The aim of this thesis is the development and implementation of a controlled testing platform for the Robust Intelligent Sensing and Controls (RISC) Lab at Utah State University (USU). This will be an open source adaptable expandable robotics platform usable for both education and research. This differs from the many other platforms developed in that the entire platform software will be made open source. This open source software will encourage collaboration among other universities and enable researchers to essentially pick up where others have left off without the necessity of replicating months or even years of work. The expected results of this research will create a foundation for diverse robotics investigation at USU as well as enable attempts at novel methods of control, estimation and optimization. This will also contribute a complete software testbed setup to the already vibrant robotics open source research community. This thesis first outlines the platform setup and novel developments therein. The second stage provides an example of how this has been used in education, providing an example curriculum implementing modern control techniques. The third section provides some exploratory research in trajectory control and state estimation of the tip of an inverted pendulum atop a small unmanned aerial vehicle as well as bearing-only cooperative localization experimentation. Finally, a conclusion and future work is discussed.

No description available.

In this paper, the use of reinforcement learning (RL) in control systems is investigated using a rotatory inverted pendulum as an example. The control behavior of an RL controller is compared to that of traditional LQR and MPC controllers. This is done by evaluating their behavior under optimal conditions, their disturbance behavior, their robustness and their development process. All the investigated controllers are developed using MATLAB and the Simulink simulation environment and later deployed to a real pendulum model powered by a Raspberry Pi. The RL algorithm used is Proximal Policy Optimization (PPO). The LQR controller exhibits an easy development process, an average to good control behavior and average to good robustness. A linear MPC controller could show excellent results under optimal operating conditions. However, when subjected to disturbances or deviations from the equilibrium point, it showed poor performance and sometimes instable behavior. Employing a nonlinear MPC Controller in real time was not possible due to the high computational effort involved. The RL controller exhibits by far the most versatile and robust control behavior. When operated in the simulation environment, it achieved a high control accuracy. When employed in the real system, however, it only shows average accuracy and a significantly greater performance loss compared to the simulation than the traditional controllers. With MATLAB, it is not yet possible to directly post-train the RL controller on the Raspberry Pi, which is an obstacle to the practical application of RL in a prototyping or teaching setting. Nevertheless, RL in general proves to be a flexible and powerful control method, which is well suited for complex or nonlinear systems where traditional controllers struggle.

No description available.

No description available.

No description available.

The Laser Interferometer Space Antenna (LISA) will be the first gravitational wave detector in space. The European Space Agency has selected LISA as a large mission scheduled to launch in the mid-2030s. The sensitivity of LISA to gravitational waves is limited at low frequencies by force disturbances acting on the otherwise free-falling test masses. Among the stray forces relevant to the LISA noise budget, we find the ones that arise from the electrostatic interaction between the test masses and the surrounding capacitive sensor. Most of such electrostatic forces scale with the electric charge deposited on the test masses. This problem is aggravated by the fact that isolated objects in space, such as the floating LISA test masses, accumulate electric charge due to the constant bombardment of cosmic rays and solar energetic particles. We, therefore, understand that if the test masses were not discharged, the electrostatic disturbances could spoil the performance of the whole mission at low frequency. The precursor LISA Pathfinder (LPF) mission proved that the test mass charge could be successfully managed with a contactless system based on photoelectric charge transfer. The light sources required for photoemission in LISA Pathfinder were mercury-vapor lamps emitting photons in the UV range. In this thesis, we will present our on-ground testing campaign of a prototype Charge Management System for LISA, which relies on UV-LEDs as light sources. LEDs, compared to mercury-vapor lamps, can emit short pulses of UV light (~10 ns), which can be synchronized with the time-varying electrostatic fields around the test mass. For this reason, we studied new discharge strategies made possible by adopting UV-LEDs characterized by pulsed illumination synced with the capacitive sensing injection bias. Our measurements indicate that UV-LEDs offer significant advantages regarding the flexibility and robustness of the Charge Management System. Moreover, the new illumination patterns offered by UV-LEDs allow fine-tuning the TM equilibrium potential without introducing local DC fields, easing the implementation of the continuous discharge mode to manage the TM potential. Finally, we investigated the charge noise introduced by the continuous discharge mode and verified that it could be kept within the LISA requirements. We will present hereafter the outline of the thesis. In the first chapter, we present a mandatory introduction to gravitational waves and the LISA mission. In the second chapter, we present the instrument used for our experimental campaign, namely the four-test-masses torsion pendulum at the University of Trento. We also present the electrostatic model and the measurement techniques used to evaluate the electric charge on the pendulum test mass. In the third chapter, we introduce the concept of apparent yield, which is a figure of merit of the charge management system performance. We also present our experimental measurement, which encompasses tests on several UV-LEDs in different illumination patterns. In the fourth chapter, we derive a simple photoemission model, which is useful for interpreting the apparent yield data acquired. We will also use the model to fit the experimental data and extract estimates of the microscopic parameters that affect the photoemission from metallic surfaces, e.g. work function or quantum yield. In the fifth chapter, we present a model and our torsion pendulum measurements for the charge noise induced on the test masses when continuously illuminated with UV light. Such noise arises from the discrete and intrinsically stochastic nature of photoelectric charge transfer. Finally, in the last chapter, we will wrap up by presenting the problems encountered and the "lessons learned" during the years-long experimental endeavor.

M.S.

A novel torsion pendulum thrust test stand for micro-Newton-scale spacecraft thrusters is described. The stand is designed to be robust against electromagnetic interference effects internal or external to the thruster being tested. The design and testing of a gas-dynamic calibration thruster is included. This thruster is fully self-contained on the pendulum arm, with no external wires or feedlines connected to the device and impacting the dynamic response of the underlying pendulum. Initial calibration results are shown. Zero drift and hysteresis are present in the results, evidenced by a constant steady-state displacement drift and a return to a different displacement after shutdown of the calibration thruster. Results are compared to theoretical solutions of the equation of motion. An external forcing function of facility effects is described for discrepancies between results and the theoretical solution. Further work to eliminate these effects and add damping are proposed. / MS / Many recently proposed space missions require very fine vehicle attitude and position control in support of their science objectives. Thrusters with the ability to provide this control are currently in development, from laboratory proofs of concept to initial test flights on pathfinding missions. The low levels of thrust produced by these devices, in the range of less than the weight of a mosquito, require specialized test stands with very fine resolution. This thesis describes a novel torsion pendulum design for measuring these thrusters as well as initial validation results from its calibration system using rarefied gas flow. This calibration device is fully-contained on the device’s arm, removing many common sources of compensation factors which are often needed for other test stand designs. A custom-built displacement measuring system for determining angular motion of the pendulum arm is described which allows for measuring angular displacements of the arm to the level of arcseconds and potentially fractions thereof. Initial results suggest measurement of the expected levels of thrust, while some work remains to remove lingering sources of error and achieve more precise thrust data.

A pendulum with its supporting point vibrating in both the x and the y direction is analyzed. Numerical simulation by computer is used to analyze the motion of the pendulum. Chaotic motion of the system is observed. Threshold values for chaos are obtained by simulation. The Lyapunov exponent and the fast Fourier transform ( FFT ) are used as the criteria to determine if the system is chaotic. Two predictive theoretical criteria, the Melnikov criterion and a period-doubling criterion, are then applied to the system. The results obtained by simulation and by theoretical criteria are shown to be in good agreement. A brute-force approach is used to supplement the results. It is found that the motion of this simple driven pendulum will have very complicated behavior. Multiple attractors can be shown to coexist. / Master of Science