Recovery of Desired Flying Characteristics with an L1 Adaptive Control Law: Flight Test Results on Calspan's VSS Learjet

Ackerman, Kasey A.; Navarro, Javier Puig; Hovakimyan, Naira; Cotting, M. Christopher; Duke, Dustin J.; Carrera, Miguel J.; McCaskey, Nathan C.; Esposito, Dario; Peterson, Jessica M.; Tellefsen, Jonathan R.

doi:10.2514/6.2019-1084

Cited by 13 publications

(6 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Articulating the mission, goals and objectives, roles and responsibilities, phases and stages, and implementation in a careful and planned manner are important factors in achieving goals (Schnugg 2019). This approach is also referred to as "planning from goals" (Ackerman et al 2019). The second approach, also referred to as the "sudden" approach, is called "planning from thrust" (Ackerman et al 2019).…”

Section: Planningmentioning

confidence: 99%

Untitled

Bokayev¹,

Kaimoldiyev²

2022

VPA

View full text Add to dashboard Cite

The purpose of this study is to analyze collaboration in the implementation of food security policies in Indonesia. This research method uses qualitative research design with a case study research strategy. The respondents in this study were: the Regent, officials, and employees of the Food Security Service; members of the DPRD (Regional People's Representative Council); business actors in the food sector; farmers (Farmer Group Association/Gapoktan); and the community as food consumers. In this study, the data collection techniques used were: observation, in-depth interviews, FGD, and document analysis. The results showed cross-sector collaboration in the implementation of food security policies and collaborative policy implementation involving various stakeholders including food agencies, agricultural agencies, sub-district/village governments, agricultural extension agents, POKTAN (farmer groups), and economic actors related to this. Conflict of interest between these stakeholders exists, and sectoral egos are still prioritized through stakeholders' respective programs. The result is that no program planning has emerged as a result of the accumulation of resources. Therefore, for the success of a policy that involves multiple actors, collaborative policy implementation is needed that enables a collaborative program to be created so that local food government can be realized to overcome food problems at both the local and national levels.

show abstract

Section: Planningmentioning

confidence: 99%

Untitled

Bokayev¹,

Kaimoldiyev²

2022

VPA

View full text Add to dashboard Cite

show abstract

“…Remark 4. Variations of the proposed L 1 AC law ( 4)-( 6) have been used to augment other baseline controllers (e.g., PID, linear quadratic regulator, MPC), as demonstrated in numerous applications and flight tests, [13], [34], [35].…”

Section: L 1 Augmentation For Policy Robustificationmentioning

confidence: 99%

Robustifying Reinforcement Learning Policies with $\mathcal{L}_1$ Adaptive Control

Cheng¹,

Zhao²,

Gandhi³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

A reinforcement learning (RL) policy trained in a nominal environment could fail in a new/perturbed environment due to the existence of dynamic variations. Existing robust methods try to obtain a fixed policy for all envisioned dynamic variation scenarios through robust or adversarial training. These methods could lead to conservative performance due to emphasis on the worst case, and often involve tedious modifications to the training environment. We propose an approach to robustifying a pre-trained non-robust RL policy with L1 adaptive control (L1AC). Leveraging the capability of an L1AC law in fast estimation of and active compensation for dynamic variations, our approach can significantly improve the robustness of an RL policy trained in a standard (i.e., non-robust) way, either in a simulator or in the real world. Numerical experiments are provided to validate the efficacy of the proposed approach.

show abstract

“…Remark 5. Variations of the proposed L 1 AC law ( 9), (10) and (12) have been used to augment other baseline controllers (e.g., PID, linear quadratic regulator, MPC), as demonstrated in numerous applications and flight tests, [15], [36], [37].…”

Section: L 1 Adaptive Augmentation For Policy Robustificationmentioning

confidence: 99%

Improving the Robustness of Reinforcement Learning Policies with $\mathcal{L}_{1}$ Adaptive Control

Cheng,

Zhao,

Wang

et al. 2021

Preprint

View full text Add to dashboard Cite

A reinforcement learning (RL) control policy trained in a nominal environment could fail in a new/perturbed environment due to the existence of dynamic variations. For controlling systems with continuous state and action spaces, we propose an add-on approach to robustifying a pre-trained RL policy by augmenting it with an L1 adaptive controller (L1AC). Leveraging the capability of an L1AC for fast estimation and active compensation of dynamic variations, the proposed approach can improve the robustness of an RL policy which is trained either in a simulator or in the real world without consideration of a broad class of dynamic variations. Numerical and real-world experiments empirically demonstrate the efficacy of the proposed approach in robustifying RL policies trained using both model-free and model-based methods. A video for the experiments on a real Pendubot setup is available at https://youtu.be/xgOB9vpyUgE.

show abstract

Recovery of Desired Flying Characteristics with an L1 Adaptive Control Law: Flight Test Results on Calspan's VSS Learjet

Cited by 13 publications

References 9 publications

Untitled

Untitled

Robustifying Reinforcement Learning Policies with $\mathcal{L}_1$ Adaptive Control

Improving the Robustness of Reinforcement Learning Policies with $\mathcal{L}_{1}$ Adaptive Control

Contact Info

Product

Resources

About