Online Nonstochastic Control with Adversarial and Static Constraints

Authors: Xin Liu, Zixian Yang, Lei Ying

ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 5. Experiment In this section, we test our algorithms on a quadrotor vertical flight (QVF) control under an adversarial environment, which is modified from (Li et al., 2023). ... Figure 1 shows the experiment results for QVT control with winds blowing down wt U( 5.5, 4.5).
Researcher Affiliation Academia 1The School of Information Science and Technology, Shanghai Tech University, Shanghai, China. 2The Electrical Engineering and Computer Science Department, The University of Michigan, Ann Arbor, Ann Arbor, USA.
Pseudocode Yes Constrained Online Nonstochastic Control Algorithm Initialize: a (κ, ρ) stable controller K and the proper learning rates in COCO-Solver. for t = 1, , T, do Observe state xt and compute the disturbance wt 1. Apply control ut = Kxt + PH i=1 M[i] t wt i. Receive feedback including cost function ct(xt, ut) and constraint functions dt(xt, ut) and l(xt, ut). Compute the approximated cost function ct( ) and constraint functions dt( ) and l( ). Invoke the COCO-Solver(Mt, Qt, ct( ), dt( ), l( )) to obtain Mt+1 and Qt+1. end for
Open Source Code No The paper does not provide any statement or link regarding the public availability of its source code.
Open Datasets No The paper describes experiments on 'quadrotor vertical flight (QVF) control' and 'Heating Ventilation and Air Conditioning (HVAC) control', which are simulated environments or control systems rather than publicly available datasets with specific access information.
Dataset Splits No The paper describes control system simulations and experiments but does not provide specific dataset split information (e.g., percentages or counts for training, validation, or test sets).
Hardware Specification No The paper does not provide specific hardware details (such as exact GPU/CPU models, processor types, or memory) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes Let m = 1kg, g = 9.8m/s2, and Ia = 0.25kg/s. The system is discretized with t = 1s. We impose time-varying constraints, zt 0.3 + 0.3 sin(t/10), to emulate the complicated time-varying obstacles on the ground. The static affine constraints are zt 1.7 and 0 vt 12. We consider a time-varying quadratic cost function 0.1(zt 0.7)2 + 0.1 z2 t + χt(vt 9.8)2, where χt U(0.1, 0.2). We simulate two different wind conditions wt U( 5.5, 4.5) (winds blow down) and wt U(4.5, 5.5) (winds blow up), respectively.