Matching Items (4)
Filtering by
- All Subjects: Reinforcement Learning
- Genre: Academic theses
- Creators: Si, Jennie
- Creators: Yang, Yezhou
- Status: Published
Description
This dissertation focuses on reinforcement learning (RL) controller design aiming for real-life applications in continuous state and control problems. It involves three major research investigations in the aspect of design, analysis, implementation, and evaluation. The application case addresses automatically configuring robotic prosthesis impedance parameters. Major contributions of the dissertation include the following. 1) An “echo control” using the intact knee profile as target is designed to overcome the limitation of a designer prescribed robotic knee profile. 2) Collaborative multiagent reinforcement learning (cMARL) is proposed to directly take into account human influence in the robot control design. 3) A phased actor in actor-critic (PAAC) reinforcement learning method is developed to reduce learning variance in RL. The design of an “echo control” is based on a new formulation of direct heuristic dynamic programming (dHDP) for tracking control of a robotic knee prosthesis to mimic the intact knee profile. A systematic simulation of the proposed control is provided using a human-robot system simulation in OpenSim. The tracking controller is then tested on able-bodied and amputee subjects. This is the first real-time human testing of RL tracking control of a robotic knee to mirror the profile of an intact knee. The cMARL is a new solution framework for the human-prosthesis collaboration (HPC) problem. This is the first attempt at considering human influence on human-robot walking with the presence of a reinforcement learning controlled lower limb prosthesis. Results show that treating the human and robot as coupled and collaborating agents and using an estimated human adaptation in robot control design help improve human walking performance. The above studies have demonstrated great potential of RL control in solving continuous problems. To solve more complex real-life tasks with multiple control inputs and high dimensional state space, high variance, low data efficiency, slow learning or even instability are major roadblocks to be addressed. A novel PAAC method is proposed to improve learning performance in policy gradient RL by accounting for both Q value and TD error in actor updates. Systematical and comprehensive demonstrations show its effectiveness by qualitative analysis and quantitative evaluation in DeepMind Control Suite.
ContributorsWu, Ruofan (Author) / Si, Jennie (Thesis advisor) / Huang, He (Committee member) / Santello, Marco (Committee member) / Papandreou- Suppappola, Antonia (Committee member) / Arizona State University (Publisher)
Created2023
Description
Robotic lower limb prostheses provide new opportunities to help transfemoral amputees regain mobility. However, their application is impeded by that the impedance control parameters need to be tuned and optimized manually by prosthetists for each individual user in different task environments. Reinforcement learning (RL) is capable of automatically learning from interacting with the environment. It becomes a natural candidate to replace human prosthetists to customize the control parameters. However, neither traditional RL approaches nor the popular deep RL approaches are readily suitable for learning with limited number of samples and samples with large variations. This dissertation aims to explore new RL based adaptive solutions that are data-efficient for controlling robotic prostheses.
This dissertation begins by proposing a new flexible policy iteration (FPI) framework. To improve sample efficiency, FPI can utilize either on-policy or off-policy learning strategy, can learn from either online or offline data, and can even adopt exiting knowledge of an external critic. Approximate convergence to Bellman optimal solutions are guaranteed under mild conditions. Simulation studies validated that FPI was data efficient compared to several established RL methods. Furthermore, a simplified version of FPI was implemented to learn from offline data, and then the learned policy was successfully tested for tuning the control parameters online on a human subject.
Next, the dissertation discusses RL control with information transfer (RL-IT), or knowledge-guided RL (KG-RL), which is motivated to benefit from transferring knowledge acquired from one subject to another. To explore its feasibility, knowledge was extracted from data measurements of able-bodied (AB) subjects, and transferred to guide Q-learning control for an amputee in OpenSim simulations. This result again demonstrated that data and time efficiency were improved using previous knowledge.
While the present study is new and promising, there are still many open questions to be addressed in future research. To account for human adaption, the learning control objective function may be designed to incorporate human-prosthesis performance feedback such as symmetry, user comfort level and satisfaction, and user energy consumption. To make the RL based control parameter tuning practical in real life, it should be further developed and tested in different use environments, such as from level ground walking to stair ascending or descending, and from walking to running.
This dissertation begins by proposing a new flexible policy iteration (FPI) framework. To improve sample efficiency, FPI can utilize either on-policy or off-policy learning strategy, can learn from either online or offline data, and can even adopt exiting knowledge of an external critic. Approximate convergence to Bellman optimal solutions are guaranteed under mild conditions. Simulation studies validated that FPI was data efficient compared to several established RL methods. Furthermore, a simplified version of FPI was implemented to learn from offline data, and then the learned policy was successfully tested for tuning the control parameters online on a human subject.
Next, the dissertation discusses RL control with information transfer (RL-IT), or knowledge-guided RL (KG-RL), which is motivated to benefit from transferring knowledge acquired from one subject to another. To explore its feasibility, knowledge was extracted from data measurements of able-bodied (AB) subjects, and transferred to guide Q-learning control for an amputee in OpenSim simulations. This result again demonstrated that data and time efficiency were improved using previous knowledge.
While the present study is new and promising, there are still many open questions to be addressed in future research. To account for human adaption, the learning control objective function may be designed to incorporate human-prosthesis performance feedback such as symmetry, user comfort level and satisfaction, and user energy consumption. To make the RL based control parameter tuning practical in real life, it should be further developed and tested in different use environments, such as from level ground walking to stair ascending or descending, and from walking to running.
ContributorsGao, Xiang (Author) / Si, Jennie (Thesis advisor) / Huang, He Helen (Committee member) / Santello, Marco (Committee member) / Papandreou-Suppappola, Antonia (Committee member) / Arizona State University (Publisher)
Created2020
Description
The goal of reinforcement learning is to enable systems to autonomously solve tasks in the real world, even in the absence of prior data. To succeed in such situations, reinforcement learning algorithms collect new experience through interactions with the environment to further the learning process. The behaviour is optimized by maximizing a reward function, which assigns high numerical values to desired behaviours. Especially in robotics, such interactions with the environment are expensive in terms of the required execution time, human involvement, and mechanical degradation of the system itself. Therefore, this thesis aims to introduce sample-efficient reinforcement learning methods which are applicable to real-world settings and control tasks such as bimanual manipulation and locomotion. Sample efficiency is achieved through directed exploration, either by using dimensionality reduction or trajectory optimization methods. Finally, it is demonstrated how data-efficient reinforcement learning methods can be used to optimize the behaviour and morphology of robots at the same time.
ContributorsLuck, Kevin Sebastian (Author) / Ben Amor, Hani (Thesis advisor) / Aukes, Daniel (Committee member) / Fainekos, Georgios (Committee member) / Scholz, Jonathan (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2019
Description
Traditional Reinforcement Learning (RL) assumes to learn policies with respect to reward available from the environment but sometimes learning in a complex domain requires wisdom which comes from a wide range of experience. In behavior based robotics, it is observed that a complex behavior can be described by a combination of simpler behaviors. It is tempting to apply similar idea such that simpler behaviors can be combined in a meaningful way to tailor the complex combination. Such an approach would enable faster learning and modular design of behaviors. Complex behaviors can be combined with other behaviors to create even more advanced behaviors resulting in a rich set of possibilities. Similar to RL, combined behavior can keep evolving by interacting with the environment. The requirement of this method is to specify a reasonable set of simple behaviors. In this research, I present an algorithm that aims at combining behavior such that the resulting behavior has characteristics of each individual behavior. This approach has been inspired by behavior based robotics, such as the subsumption architecture and motor schema-based design. The combination algorithm outputs n weights to combine behaviors linearly. The weights are state dependent and change dynamically at every step in an episode. This idea is tested on discrete and continuous environments like OpenAI’s “Lunar Lander” and “Biped Walker”. Results are compared with related domains like Multi-objective RL, Hierarchical RL, Transfer learning, and basic RL. It is observed that the combination of behaviors is a novel way of learning which helps the agent achieve required characteristics. A combination is learned for a given state and so the agent is able to learn faster in an efficient manner compared to other similar approaches. Agent beautifully demonstrates characteristics of multiple behaviors which helps the agent to learn and adapt to the environment. Future directions are also suggested as possible extensions to this research.
ContributorsVora, Kevin Jatin (Author) / Zhang, Yu (Thesis advisor) / Yang, Yezhou (Committee member) / Praharaj, Sarbeswar (Committee member) / Arizona State University (Publisher)
Created2021