Matching Items (138)
Filtering by
- All Subjects: robotics
- Genre: Masters Thesis
- Creators: Arizona State University
- Status: Published
Description
Many real-world planning problems can be modeled as Markov Decision Processes (MDPs) which provide a framework for handling uncertainty in outcomes of action executions. A solution to such a planning problem is a policy that handles possible contingencies that could arise during execution. MDP solvers typically construct policies for a problem instance without re-using information from previously solved instances. Research in generalized planning has demonstrated the utility of constructing algorithm-like plans that reuse such information. However, using such techniques in an MDP setting has not been adequately explored.
This thesis presents a novel approach for learning generalized partial policies that can be used to solve problems with different object names and/or object quantities using very few example policies for learning. This approach uses abstraction for state representation, which allows the identification of patterns in solutions such as loops that are agnostic to problem-specific properties. This thesis also presents some theoretical results related to the uniqueness and succinctness of the policies computed using such a representation. The presented algorithm can be used as fast, yet greedy and incomplete method for policy computation while falling back to a complete policy search algorithm when needed. Extensive empirical evaluation on discrete MDP benchmarks shows that this approach generalizes effectively and is often able to solve problems much faster than existing state-of-art discrete MDP solvers. Finally, the practical applicability of this approach is demonstrated by incorporating it in an anytime stochastic task and motion planning framework to successfully construct free-standing tower structures using Keva planks.
This thesis presents a novel approach for learning generalized partial policies that can be used to solve problems with different object names and/or object quantities using very few example policies for learning. This approach uses abstraction for state representation, which allows the identification of patterns in solutions such as loops that are agnostic to problem-specific properties. This thesis also presents some theoretical results related to the uniqueness and succinctness of the policies computed using such a representation. The presented algorithm can be used as fast, yet greedy and incomplete method for policy computation while falling back to a complete policy search algorithm when needed. Extensive empirical evaluation on discrete MDP benchmarks shows that this approach generalizes effectively and is often able to solve problems much faster than existing state-of-art discrete MDP solvers. Finally, the practical applicability of this approach is demonstrated by incorporating it in an anytime stochastic task and motion planning framework to successfully construct free-standing tower structures using Keva planks.
ContributorsKala Vasudevan, Deepak (Author) / Srivastava, Siddharth (Thesis advisor) / Zhang, Yu (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2020
Description
Most planning agents assume complete knowledge of the domain, which may not be the case in scenarios where certain domain knowledge is missing. This problem could be due to design flaws or arise from domain ramifications or qualifications. In such cases, planning algorithms could produce highly undesirable behaviors. Planning with incomplete domain knowledge is more challenging than partial observability in the sense that the planning agent is unaware of the existence of such knowledge, in contrast to it being just unobservable or partially observable. That is the difference between known unknowns and unknown unknowns.
In this thesis, I introduce and formulate this as the problem of Domain Concretization, which is inverse to domain abstraction studied extensively before. Furthermore, I present a solution that starts from the incomplete domain model provided to the agent by the designer and uses teacher traces from human users to determine the candidate model set under a minimalistic model assumption. A robust plan is then generated for the maximum probability of success under the set of candidate models. In addition to a standard search formulation in the model-space, I propose a sample-based search method and also an online version of it to improve search time. The solution presented has been evaluated on various International Planning Competition domains where incompleteness was introduced by deleting certain predicates from the complete domain model. The solution is also tested in a robot simulation domain to illustrate its effectiveness in handling incomplete domain knowledge. The results show that the plan generated by the algorithm increases the plan success rate without impacting action cost too much.
In this thesis, I introduce and formulate this as the problem of Domain Concretization, which is inverse to domain abstraction studied extensively before. Furthermore, I present a solution that starts from the incomplete domain model provided to the agent by the designer and uses teacher traces from human users to determine the candidate model set under a minimalistic model assumption. A robust plan is then generated for the maximum probability of success under the set of candidate models. In addition to a standard search formulation in the model-space, I propose a sample-based search method and also an online version of it to improve search time. The solution presented has been evaluated on various International Planning Competition domains where incompleteness was introduced by deleting certain predicates from the complete domain model. The solution is also tested in a robot simulation domain to illustrate its effectiveness in handling incomplete domain knowledge. The results show that the plan generated by the algorithm increases the plan success rate without impacting action cost too much.
ContributorsSharma, Akshay (Author) / Zhang, Yu (Thesis advisor) / Fainekos, Georgios (Committee member) / Srivastava, Siddharth (Committee member) / Arizona State University (Publisher)
Created2020
Description
The human ankle is a critical joint required for mobility and stability of the body during static and dynamic activity. The absence of necessary torque output by the ankle due to neurological disorder or near-fatal injury can severely restrict locomotion and cause an inability to perform daily tasks. Physical Human-Robot Interaction (pHRI) has explored the potential of controlled actuators to positively impact human joints and partly restoring the required torque and stability at the joint to perform a task. However, a trade-off between agility and stability of the control technique of these devices can reduce the complete utilization of the performance to create a desirable impact on human joints. This research focuses on two control techniques of an Active Ankle Foot Orthosis (AFO) namely, Variable Stiffness (VS) and Variable Damping (VD) controllers to modulate ankle during walking. The VS controller is active during the stance phase and is used to restore the ankle trajectory of healthy participants that has been altered by adding a dead-weight of 2 Kgs. The VD controller is active during the terminal stance and early-swing phase and provides augmentative force during push-off that results in increased propulsion and stabilizes the ankle based on user-intuitions. Both controllers have a positive impact on Medial Gastrocnemius (GAS) muscle and Soleus (SOL) muscle which are powerful plantar - flexors critical to propulsion and kinematic properties during walking. The VS controller has recorded an 8.18% decrease in GAS and an 9.63 % decrease in SOL muscle activity during the stance phase amongst participants while decreasing mean ankle position error by 22.28 % and peak ankle position error by 17.43%. The VD controller demonstrated a 7.59 % decrease in GAS muscle and a 10.15 % decrease in SOL muscle activity during push-off amongst the participants while increasing the range-of-motion (ROM) by 7.84 %. Comprehensively, the study has shown a positive impact on ankle trajectory and the corresponding muscle effort at respective stages of the controller activity.
ContributorsSave, Omik Milind (Author) / Lee, Hyunglae (Thesis advisor) / Marvi, Hamidreza (Committee member) / Yong, Sze (Committee member) / Arizona State University (Publisher)
Created2020
Description
A novel underwater, open source, and configurable vehicle that mimics and leverages advances in quad-copter controls and dynamics, called the uDrone, was designed, built and tested. This vehicle was developed to aid coral reef researchers in collecting underwater spectroscopic data for the purpose of monitoring coral reef health. It is designed with an on-board integrated sensor system to support both automated navigation in close proximity to reefs and environmental observation. Additionally, the vehicle can serve as a testbed for future research in the realm of programming for autonomous underwater navigation and data collection, given the open-source simulation and software environment in which it was developed. This thesis presents the motivation for and design components of the new vehicle, a model governing vehicle dynamics, and the results of two proof-of-concept simulation for automated control.
ContributorsGoldman, Alex (Author) / Das, Jnaneshwar (Thesis advisor) / Asner, Greg (Committee member) / Marvi, Hamidreza (Committee member) / Arizona State University (Publisher)
Created2020
Description
Riding a bicycle requires accurately performing several tasks, such as balancing and navigation, which may be difficult or even impossible for persons with disabilities. These difficulties may be partly alleviated by providing active balance and steering assistance to the rider. In order to provide this assistance while maintaining free maneuverability, it is necessary to measure the position of the rider on the bicycle and to understand the rider's intent. Applying autonomy to bicycles also has the potential to address some of the challenges posed by traditional automobiles, including CO2 emissions, land use for roads and parking, pedestrian safety, high ownership cost, and difficulty traversing narrow or partially obstructed paths.
The Smart Bike research platform provides a set of sensors and actuators designed to aid in understanding human-bicycle interaction and to provide active balance control to the bicycle. The platform consists of two specially outfitted bicycles, one with force and inertial measurement sensors and the other with robotic steering and a control moment gyroscope, along with the associated software for collecting useful data and running controlled experiments. Each bicycle operates as a self-contained embedded system, which can be used for untethered field testing or can be linked to a remote user interface for real-time monitoring and configuration. Testing with both systems reveals promising capability for applications in human-bicycle interaction and robotics research.
The Smart Bike research platform provides a set of sensors and actuators designed to aid in understanding human-bicycle interaction and to provide active balance control to the bicycle. The platform consists of two specially outfitted bicycles, one with force and inertial measurement sensors and the other with robotic steering and a control moment gyroscope, along with the associated software for collecting useful data and running controlled experiments. Each bicycle operates as a self-contained embedded system, which can be used for untethered field testing or can be linked to a remote user interface for real-time monitoring and configuration. Testing with both systems reveals promising capability for applications in human-bicycle interaction and robotics research.
ContributorsBush, Jonathan Ernest (Author) / Zhang, Wenlong (Thesis advisor) / Heinrichs, Robert (Thesis advisor) / Sandy, Douglas (Committee member) / Arizona State University (Publisher)
Created2020
Description
As the world population continues to age, the demand for treatment and rehabilitation of long-term age-related ailments will rise. Healthcare technology must keep up with this demand, and existing solutions must become more readily available to the populace. Conditions such as impairment due to stroke currently take months or years of physical therapy to overcome, but rehabilitative exoskeletons can be used to greatly extend a physical therapist’s capabilities.
In this thesis, a rehabilitative knee exoskeleton was designed which is significantly lighter, more portable and less costly to manufacture than existing designs. It accomplishes this performance by making use of high-powered and weight-optimized brushless DC (BLDC) electric motors designed for drones, open-source hardware and software solutions for robotic motion control, and rapid prototyping technologies such as 3D printing and laser cutting.
The exoskeleton is made from a series of laser cut aluminum plates spaced apart with off-the-shelf standoffs. A drone motor with a torque of 1.32 Nm powers an 18.5:1 reduction two-stage belt drive, giving a maximum torque of 24.4 Nm at the output. The bearings for the belt drive are installed into 3D printed bearing mounts, which act as a snug intermediary between the bearing and the aluminum plate. The system is powered off a 24 volt, 1,500 MAh lithium battery, which can provide power for around an hour of walking activity.
The exoskeleton is controlled with an ODrive motor controller connected to a Raspberry Pi. Hip angle data is provided by an IMU, and the knee angle is provided by an encoder on the output shaft. A compact Rotary Series Elastic Actuator (cRSEA) device is mounted on the output shaft as well, to accurately measure the output torque going to the wearer. A Proportional-Derivative (PD) controller with feedforward relates the input current with the output torque. The device was tested on a treadmill and found to have an average backdrive torque of 0.39 Nm, significantly lower than the current state of the art. A gravity compensation controller and impedance controller were implemented to assist during swing and stance phases respectively. The results were compared to the muscular exertion of the knee measured via Electromyography (EMG).
In this thesis, a rehabilitative knee exoskeleton was designed which is significantly lighter, more portable and less costly to manufacture than existing designs. It accomplishes this performance by making use of high-powered and weight-optimized brushless DC (BLDC) electric motors designed for drones, open-source hardware and software solutions for robotic motion control, and rapid prototyping technologies such as 3D printing and laser cutting.
The exoskeleton is made from a series of laser cut aluminum plates spaced apart with off-the-shelf standoffs. A drone motor with a torque of 1.32 Nm powers an 18.5:1 reduction two-stage belt drive, giving a maximum torque of 24.4 Nm at the output. The bearings for the belt drive are installed into 3D printed bearing mounts, which act as a snug intermediary between the bearing and the aluminum plate. The system is powered off a 24 volt, 1,500 MAh lithium battery, which can provide power for around an hour of walking activity.
The exoskeleton is controlled with an ODrive motor controller connected to a Raspberry Pi. Hip angle data is provided by an IMU, and the knee angle is provided by an encoder on the output shaft. A compact Rotary Series Elastic Actuator (cRSEA) device is mounted on the output shaft as well, to accurately measure the output torque going to the wearer. A Proportional-Derivative (PD) controller with feedforward relates the input current with the output torque. The device was tested on a treadmill and found to have an average backdrive torque of 0.39 Nm, significantly lower than the current state of the art. A gravity compensation controller and impedance controller were implemented to assist during swing and stance phases respectively. The results were compared to the muscular exertion of the knee measured via Electromyography (EMG).
ContributorsParmentier, Robin W (Author) / Zhang, Wenlong (Thesis advisor) / Sugar, Thomas (Committee member) / Lee, Hyunglae (Committee member) / Arizona State University (Publisher)
Created2020
Description
Bicycles and motorcycles offer maneuverability, energy efficiency and acceleration that four wheeled vehicles cannot offer given similar budget for. Two wheeled vehicles have drastically different dynamics from four wheeled vehicles due to their instability and gyroscopic effect from their wheels.
This thesis focuses on self-stabilization of a motorcycle using an active control momentum gyroscope (CMG) and validation of this multi-degree-of-freedom system’s mathematical model. Physical platform was created to mimic the simulation as accurately as possible and all components used were justified. This process involves derivation of a 3 Degree-of-Freedom (DOF) system’s forward kinematics and its Jacobian matrix, simulation analysis of different controller algorithms, setting the system and subsystem specifications, and real system experimentation and data analysis.
A Jacobian matrix was used to calculate accurately decomposed resultant angular velocities which are used to create the dynamics model of the system torque using the Euler-Lagrange method. This produces a nonlinear second order differential equation that is modeled using MATLAB/Simulink. PID, and cascaded feedback loop are tested in this Simulink model. Cascaded feedback loop shows most promises in the simulation analysis. Therefore, system specifications are calculated according to the data produced by this controller method. The model validation is executed using the Vicon motion capture system which captured the roll angle of the motorcycle. This work contributes to creating a set of procedures for creating a validated dynamic model for a CMG stabilized motorcycle which can be used to create variants of other self-stabilizing motorcycle system.
This thesis focuses on self-stabilization of a motorcycle using an active control momentum gyroscope (CMG) and validation of this multi-degree-of-freedom system’s mathematical model. Physical platform was created to mimic the simulation as accurately as possible and all components used were justified. This process involves derivation of a 3 Degree-of-Freedom (DOF) system’s forward kinematics and its Jacobian matrix, simulation analysis of different controller algorithms, setting the system and subsystem specifications, and real system experimentation and data analysis.
A Jacobian matrix was used to calculate accurately decomposed resultant angular velocities which are used to create the dynamics model of the system torque using the Euler-Lagrange method. This produces a nonlinear second order differential equation that is modeled using MATLAB/Simulink. PID, and cascaded feedback loop are tested in this Simulink model. Cascaded feedback loop shows most promises in the simulation analysis. Therefore, system specifications are calculated according to the data produced by this controller method. The model validation is executed using the Vicon motion capture system which captured the roll angle of the motorcycle. This work contributes to creating a set of procedures for creating a validated dynamic model for a CMG stabilized motorcycle which can be used to create variants of other self-stabilizing motorcycle system.
ContributorsMoon, Hansol (Author) / Zhang, Wenlong (Thesis advisor) / Frank, Daniel (Committee member) / Delp, Deana (Committee member) / Sugar, Thomas (Committee member) / Arizona State University (Publisher)
Created2020
Description
In certain ant species, groups of ants work together to transport food and materials back to their nests. In some cases, the group exhibits a leader-follower behavior in which a single ant guides the entire group based on its knowledge of the destination. In some cases, the leader role is occupied temporarily by an ant, only to be replaced when an ant with new information arrives. This kind of behavior can be very useful in uncertain environments where robot teams work together to transport a heavy or bulky payload. The purpose of this research was to study ways to implement this behavior on robot teams.
In this work, I combined existing dynamical models of collective transport in ants to create a stochastic model that describes these behaviors and can be used to control multi-robot systems to perform collective transport. In this model, each agent transitions stochastically between roles based on the force that it senses the other agents are applying to the load. The agent’s motion is governed by a proportional controller that updates its applied force based on the load velocity. I developed agent-based simulations of this model in NetLogo and explored leader-follower scenarios in which agents receive information about the transport destination by a newly informed agent (leader) joining the team. From these simulations, I derived the mean allocations of agents between “puller” and “lifter” roles and the mean forces applied by the agents throughout the motion.
From the simulation results obtained, we show that the mean ratio of lifter to puller populations is approximately 1:1. We also show that agents using the role update procedure based on forces are required to exert less force than agents that select their role based on their position on the load, although both strategies achieve similar transport speeds.
In this work, I combined existing dynamical models of collective transport in ants to create a stochastic model that describes these behaviors and can be used to control multi-robot systems to perform collective transport. In this model, each agent transitions stochastically between roles based on the force that it senses the other agents are applying to the load. The agent’s motion is governed by a proportional controller that updates its applied force based on the load velocity. I developed agent-based simulations of this model in NetLogo and explored leader-follower scenarios in which agents receive information about the transport destination by a newly informed agent (leader) joining the team. From these simulations, I derived the mean allocations of agents between “puller” and “lifter” roles and the mean forces applied by the agents throughout the motion.
From the simulation results obtained, we show that the mean ratio of lifter to puller populations is approximately 1:1. We also show that agents using the role update procedure based on forces are required to exert less force than agents that select their role based on their position on the load, although both strategies achieve similar transport speeds.
ContributorsGah, Elikplim (Author) / Berman, Spring M (Thesis advisor, Committee member) / Pavlic, Theodore (Committee member) / Marvi, Hamidreza (Committee member) / Arizona State University (Publisher)
Created2020
Description
Robot motion planning requires computing a sequence of waypoints from an initial configuration of the robot to the goal configuration. Solving a motion planning problem optimally is proven to be NP-Complete. Sampling-based motion planners efficiently compute an approximation of the optimal solution. They sample the configuration space uniformly and hence fail to sample regions of the environment that have narrow passages or pinch points. These critical regions are analogous to landmarks from planning literature as the robot is required to pass through them to reach the goal.
This work proposes a deep learning approach that identifies critical regions in the environment and learns a sampling distribution to effectively sample them in high dimensional configuration spaces.
A classification-based approach is used to learn the distributions. The robot degrees of freedom (DOF) limits are binned and a distribution is generated from sampling motion plan solutions. Conditional information like goal configuration and robot location encoded in the network inputs showcase the network learning to bias the identified critical regions towards the goal configuration. Empirical evaluations are performed against the state of the art sampling-based motion planners on a variety of tasks requiring the robot to pass through critical regions. An empirical analysis of robotic systems with three to eight degrees of freedom indicates that this approach effectively improves planning performance.
This work proposes a deep learning approach that identifies critical regions in the environment and learns a sampling distribution to effectively sample them in high dimensional configuration spaces.
A classification-based approach is used to learn the distributions. The robot degrees of freedom (DOF) limits are binned and a distribution is generated from sampling motion plan solutions. Conditional information like goal configuration and robot location encoded in the network inputs showcase the network learning to bias the identified critical regions towards the goal configuration. Empirical evaluations are performed against the state of the art sampling-based motion planners on a variety of tasks requiring the robot to pass through critical regions. An empirical analysis of robotic systems with three to eight degrees of freedom indicates that this approach effectively improves planning performance.
ContributorsSrinet, Abhyudaya (Author) / Srivastava, Siddharth (Thesis advisor) / Zhang, Yu (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2020
Description
Despite the prevalence of teams in complex sociotechnical systems, current approaches to understanding workload tend to focus on the individual operator. However, research suggests that team workload has emergent properties and is not necessarily equivalent to the aggregate of individual workload. Assessment of communications provides a means of examining aspects of team workload in highly interdependent teams. This thesis set out to explore how communications are associated with team workload and performance under high task demand in all-human and human–autonomy teams in a command and control task. A social network analysis approach was used to analyze the communications of 30 different teams, each with three members operating in a command and control task environment of over a series of five missions. Teams were assigned to conditions differentiated by their composition with either a naïve participant, a trained confederate, or a synthetic agent in the pilot role. Social network analysis measures of centralization and intensity were used to assess differences in communications between team types and under different levels of demand, and relationships between communication measures, performance, and workload distributions were also examined. Results indicated that indegree centralization was greater in the all-human control teams than in the other team types, but degree centrality standard deviation and intensity were greatest in teams with a highly trained experimenter pilot. In all three team types, the intensity of communications and degree centrality standard deviation appeared to decrease during the high demand mission, but indegree and outdegree centralization did not. Higher communication intensity was associated with more efficient target processing and more successful target photos per mission, but a clear relationship between measures of performance and decentralization of communications was not found.
ContributorsJohnson, Craig Jonathon (Author) / Cooke, Nancy J. (Thesis advisor) / Gray, Robert (Committee member) / Gutzwiller, Robert S (Committee member) / Arizona State University (Publisher)
Created2020