This dissertation describes the design, analysis, planning, and control of a self-reconfigurable modular robotic system. The proposed robotic system mainly contains three major types of robotic modules: load carrier, manipulation module, and locomotion module. Each module is capable of navigation and interaction with the environment individually. In addition, the robotic system is proposed to reassemble autonomously into various configurations to perform complex tasks such as humanoid configuration to enable enhanced functionality to reconfigure into a configuration that would enable the system to cross over a ditch. A non-back drivable active docking mechanism with two Degrees of Freedom (DOFs) was designed to fit into the tracked units of the robot modules for achieveing the reconfiguration. The quantity and location of the docking mechanisms are customizable and selectable to satisfy various mission requirements and adapt to different environments. During the reconfiguration process, the target coupling mechanism of each module reconfigurable with each other autonomously. A Lyapunov function-based precision controller was developed to align the target docking mechanisms in a close range and high precision for assembling the robot modules autonomously into other configurations. Additionally, an trajectory optimization algorithm was developed to help the robot determine when to switch the locomotion modes and find the fastest path to the destination with the desired pose. / Doctor of Philosophy / Though the capabilities of individual robot platforms have advanced greatly from their original rigid construction to more modern reconfigurable platforms, it is still difficult to build a singular platform capable of adapting to all situations and environments that users may want or need it to function in. To help improve the versatility of robot systems, modular robots have become an active area of research. These modular robotic systems are made up of multiple robotic platforms capable of docking together, breaking apart, or otherwise reconfiguring to form a multitude of shapes to overcome and adapt to many diverse challenges and environments. This dissertation describes the design of a new modular robotic system with autonomous docking and reconfiguration. Existing technologies and motivations for the creation of a new modular robotic system are covered. Then the physical design, with a primary focus on the docking mechanism, is reviewed. A validation of the proposed robotic system in a virtual environment with real physical properties is demonstrated. Following this, the development of a Lyapunov function-based controller to autonomously align the docking mechanisms is presented. The overall docking process was also preliminarily verified using a prototype of a locomotion module and an active docking mechanism. In addition, the trajectory optimization and tracking methods are presented in the end.

Strategies for topology optimization of trusses and plane stress domains for minimum weight subject to stress and displacement constraints by Simultaneous Analysis and Design (SAND) are considered. The ground structure approach is used. For the truss topology optimization, a penalty function formulation of SAND is compared with an augmented Lagrangian formulation. The efficiency of SAND in handling combinations of general constraints for truss topology optimization is tested. A strategy for obtaining an optimal topology by minimizing the compliance of the truss is compared with a direct weight minimization solution to satisfy stress and displacement constraints. It is shown that for some problems, starting from the ground structure and using SAND is better than starting from a minimum compliance topology design and optimizing only the cross sections for minimum weight under stress and displacement constraints. One case where the SAND approach could not predict a singular topology obtained by compliance minimization is discussed in detail. A member elimination strategy to save CPU time is developed. For the plane stress topology optimization problem, the ground structure is obtained by using 3 noded constant stress triangular elements. A chess board pattern is observed in the optimal topologies which may be attributed to the triangular elements. Some suggestions for future research are made. / Ph. D.

Spacecraft trajectory optimization is a near-infinite problem space with a wide variety of models and optimizers. As trajectory complexity increases, so too must the capabilities of modern optimizers. Common objective cost functions for these optimizers include the propellant utilized by the spacecraft and the time the spacecraft spends in flight. One effective method of minimizing these costs is the utilization of one or multiple gravity assists. Due to the phenomenon known as the Oberth effect, fuel burned at a high velocity results in a larger change in orbital energy than fuel burned at a low velocity. Since a spacecraft is flying fastest at the periapsis of its orbit, application of impulsive thrust at this closest approach is demonstrably capable of generating a greater change in orbital energy than at any other location in a trajectory. Harnessing this extra energy in order to lower relevant cost functions requires the modeling of these “powered flybys” or “powered gravity assists” (PGAs) within an interplanetary trajectory optimizer. This paper will discuss the use and modification of the Spacecraft Trajectory Optimization Suite, an optimizer built on evolutionary algorithms and the island model paradigm from the Parallel Global Multi-Objective Optimizer (PaGMO). This variant of STOpS enhances the STOpS library of tools with the capability of modeling and optimizing single and multiple powered gravity assist trajectories. Due to its functionality as a tool to optimize powered flybys, this variant of STOpS is named the Spacecraft Trajectory Optimization Suite - Flybys with Impulsive Thrust Engines (STOpS-FLITE). In three test scenarios, the PGA algorithm was able to converge to comparable or superior solutions to the unpowered gravity assist (uPGA) modeling used in previous STOpS versions, while providing extra options of trades between time of flight and propellant burned. Further, the PGA algorithm was able to find trajectories utilizing a PGA where uPGA trajectories were impossible due to limitations on time of flight and flyby altitude. Finally, STOpS-FLITE was able to converge to a uPGA trajectory when it was the most optimal solution, suggesting the algorithm does include and properly considers the uPGA case within its search space.

by Chiu, Kit Chau. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2005. / Includes bibliographical references (leaves 90-92). / Abstracts in English and Chinese. / ABSTRACT --- p.02 / 摘要 --- p.04 / ACKNOWLEDGEMENTS --- p.06 / TABLE OF CONTENTS --- p.07 / LIST OF FIGURES --- p.10 / LIST OF TABLES --- p.12 / Chapter 1 --- INTRODUCTION --- p.13 / Chapter 1.1 --- Application with formation flying --- p.14 / Chapter 1.2 --- Previous work --- p.16 / Chapter 1.3 --- The present work --- p.18 / Chapter 1.4 --- Thesis outline --- p.19 / Chapter 2 --- OPTIMIZATION IN DESIRED TRAJECTORY --- p.21 / Chapter 2.1 --- Problem formulation --- p.21 / Chapter 2.1.1 --- System model --- p.21 / Chapter 2.1.2 --- System constraints --- p.22 / Chapter 2.1.3 --- Cost function of the system --- p.23 / Chapter 2.2 --- Reformation as optimal control problem --- p.23 / Chapter 2.2.1 --- Polynomial form for input --- p.24 / Chapter 2.2.2 --- Problem simplification --- p.26 / Chapter 2.3 --- Numerical case studies --- p.27 / Chapter 2.3.1 --- Case study 2-1: Equal weightings in all units and directions --- p.27 / Chapter 2.3.2 --- Case study 2-2: Equal weightings in all directions but different weightings in control units --- p.30 / Chapter 2.3.3 --- Case 2-3: Different weightings in x-y-z directions but equal weightings in all control units --- p.33 / Chapter 2.4 --- Chapter summary --- p.35 / Chapter 3 --- OBSTACLE AVOIDANCE --- p.36 / Chapter 3.1 --- Additions of obstacle constraints --- p.36 / Chapter 3.2 --- Simulation case studies --- p.37 / Chapter 3.2.1 --- Case study 3-1: No obstacle --- p.38 / Chapter 3.2.2 --- Case study 3-2: Single obstacles --- p.40 / Chapter 3.2.3 --- Case study 3-3: Two obstacles --- p.42 / Chapter 3.2.4 --- Case study 3-4: Two obstacles and optimal velocity --- p.48 / Chapter 3.3 --- Chapter summary --- p.51 / Chapter 4 --- FUZZY INTERACTIVE FORCE BETWEEN ELEMENTS --- p.52 / Chapter 4.1 --- Region of repulsive force --- p.52 / Chapter 4.2 --- Region of attractive force --- p.53 / Chapter 4.3 --- Beyond the attractive region --- p.53 / Chapter 4.4 --- Interactive force as function of separation --- p.54 / Chapter 4.5 --- Fuzzy mapping --- p.55 / Chapter 4.6 --- Chapter summary --- p.58 / Chapter 5 --- VIRTUAL LEADER --- p.59 / Chapter 5.1 --- Virtual leader --- p.59 / Chapter 5.2 --- Two maneuverable elements and two virtual leaders --- p.60 / Chapter 5.3 --- Rotational Trajectories for the two virtual leaders --- p.61 / Chapter 5.4 --- Chapter summary --- p.65 / Chapter 6 --- OPIMIZATION BY INTERACTIVE FORCE --- p.66 / Chapter 6.1 --- Narrow channel passage --- p.66 / Chapter 6.2 --- Interactive forces --- p.68 / Chapter 6.3 --- Definition of interactive force --- p.69 / Chapter 6.4 --- Formulation as optimization problem --- p.71 / Chapter 6.4.1 --- Parameterization of f1 and f2 --- p.71 / Chapter 6.4.2 --- Reformulated optimization problem --- p.73 / Chapter 6.5 --- Simulation results --- p.74 / Chapter 6.6 --- Chapter summary --- p.77 / Chapter 7 --- MODIFICATION IN OBSTACLE --- p.78 / Chapter 7.1 --- Modification for interactive force --- p.78 / Chapter 7.2 --- Modification in obstacle description --- p.79 / Chapter 7.3 --- """Shortest distance"" between control unit and obstacle" --- p.80 / Chapter 7.4 --- Simulation case studies --- p.81 / Chapter 7.4.1 --- Case study 7-1: Single triangular obstacle --- p.81 / Chapter 7.4.2 --- Case study 7-2: Two triangular obstacles --- p.83 / Chapter 7.5 --- Chapter summary --- p.85 / Chapter 8 --- Conclusions and future works --- p.86 / Chapter 8.1 --- Conclusions --- p.86 / Chapter 8.2 --- Future works --- p.88 / Chapter 8.2.1 --- Fuzzy mapping --- p.88 / Chapter 8.2.2 --- Intrinsic parameters and properties --- p.89 / BIBLIOGRAPHY --- p.90

Supercavitating vehicles are characterized by substantially reduced hydrodynamic drag with respect to fully wetted underwater vehicles. Drag is localized at the nose of the vehicle, where a cavitator generates a cavity that completely envelops the body. This causes the center of pressure to be always ahead of the center of mass, thus violating a fundamental principle of hydrodynamic stability. This unique loading configuration, the complex and non-linear nature of the interaction forces between vehicle and cavity, and the unsteady behavior of the cavity itself make the control and maneuvering of supercavitating vehicles particularly challenging. This study represents an effort towards the evaluation of optimal trajectories for this class of underwater vehicles, which often need to operate in unsteady regimes and near the boundaries of the flight envelope. Flight trajectories and maneuvering strategies for supercavitating vehicles are here obtained through the solution of an optimal control problem. Given a cost function and general constraints and bounds on states and controls, the solution of the optimal control problem yields the control time histories that maneuver the vehicle according to a desired strategy, together with the associated flight path. The optimal control problem is solved using the direct transcription method, which does not require the derivation of the equations of optimal control and leads to the solution of a discrete parameter optimization problem. Examples of maneuvers and resulting trajectories are given to demonstrate the effectiveness of the proposed methodology and the generality of the formulation.

Thesis (Ph.D)--Aerospace Engineering, Georgia Institute of Technology, 2008. / Committee Chair: Tsiotras, Panagiotis; Committee Member: Corban, Eric; Committee Member: Feron, Eric; Committee Member: Johnson, Eric; Committee Member: Vachtsevanos, George.

Thesis (M.S. in Astronautical Engineering)--Naval Postgraduate School, March 2005. / Thesis Advisor(s): I. Michael Ross. Includes bibliographical references (p. 57-58). Also available online.

No description available.

As progress pushes the boundaries of both the performance of new hardware components and the computational capacity of modern computers, the requirements on the performance of robotic systems are becoming more and more demanding. The objective of this thesis is to demonstrate that concurrent design (Co-Design) is the approach to follow to design hardware and control for such high-performance robots. In particular, this work proposes a co-design framework and an algorithm to tackle two main issues: i) how to use Co-Design to benchmark different robotic systems, and ii) how to effectively warm-start the trajectory optimization (TO) problem underlying the co-design problem aiming at global optimality. The first contribution of this thesis is a co-design framework for the energy efficiency analysis of a redundant actuation architecture combining Quasi-Direct Drive (QDD) motors and Series Elastic Actuators (SEAs). The energy consumption of the redundant actuation system is compared to that of Geared Motors (GMs) and SEAs alone. This comparison is made considering two robotic systems performing different tasks. The results show that, using the redundant actuation, one can save up to 99% of energy with respect to SEA for sinusoidal movements. This efficiency is achieved by exploiting the coupled dynamics of the two actuators, resulting in a latching-like control strategy. The analysis also shows that these large energy savings are not straightforwardly extendable to non-sinusoidal movements, but smaller savings (e.g., 7%) are nonetheless possible. The results highlight that the combination of complex hardware morphologies and advanced numerical Co-Design can lead to peak hardware performance that would be unattainable by human intuition alone. Moreover, it is also shown how to leverage Stochastic Programming (SP) to extend a similar co-design framework to design robots that are robust to disturbances by combining TO, morphology and feedback control optimization. The second contribution is a first step towards addressing the non-convexity of complex co-design optimization problems. To this aim, an algorithm for the optimal control of dynamical systems is designed that combines TO and Reinforcement Learning (RL) in a single framework. This algorithm tackles the two main limitations of TO and RL when applied to continuous-space non-linear systems to minimize a non-convex cost function: TO can get stuck in poor local minima when the search is not initialized close to a “good” minimum, whereas the RL training process may be excessively long and strongly dependent on the exploration strategy. Thus, the proposed algorithm learns a “good” control policy via TO-guided RL policy search. Using this policy to compute an initial guess for TO, makes the trajectory optimization process less prone to converge to poor local optima. The method is validated on several reaching problems featuring non-convex obstacle avoidance with different dynamical systems. The results show the great capabilities of the algorithm in escaping local minima, while being more computationally efficient than the state-of-the-art RL algorithms Deep Deterministic Policy Gradient and Proximal Policy Optimization. The current algorithm deals only with the control side of a co-design problem, but future work will extend it to include also hardware optimization. All things considered, this work advanced the state of the art on Co-Design, providing a framework and an algorithm to design both hardware and control for high-performance robots and aiming to the global optimality.

Optimal flight in the vertical plane with a vehicle model intermediate in complexity between the point-mass and energy models is studied. Flight-path angle takes on the role of a control variable. Range-open problems feature subarcs of vertical flight and singular subarcs as previously studied. The class of altitude-speed-range-time optimization problems with fuel expenditure unspecified is investigated and some interesting phenomena uncovered. The maximum-lift-to-drag glide appears as part of the family, final-time-open, with appropriate initial and terminal transient maneuvers. A family of, climb-range paths appears for thrust exceeding level-flight drag, some members exhibiting oscillations. Oscillatory paths generally fail the Jacobi test for durations exceeding a period and furnish a minimum only for short-duration problems. Minimizing paths of long duration follow a certain corridor in the V-h chart. The features of the family sharpen for the special case of thrust and drag independent of altitude, and considerable analytical attention is accorded to this for the insight it provides to the more general model. The problem of "steepest climb" is found to be ill-posed with the vehicle model under consideration, straight-vertically-upward maneuver sequences being furnished by a family of paths alternating between upward and downward vertical flight and including a limiting "chattering" member. / Ph. D.