

Knowing the semantic category of a robot's current position not only facilitates the robot's navigation, but also greatly improves its ability to serve human needs and to interpret the scene. Visual Place Categorization (VPC) is addressed in this dissertation, which refers to the problem of predicting the semantic category of a place using visual information collected from an autonomous robot platform. Census Transform (CT) histogram and Histogram Intersection Kernel (HIK) based visual codebooks are proposed to represent an image. CT histogram encodes the stable spatial structure of an image that reflects the functionality of a location. It is suitable for categorizing places and has shown better performance than commonly used descriptors such as SIFT or Gist in the VPC task. HIK has been shown to work better than the Euclidean distance in classifying histograms. We extend it in an unsupervised manner to generate visual codebooks for the CT histogram descriptor. HIK codebooks help CT histogram to deal with the huge variations in VPC and improve system accuracy. A computational method is also proposed to generate HIK codebooks in an efficient way. The first significant VPC dataset in home environments is collected and is made publicly available, which is also used to evaluate the VPC system based on the proposed techniques. The VPC system achieves promising results for this challenging problem, especially for important categories such as bedroom, bathroom, and kitchen. The proposed techniques achieved higher accuracies than competing descriptors and visual codebook generation methods.

Autonomous mobile robots are being developed with the aim of accomplishing complex tasks in different environments, including human habitats as well as less friendly places, such as distant planets and underwater regions. A major challenge faced by such robots is to make sure that their actions are executed correctly and reliably, despite the dynamics and the uncertainty inherent in their working space. This thesis is concerned with the ability of a mobile robot to reliably monitor the execution of its plans and detect failures. Existing approaches for monitoring the execution of plans rely mainly on checking the explicit effects of plan actions, i.e., effects encoded in the action model. This supposedly means that the effects to monitor are directly observable, but that is not always the case in a real-world environment. In this thesis, we propose to use semantic domain-knowledge to derive and monitor implicit expectations about the effects of actions. For instance, a robot entering a room asserted to be an office should expect to see at least a desk, a chair, and possibly a PC. These expectations are derived from knowledge about the type of the room the robot is entering. If the robot enters a kitchen instead, then it should expect to see an oven, a sink, etc. The major contributions of this thesis are as follows. • We define the notion of Semantic Knowledge-based Execution Monitoring SKEMon, and we propose a general algorithm for it based on the use of description logics for representing knowledge. • We develop a probabilistic approach of semantic Knowledge-based execution monitoring to take into account uncertainty in both acting and sensing. Specifically, we allow for sensing to be unreliable and for action models to have more than one possible outcome. We also take into consideration uncertainty about the state of the world. This development is essential to the applicability of our technique, since uncertainty is a pervasive feature in robotics. • We present a general schema to deal with situations where perceptual information relevant to SKEMon is missing. The schema includes steps for modeling and generating a course of action to actively collect such information. We describe approaches based on planning and greedy action selection to generate the information-gathering solutions. The thesis also shows how such a schema can be applied to respond to failures occurring before or while an action is executed. The failures we address are ambiguous situations that arise when the robot attempts to anchor symbolic descriptions (relevant to a plan action) in perceptual information. The work reported in this thesis has been tested and verified using a mobile robot navigating in an indoor environment. In addition, simulation experiments were conducted to evaluate the performance of SKEMon using known metrics. The results show that using semantic knowledge can lead to high performance in monitoring the execution of robot plans.

For robots to operate in human environments they must be able to make their own maps because it is unrealistic to expect a user to enter a map into the robot’s memory; existing floorplans are often incorrect; and human environments tend to change. Traditionally robots have used sonar, infra-red or laser range finders to perform the mapping task. Digital cameras have become very cheap in recent years and they have opened up new possibilities as a sensor for robot perception. Any robot that must interact with humans can reasonably be expected to have a camera for tasks such as face recognition, so it makes sense to also use the camera for navigation. Cameras have advantages over other sensors such as colour information (not available with any other sensor), better immunity to noise (compared to sonar), and not being restricted to operating in a plane (like laser range finders). However, there are disadvantages too, with the principal one being the effect of perspective. This research investigated ways to use a single colour camera as a range sensor to guide an autonomous robot and allow it to build a map of its environment, a process referred to as Simultaneous Localization and Mapping (SLAM). An experimental system was built using a robot controlled via a wireless network connection. Using the on-board camera as the only sensor, the robot successfully explored and mapped indoor office environments. The quality of the resulting maps is comparable to those that have been reported in the literature for sonar or infra-red sensors. Although the maps are not as accurate as ones created with a laser range finder, the solution using a camera is significantly cheaper and is more appropriate for toys and early domestic robots.