BinghengNUS
Research Fellow (PhD) in electrical and computer engineering with focus on learning-based control
National University of Singapore
Pinned Repositories
Auto-Multilift
Auto-Multilift is a novel learning framework for cooperative load transportation with quadrotors. It can automatically tune various MPC hyperparameters, which are modeled by DNNs and difficult to tune manually, via reinforcement learning in a distributed and closed-loop manner.
Data_sample_SIMULINK_Lab_Section2_EE4307
This file provides those who take EE4307 with the data and sample code for section 2 of the lab.
DMHE
Differentiable moving horizon estimation (DMHE) is an auto-tuning optimal estimation algorithm. This source code is a supplementery material of the accepted paper for the 60th IEEE Conference on Decision and Control.
LearningAgileFlight_SE3
Learning Agile Flights on SE(3): a novel deep SE(3) motion planning and control method for quadrotors. It learns an MPC's adaptive SE(3) decision variables parameterized by a portable DNN, encouraging the quadrotor to fly through the gate with maximum safety margins under diverse settings.
ME5701
NeuroMHE
Neural Moving Horizon Estimation (NeuroMHE) is an auto-tuning and adaptive optimal estimator. It fuses a nueral network with an MHE to render fast online adaptation to state-dependent noise. The neural network can be efficiently trained from the robot's trajectory tracking errors without the need for the ground truth data.
NeuroPID_EE4307_2022
Dear Class of EE4307 (Control System Design and Simulation, NUS) Please find the uploaded sample python code for the mini-project in your lab manual. You can use this code but are encouraged to modify it yourselves according to the control specification discussed in the manual. Best, Bingheng (Graduate Assistant of EE4307)
SUSTech_NeZha
Scource code of the paper entitled 'Underactuated Motion Planning and Control for Jumping With Wheeled-Bipedal Robots'
TR-NeuroMHE
NeuroMHE
Neural Moving Horizon Estimation (NeuroMHE) is an auto-tuning and adaptive optimal estimator. It fuses a nueral network with an MHE to render fast online adaptation to state-dependent noise. The neural network can be efficiently trained from the robot's trajectory tracking errors without the need for the ground truth data.
BinghengNUS's Repositories
BinghengNUS/LearningAgileFlight_SE3
Learning Agile Flights on SE(3): a novel deep SE(3) motion planning and control method for quadrotors. It learns an MPC's adaptive SE(3) decision variables parameterized by a portable DNN, encouraging the quadrotor to fly through the gate with maximum safety margins under diverse settings.
BinghengNUS/TR-NeuroMHE
BinghengNUS/NeuroPID_EE4307_2022
Dear Class of EE4307 (Control System Design and Simulation, NUS) Please find the uploaded sample python code for the mini-project in your lab manual. You can use this code but are encouraged to modify it yourselves according to the control specification discussed in the manual. Best, Bingheng (Graduate Assistant of EE4307)
BinghengNUS/SUSTech_NeZha
Scource code of the paper entitled 'Underactuated Motion Planning and Control for Jumping With Wheeled-Bipedal Robots'
BinghengNUS/Auto-Multilift
Auto-Multilift is a novel learning framework for cooperative load transportation with quadrotors. It can automatically tune various MPC hyperparameters, which are modeled by DNNs and difficult to tune manually, via reinforcement learning in a distributed and closed-loop manner.
BinghengNUS/Data_sample_SIMULINK_Lab_Section2_EE4307
This file provides those who take EE4307 with the data and sample code for section 2 of the lab.
BinghengNUS/DMHE
Differentiable moving horizon estimation (DMHE) is an auto-tuning optimal estimation algorithm. This source code is a supplementery material of the accepted paper for the 60th IEEE Conference on Decision and Control.
BinghengNUS/ME5701
BinghengNUS/NeuroMHE
Neural Moving Horizon Estimation (NeuroMHE) is an auto-tuning and adaptive optimal estimator. It fuses a nueral network with an MHE to render fast online adaptation to state-dependent noise. The neural network can be efficiently trained from the robot's trajectory tracking errors without the need for the ground truth data.