In this thesis, a special control strategy is proposed which treats manipulators as a unique class of systems with their own unique characteristics. The objectives of proposing the control strategy are to obtain 1) no or tolerable overshoot; 2) smaller settling time; and 3) higher accuracy. Two two-link manipulators and one three-link manipulator were simulated and controlled by the proposed control strategy. The results are provided in the thesis. The proposed control strategy has the structure of local and global control. The local control takes care of the links. Its objective is to drive the links to the desired positions, disregarding the effects caused by gravity torques and the coupling and the interaction between links. The objective of the global control is to compensate for the effects ignored by the local control. The local control law has the feature of variable structure control. The decision concerning the control to be switched to depends on the region in the phase-plane the trajectory of the subsystem (link) is in. The switching time is determined by the state of the subsystem in the phase-plane via a simple criterion. A simplified model is also proposed and used to construct the control strategy. The model is built up with emphasis on the effective moment of inertia and gravity torques. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

An important problem in robotics research is the automatic off-line planning of optimal robot trajectories to perform specified tasks while satisfying physical constraints. This thesis proposes a method for finding an optimal geometric robot trajectory subject to the constraints of joint displacement limits and interference avoidance. A geometric method for calculating the distance between convex polyhedra is presented, and the method is implemented in two dimensions for the calculation of interference. Point-to-point trajectory planning is posed as a two-point boundary value problem in the calculus of variations. The kinematic constraints are formulated as exterior penalty functions and are combined with other optimization criteria to form a cost functional. The problem is solved by discretizing the problem and numerically minimizing the cost functional by using a steepest-descent approach to iteratively modify the trajectory. Any starting trajectory which satisfies the boundary conditions is acceptable, but different starting trajectories may converge to different locally optimal final trajectories. The method has been implemented for the two-dimensional case by an interactive FORTRAN program running on a VAX 11/750 computer. Successful results were obtained for a number of test cases, and further work has been identified to allow application of the method to a wide range of problems. / Applied Science, Faculty of / Mechanical Engineering, Department of / Graduate

This dissertation presents an approach to robot programming by demonstration based on two key concepts: demonstrator intent is the most meaningful signal that the robot can observe, and the robot should have a basic level of behavioral competency from which to interpret observed actions. Intent is a teleological, robust teaching signal invariant to many common sources of noise in training. The robot can use the knowledge encapsulated in sensorimotor schemas to interpret the demonstration. Furthermore, knowledge gained in prior demonstrations can be applied to future sessions. I argue that programming by demonstration be organized into declarative and procedural components. The declarative component represents a reusable outline of underlying behavior that can be applied to many different contexts. The procedural component represents the dynamic portion of the task that is based on features observed at run time. I describe how statistical models, and Bayesian methods in particular, can be used to model these components. These models have many features that are beneficial for learning in this domain, such as tolerance for uncertainty, and the ability to incorporate prior knowledge into inferences. I demonstrate this architecture through experiments on a bimanual humanoid robot using tasks from the pick and place domain. Additionally, I develop and experimentally validate a model for generating grasp candidates using visual features that is learned from demonstration data. This model is especially useful in the context of pick and place tasks.

Stereo vision uses two adjacent cameras to create a 3D image of the world. A depth map can be created by comparing the offset of the corresponding pixels from the two cameras. However, for real-time stereo vision, the image data needs to be processed at a reasonable frame rate. Real-time stereo vision allows for mobile robots to more easily navigate terrain and interact with objects by providing both the images from the cameras and the depth of the objects. Fortunately, the image processing can be parallelized in order to increase the processing speed. Field-programmable gate arrays (FPGAs) are highly parallelizable and lend themselves well to this problem. This thesis presents a stereo vision module which uses the Sum of Absolute Differences (SAD) algorithm. The SAD algorithm uses regions of pixels called windows to compare pixels to find matching pairs for determining depth. Two implementations are presented that utilize the SAD algorithm differently. The first implementation uses a 9x9 window for comparison and is able to process 4 pixels simultaneously. The second implementation uses a 7x7 window and processes 2 pixels simultaneously, but parallelizes each SAD algorithm for faster processing. The 9x9 implementation creates a better depth image with less noise, but the 7x7 implementation processes images at a higher frame rate. It has been shown through simulation that the 9x9 and 7x7 are able to process an image size of 640x480 at a frame rate of 15.73 and 29.32, respectively.

abstract: At modern-day intersections, traffic lights and stop signs assist human drivers to cross the intersection safely. Traffic congestion in urban road networks is a costly problem that affects all major cities. Efficiently operating intersections is largely dependent on accuracy and precision of human drivers, engendering a lingering uncertainty of attaining safety and high throughput. To improve the efficiency of the existing traffic network and mitigate the effects of human error in the intersection, many studies have proposed autonomous, intelligent transportation systems. These studies often involve utilizing connected autonomous vehicles, implementing a supervisory system, or both. Implementing a supervisory system is relatively more popular due to the security concerns of vehicle-to-vehicle communication. Even though supervisory systems are a step in the right direction for security, many supervisory systems’ safe operation solely relies on the promise of connected data being correct, making system reliability difficult to achieve. To increase fault-tolerance and decrease the effects of position uncertainty, this thesis proposes the Reliable and Robust Intersection Manager, a supervisory system that uses a separate surveillance system to dependably detect vehicles present in the intersection in order to create data redundancy for more accurate scheduling of connected autonomous vehicles. Adding the Surveillance System ensures that the temporal safety buffers between arrival times of connected autonomous vehicles are maintained. This guarantees that connected autonomous vehicles can traverse the intersection safely in the event of large vehicle controller error, a single rogue car entering the intersection, or a sybil attack. To test the proposed system given these fault-models, MATLAB® was used to create simulations in order to observe the functionality of R2IM compared to the state-of-the-art supervisory system, Robust Intersection Manager. Though R2IM is less efficient than the Robust Intersection Manager, it considers more fault models. The Robust Intersection Manager failed to maintain safety in the event of large vehicle controller errors and rogue cars, however R2IM resulted in zero collisions. / Dissertation/Thesis / Masters Thesis Computer Engineering 2019

Trajectory tracking is an integral component of most approaches to robot motion planning. There are a variety of techniques that have been developed over the years. However, without clear guidance in the literature, users are often left to guess what algorithm may work best for them. In this thesis, we quantitatively compare three trajectory tracking controllers for a differential drive wheeled robot: Linear Quadratic Regulator, Model Predictive Control and Sliding Mode Control. We compare the performance of these controllers through the metrics of tracking error, control energy, and rate of convergence. Through both simulations and experiments, our results show that LQR and MPC outperform SMC. In addition, SMC is difficult to tune and less robust than the other two. While LQR and MPC have similar performance. LQR is easier to implement while MPC enforces the physical constraints of the robot hardware.

The flexibility of animals limbs combined with the robustness of locomotion traversing unstructured terrains allow animals such as geckos and cockroaches to move dynamically in both horizontal and vertical domains. To capture this biological inspiration and the features which enable the performance, several approaches have been used including attempting to capture the morphological complexity of animals as well as determining shared fundamental dynamics of numerous animal in a single domain. These tools have lead to numerous robotics platforms along with the associated controllers which are able to either approach the flexibility or locomotive speeds of their biological inspiration. However, the robustness seen with biology has not been achieved with animal like speeds. This motivates the exploration of the dynamic understanding required to develop control which supplements and does not interfere with the passive dynamics representative of animal locomotion. This dissertation focuses on several topics related to the dynamics and control within the horizontal and vertical domains. These include defining three distinct dynamic gaits within the vertical domain (walking, running with a flight phase, and compliant running) along the associated implications of these gaits, determining the impact of power input on the gaits and stability selected as slope is varied from horizontal to vertical, and testing the robustness of a new dynamic decomposition based control for horizontal domain running. The resulting insights of this work will enable dynamic decomposition based control to be extended for the first time to the vertical domain and will provide insights into the design of future multi-modal robotic platforms / A Dissertation submitted to the Department of Mechanical Engineering in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Spring Semester 2019. / April 2, 2019. / Biologically Inspired Robotics, Climbing Robotics, Control, Dynamic, Legged Robotics, Simulation / Includes bibliographical references. / Jonathan Clark, Professor Directing Thesis; Rodney Roberts, University Representative; William Oates, Committee Member; Patrick Hollis, Committee Member; Camilo Ordonez, Committee Member.

Motion planning and controller design are essential in many autonomous applications, such as self-driving cars, surveillance with drones, and assistive robotics for home or medical applications. In these areas, the dynamical systems can be characterized as cyber-physical systems, in which the physical parts can be modeled by differential equations in continuous time, while digital computers are used to perform discrete state sampling and control updates. However, many existing frameworks were developed solely for the cyber part of the system. Ignoring continuous phenomena during discrete-time implementation could render undesired and even unsafe behavior. In this dissertation, I developed a motion planning and control framework for safety-critical systems to satisfy spatial and temporal constraints under a zeroth-order hold control implementation. In the first part of the dissertation, I introduce a self-triggered controller, with control barrier functions constraints for safety and control Lyapunov functions constraints for stability. The controls are generated by solving quadratic programs sequentially over time. Instead of using a periodic control mechanism, where the controls are updated at a constant rate, I proactively calculate the next update time instant, given the constraints, to ensure continuous-time safety and stability. In the second part of the dissertation, I introduce a correct-by-construction control synthesis framework where the system is required to satisfy Signal Temporal Logic formulas. I utilize the lower bounds of the control barrier functions and the predicate functions over a time interval to ensure continuous-time satisfaction of a formula. The focus of the third part of the dissertation is a sampling-based motion planner. The motion planner is based on Rapidly-exploring Random Trees. I introduce a function that generates collision-free trajectories and controls on-the-fly to avoid separate collision checks during edge extensions. The developed motion planner works for nonlinear systems and can be efficiently solved in real-time.

Robotic mapping has been a highly active research area in robotics and AI for at least two decades. Robot mapping addresses the problem of acquiring spatial models of physical environments through mobile robots. The mapping problem is generally regarded as one of the most important problems in the pursuit of building truly autonomous mobile robots. The effectiveness of current algorithms is driven in large part by the sheer volume of data available, acquired at high sample rates using sensors such as high-resolution laser rangefinders or cameras. Storing and operating on such large volumes of data can be costly in terms of storage space and computational power. While this may not be an issue for many robots, recently, there has been a surge of interest in micro aerial vehicles with limited onboard resources and payload capacity. It is not feasible for these systems to carry the high powered sensors that are typically needed for mapping or to store the large volumes of data algorithms currently need. The goal of this thesis is to study and evaluate compressive sensing methods to enable low resolution, lightweight sensors to create maps of sufficient quality for robot navigation tasks. The first part of this work focuses on two Compressive Sensing (CS)-based approaches for producing maps from highly sub-sampled data: 1) A total-variation minimization and 2) A basis pursuit problem. The total variation reconstruction is shown to provide high quality reconstruction of the map, but suffers from the fact that the regularization parameter is highly dependent on the structure of the mapped environment. The basis pursuit reconstruction, however, is more robust to the environment. The algorithm, previously introduced in the literature, uses a re-weighted basis pursuit algorithm to reconstruct a full lidar scan from a sub-sampled set of points. This reconstruction is then passed on to any standard mapping algorithm to produce a map of the environment. The re-weighting is done based on the slope of the depth profile around each sampled point. The second part of this work focuses on empirically evaluating the maps obtained by simulated experiments from the re-weighted basis pursuit algorithm and that obtained from linear interpolation as a function of the amount of sub-sampling. We considered three different subsampling approaches: subsampling in space (i.e., subsampling an individual lidar scan that in an ideal case would be comprised of 180 individual beams), and subsampling in time (i.e., using a reduced number of measurement locations in the environment), and a combination of both. For the evaluation metric, we picked 150 random pairs of points and constructed optimal paths between them using the A-star graph search algorithm, and compared how ``close'' the lengths of these paths were to the paths constructed on the ground truth map. The sum of squared error metric was used to measure this ``closeness''. In essence, if a reconstructed map contains too many false walls or randomly filled locations, the A-star paths should be longer than on the true map while if the reconstruction fails to produce complete walls, the paths should be shorter. The simulated experiments show that the re-weighted basis pursuit algorithm outperforms linear interpolation in severely under sampled datasets. The experiment also shows that the best trajectory for optimal reconstruction is if the robot stays close to the walls of the environment. / 2021-06-03T00:00:00Z

Collisions are typically seen as adverse events for quadrotors, with numerous researchers designing cages for minimizing the effect of impacts on the vehicles. In this thesis, we reverse this paradigm, treating collisions as an additional opportunity for localization. In order to gather collision information, a touch-only sensor with a protective frame is designed to sense the 2-D relative displacement due to inertial force between vehicle and the frame while collision happens. We provide a mathematical model that represents the displacement in terms of the poses of the protective frame and quadrotor is constructed and solved numerically, which helps analyze the distance from obstacles to the vehicle and collision direction. At last, a particle filter based localization observing the collision status is simulated, which verifies the theoretical feasibility utilizing collision information to perform localization in a known environment.