In complex systems, it has been observed that the parts within the system are "loosely coupled". Loose coupling means that the parts of the system interact in some way, and as long as this interaction is maintained the parts can evolve independently. Detrimental evolutionary changes within one part of the system do not negatively affect other parts. Overall system functionality is maintained, leading to faster evolution. In swarm robotic systems there are multiple robots working together to achieve a shared goal. However it is not always obvious how to program the actions of the robots such that the desired aggregate behaviour emerges. One solution is to use a genetic algorithm to evolve robot controllers, this approach is called "Evolutionary Swarm Robotics". This thesis makes the case that swarm robotic systems are complex systems, and hypothesises that loose coupling between the robots in a swarm would lead to faster evolution. Robot swarms are investigated where robots describe environmental features to each other as part of a foraging task. Multiple descriptions can be used to describe a feature. The mappings between feature descriptions, and the signals used to express those descriptions, are manipulated. By doing this, the interactions between robots can change over time or stay the same. Results show that loose coupling leads to higher swarmfitnesses because it makes the communicated information easier to interpret. However there are some subtleties in its working. We also observe that if some of the information is not useful for completing the task, this negatively affects swarm fitness regardless of coupling. This problem can be mitigated by using loose coupling. This research has implications for the design of communication within robot swarms. Before evolution, it is difficult to know what information is relevant. This research shows that sharing unnecessary information between robots is detrimental to swarm fitness because the cost of interpreting information can be greater than the benefit gained from the information. Loose coupling can reduce, but not eliminate, the evolutionary cost of interpreting multiple pieces of information in exchange for slower message transmission.

In the thesis we solve the shared-control problem for three classes of systems: a class of linear mechanical systems, mobile robots and rear wheel drive cars, via full state feedback or output feedback while ensuring that all the state constraints on the closed-loop systems are satisfied. To design the feedback controller for a system with state constraints we firstly remove all the constraints by changing the coordinates through a logarithmic function. Then the back-stepping method is used to design the controller and a Lyapunov-like analysis is used to prove stability properties of the closed-loop system. The shared-control algorithm is based on a hysteresis switch which reduces oscillations when changing the control authority from the human operator to the feedback controller or vice-versa. Unlike other shared-control methods, formal properties of the closed-loop systems with the shared-control have been rigorously established. We start the design of the full state-feedback shared-controller with the assumption that the admissible Cartesian configuration set Pa of the system is a time-invariant convex set defined by a group of linear inequalities. Then the results are extended to the design of shared-controllers via output feedback. In the cases in which only output feedback is available, we can solve the problem by either developing an observer or 'remodeling' the system. Through system remodeling we are able to deal with any shape of the admissible configuration set Pa, even time-varying ones. Simulation results help to illustrate how the shared-controller works and show its effectiveness. The state of the closed-loop system with the shared-control never violates the constraints. Experiments done on a mobile robot also demonstrate that the shared-control algorithm works well in practice and meets all safety requirements. In addition, the experimental results match the simulation ones, indicating that the modeling approximations are reasonable and suitable.

This thesis is concerned with the task of efficiently recognising the particular instance of a scene depicted in a query image, with applications in robot navigation including loop closure, global localisation and topological navigation. Three novel frameworks are proposed, each based on learning scene models by tracking local features to form sets of landmarks. Recognition then proceeds by considering 2D constraints between pairs of local feature correspondences to efficiently approximate global scene geometry. First, the inter-image and intra-image pairwise geometries are considered to reduce feature correspondences to a more succinct set for a RANSAC-based 3D geometry constraint. A Hough-transform voting scheme based on inter-image correspondences weakly prunes the set of correspondences, after which intra-image geometries constrain the relative image positions of correspondences to eliminate unrealistic configurations. This idea is first proposed in an image retrieval application, and then extended to scene recognition where relative landmark positions are learned explicitly per scene. Second, a method is introduced to embed 2D pairwise geometry directly in an inverted index, to allow for fast scene recognition without 3D estimations. A set of discrete geometric words are extracted for a query image, and passed through the index to find examples of such pairwise configurations in the database. A global geometry constraint is then proposed by considering a maximum-clique approach to an adjacency matrix of correspondences. Third, a global topological localisation system is investigated which learns a naive Bayesian network for each landmark, to efficiently approximate global geometry without a fully-connected model. Long-term robot navigation is then addressed by learning scene models in an incremental manner, and updating the dynamic properties of landmarks accordingly. Experiments were performed on a new challenging dataset obtained by manually walking along a 7km path in a park and urban district, to capture long-term effects over an 8 month period.

This thesis concerns the automated reverse engineering of agent behaviors. It proposes a metric-free coevolutionary approach - Turing Learning, which allows a machine to infer the behaviors of agents (simulated or physical ones), in a fully automated way. Turing Learning consists of two populations. A population of models competitively coevolves with a population of classifiers. The classifiers observe the models and agents. The fitness of the classifiers depends solely on their ability to distinguish between them. The models, on the other hand, are evolved to mimic the behavior of the agents and mislead the judgment of the classifiers. The fitness of the models depends solely on their ability to 'trick' the classifiers into categorizing them as agents. Unlike other methods for system identification, Turing Learning does not require any predefined metrics to quantitatively measure the difference between the models and agents. The merits of Turing Learning are demonstrated using three case studies. In the first case study, a machine automatically infers the behavioral rules of a group of homogeneous agents through observation. A replica, which resembles the agents under investigation in terms of behavioral capabilities, is mixed into the group. The models are executed on the replica. This case study is conducted with swarms of both simulated and physical robots. In the second and third case studies, Turing Learning is applied to infer deterministic and stochastic behaviors of a single agent through controlled interaction, respectively. In particular, the machine is able to modify the environmental stimuli that the agent responds to. In the case study of inferring deterministic behavior, the machine can construct static patterns of stimuli that facilitate the learning process. In the case study of inferring stochastic behavior, the machine needs to interact with the agent on the fly through dynamically changing patterns of stimuli. This allows the machine to explore the agent's hidden information and thus reveal its entire behavioral repertoire. This interactive approach proves superior to learning only through observation.

This thesis presents the development of a novel two-wheeled robotic vehicle with a movable payload and able to manoeuvre in different environments and terrains. The vehicle structure is based on the double inverted pendulum on cart mechanism. The system has five degrees of freedom that allow the vehicle to serve as a basis for new mobility solution applications. In this study, the vehicle model is derived mathematically using the Euler-Lagrange approach to describe the system dynamics. A hybrid fuzzy logic control approach is designed to stabilise and drive the vehicle on different terrains with different inclination angles. The Matlab Simulink environment is used to simulate the vehicle system. A hybrid spiral dynamic bacteria chemotaxis optimisation algorithm is used to optimise the control parameters to achieve the least mean square error of system response and to reduce the amount of exerted control effort. Various simulation scenarios are considered to demonstrate the vehicle’s ability to work on smooth and frictional surfaces. Disturbances are applied to the vehicle to evaluate the performance of the developed control system in coping with disturbances of variable amplitudes and durations. It is shown that the vehicle exhibits a stable response and a high degree of control robustness. A steering mechanism is implemented to drive the vehicle in different environments and terrains encountered in real life. Environment modelling has been incorporated into the vehicle system to simulate various ground types and levels of frictional forces. It is demonstrated that the vehicle is able to successfully manoeuvre in indoor and outdoor environments and on flat and sloped surfaces fulfilling the aims and objectives of the research.

No description available.

Search and rescue (SAR) mission always takes place when disasters happen. Disaster could be defined into two categories, namely natural disaster and man-made disaster. Natural disasters normally cover a large area making the SAR mission’s team require an aerial view from airplane. This is because it changes the geographical landscape of the affected areas in huge perimeters. The impact is not only changing the whole landscapes, but it also impacts on residences, commercial buildings, transportations and communication infrastructures. This is always the primary reason of choosing an air vehicle as a first respond for any natural disaster. Meanwhile, made-made disasters occur in small areas relative to natural disaster. Terrorist bombing, structural collapse because of human failures or serious accident are some examples of man-made disaster. In addition, the effect from natural disaster such as earthquake also resulting horrendous structural collapse. The challenges for this rescue operation are focused on the interior of the rubble and entire external extent of the damage often not as primary interest because most victims are trapped inside, under the rubbles. Locating, extracting and rescuing any survivors will become the main goal for any rescue mission. Besides, the mission also deals with a lot of potentially dangerous situation such as further collapse, explosions, hazardous gas leaks and fire. Extreme high temperature from fire or explosions prevent rescuer to go down further into rubbles. Urban search and rescue (USAR) is the term that is being used recently for the rescue operation after man-made disaster. Conventionally, dog has been used to identify location of any potential survivors in the rubble. Again, capabilities of dog rescuer are restricted by certain working temperature, uncertainty of void size and fatigue factor. USAR operation is like race against time where trapped survivors cannot wait any longer. Further collapse or explosion may happen anytime. Therefore, the rescue team should have ideal strategy and tactic in order to maximize numbers of survivors being extract from rubble but also minimise the risk face by rescuer. Hazards is everywhere at disaster site. Human rescuers as well as dog are exposed to danger such as further collapse which would trap them in rubble and resulting an increased number of victims. This kind of situation make USAR uncertain. As a result, hazard identification and situation awareness have to be conducted concurrently with finding survivors. Robotic system, in many ways, have shown its versatility in wide range of applications. For instance, a modern and sophisticated automative assembly plant employ robotic systems in the production line in order to fullfil a specific part assembly task. On the other hand, robotic systems also started to be used as exploration vehicle in unknown world such as deep sea and outer space exploration purposes. In many aspects, the implementation of robotic systems in these applications have a significant impact to the overall process flow of the specific application. Having said that, mobile robots use in many deep sea explorations help scientists to discover the 'underworld' where human cannot explore. Therefore, implementation of robotic system in USAR operation is inevitable. In fact, it has been used in several USAR operation including the 9/11 World Trade Centre tragedy and Fukushima Daiichi Nuclear Power Station. The physical design of mobile robot is one of the main challenges to implement robotic system in USAR operation. The ability to manouver and negotiate with rough terrain is highly essential. In addition, the physical design is also need appropriate sensors in order to sense the environment. Therefore, the overall mechatronic structure must consist a robust platform equipped with sensors and actuators and able to navigate seemlessly on extreme rough terrain as well as perform designated task (e.g., find survivors, clean debris or conduct onsite disaster assessment). In order for a mobile robot to operate in unknown world such as USAR environment, it is crucially important for the mobile robot to have certain level of autonomy to plan a desired behaviour and act according to the surrounding. Eventhough it is very challenging to program a mobile robot for this type of environment, a comprehensive control architecture provide a systematic overview of the overall programming structure whilst simplifiying the programming procedure. On top of that, the ability of the overall robot system to plan and track its mission is clearly present in the control architecture. Navigation problem, which is one of the common problem in any exploration robot, also can be solved systematically. In general navigation task, a mobile robot is required to move according to the prior designated trajectory, normally in 2-dimension (flat surface). However, a mobile robot that is design for the USAR operation should be able work and navigate in unknown, uncertain and complex environment. This thesis describes the development of a mobile robot system motivated by the shape-shifting or variable geometry tracked vehicle (VGTV) configuration. The mobile robot is designed with expectation to be able to traverse on various types of terrain and enhance stability to prevent tip-over mishap. The practical work is evaluated by experimental trials on prepared terrains such as staircase, ramp and curb. On top of that, the control framework is outlined to set the objectives of the mobile robot system based on the control hierarchy. This set of works is further simulated with the aim to solve navigation problems as well as to determine the mobile robot behaviour when it is required to travel on uneven surface.

The work presented in this thesis deals with the design and real-time implementation of a control system for a new configuration of two-wheeled robot. In real life applications, two wheeled robots are considered to carry payloads of different sizes at different positions and different motion speeds along the vertical axis. Such parameters will have an impact on the robot stability and the control mechanism, hence their detailed study is essential for robust performance of the system. This work investigates the impact of the dynamic change of the payload position on the system damping characteristics whilst the robot is in its balancing state. A mathematical model is developed to describe and study the dynamics of the two-wheeled robot system. Accurate tilt angle measurement is achieved through applying a sensor fusion realized by complementary filter. A PID control strategy is considered to balance and position the robot assuming a fixed location of the payload. Then the controller is extended to provide the robot with the ability of self-standing from its rest position without human interaction. The developed controller and the sensor fusion filter are implemented on a microcontroller development board. The results show that the implementation of the controller fulfils the requirement to balance the robot. The two-wheeled robot test rig was modified to accommodate the payload actuation unit and improve the overall performance. Consequentially, the controller of the robot is also modified with the sensor fusion algorithm enhanced by implementing Kalman filter. The controller comprise a PID feedback with a feedforward approach. Furthermore, gain scheduling approach is utilized to ensure smooth and fast braking of the two-wheeled robot. The control approach is extended to an intelligent controller, where a PD-like fuzzy controller is design. A binary coding technique is developed for real-time implementation of the fuzzy controller. Such coding and implementation eliminates the need to store a big lookup table for the controller rules. The controller is tested in laboratory experiments and its robustness is demonstrated with application of various disturbance forces on the system. In order to evaluate the performance of the new configuration two-wheeled robot, various experiments are conducted under different conditions. The results demonstrate that the proposed two-wheeled robot configuration and the proposed control approaches form a solid foundation and a framework for assisted mobility of disabled and elderly people.

The aim of this PhD study was to develop a locomotion system for use on a robotic device that can traverse a liquid filled colon for atraumatic inspection and biopsy tasks. The PhD was undertaken as part of a larger two-centre EU project, which aimed to bring about a change in the way colonoscopy is done by moving to “robotic hydro-colonoscopy”. Current colonoscopes have seen little change or innovation throughout their 40 years of use with patient discomfort still limiting success and adherence of the procedure. In this thesis the initial development and testing of an amphibious locomotion concept for use in a liquid filled colon during a procedure known as hydrocolonoscopy is described. The locomotion system is comprised of four Archimedes’ screws arranged in two counter rotating pairs. These aim to provide propulsion through a liquid filled colon as well as provide locomotive traction against colonic tissue in partially fluid filled or collapsed sections of the colon, such as the splenic flexure. Experimental studies were carried out on a single screw system in fluid and dual counter-rotating screws in contact conditions. These show the system’s ability to generate thrust in the two discrete modes of locomotion of the amphibious system. A 2:1 scale prototype of the proposed device was produced from multiple materials using additive layer manufacturing processes and commercially available components. The prototype features compliant screw threads to provide atraumatic locomotion; through material selection the components will yield before damage to tissue occurs. The scale prototype device was tested in an ex-vivo porcine colon which demonstrated that this concept has the potential to be used for an intra-luminal robotic device. The key contributions of this research are: variable geometry locomotion system; amphibious locomotion using Archimedes’ screws; experimental assessment of the locomotion in fluid, contact and amphibious states; and analysis of the contact dynamics against tissue.

This thesis presents, explains and analyses a novel design of a locomotion mechanism for a miniature robot envisaged for assisting surgeons during minimally invasive procedures in abdominal surgery. Minimally invasive procedures have proved to be beneficial for hospitals and patients and are currently applied successfully in many surgical operations. Robotic arms mounted outside the body are currently used in order to move the surgical tools inside the body and some research prototypes move fully inside the abdomen. In order to fully realise the potential of minimally invasive robotic surgery, the robotic assistant should operate at a distance from intense surgical activity and attach to tissue, moving stably within the abdomen. This thesis presents the conceptual design of a miniature robot which uses four adhesive pads to attach to the surface of the abdominal wall, a vantage point within the abdominal cavity. The adhesive pads use a micro-structured surface inspired by tree frogs in order to obtain smooth and repeatable attachment to biological tissue and enable the robot to move in inverted locomotion. The design of the locomotion mechanism of the robot also takes inspiration from tree frogs and geckoes in the way the pads are peeled off the tissue in order to detach them. Inspired by amoeboid locomotion, the robot detaches one pad at a time and changes the shape of the locomotion mechanism in the horizontal plane in order to move the pads across the tissue. The implementation and testing of the robot resulted in a proof-of-concept prototype, able to move consistently using magnetic pads with an adhesion force similar to the bio-mimetic pads. The robot also managed to attach and move the pads while attached to tissue with the bio-mimetic pads. The analysis of the locomotion mechanism resulted in the definition of a peeling model for the adhesive pad, and a stability criterion and control strategies for adhesion-reliant robots.