Neural Moving Horizon Estimation for Robust Flight Control

@article{Wang2022NeuralMH,
  title={Neural Moving Horizon Estimation for Robust Flight Control},
  author={Bingheng Wang and Zhengtian Ma and Shupeng Lai and Lin Zhao},
  journal={ArXiv},
  year={2022},
  volume={abs/2206.10397}
}
—Estimating and reacting to external disturbances is crucial for robust flight control of quadrotors. Existing estimators typically require significant tuning for a specific flight scenario or training with extensive ground-truth disturbance data to achieve satisfactory performance. In this paper, we propose a neural moving horizon estimator (NeuroMHE) that can automatically tune the key parameters modeled by a neural network and adapt to different flight scenarios. We achieve this by deriving the… 

References

SHOWING 1-10 OF 52 REFERENCES

Differentiable Moving Horizon Estimation for Robust Flight Control

TLDR
A data-efficient differentiable moving horizon estimation (DMHE) algorithm that can automatically tune the MHE parameters online and also adapt to different scenarios and a model-based policy gradient algorithm to learn the parameters directly from the trajectory tracking errors without the need for the ground truth is developed.

Learning-based Moving Horizon Estimation through Differentiable Convex Optimization Layers

TLDR
This work addresses the problem of state estimation in the case of constrained linear systems with parametric uncertainty, and makes use of differentiable convex optimization layers to formulate an MHE state estimator for systems with uncertain parameters.

External Force Estimation based on Nonlinear Moving Horizon Estimation for MAV Navigation

TLDR
A novel Nonlinear Moving Horizon Estimation (NMHE) for estimating the external forces without adding extra sensor to the MAV or using external sensor for force measurements, while avoiding the need to know aerodynamic models or other parameters of theMAV is proposed.

Reinforcement Learning based on MPC/MHE for Unmodeled and Partially Observable Dynamics

TLDR
For the sake of tackling the unmodeled and partially observable dynamics, the Reinforcement Learning (RL) is leveraged to tune the parameters of MPC and MHE schemes jointly, with the closed-loop performance of the policy as a goal rather than model fitting or the MHE performance.

Performance, Precision, and Payloads: Adaptive Nonlinear MPC for Quadrotors

TLDR
This letter proposes a novel hybrid adaptive NMPC to learn model uncertainties online and immediately compensate for them, drastically improving performance over the non-adaptive baseline with minimal computational overhead.

Quadcopter Trajectory Tracking in the Presence of 4 Faulty Actuators: A Nonlinear MHE and MPC Approach

TLDR
The proposed framework is capable of achieving trajectory tracking, estimating states, as well as estimating and compensating for all four actuators multiplicative faults, and combines two optimization-based nonlinear estimation and control techniques; Nonlinear Moving Horizon Estimation (NMHE) and Nonlinear Model Predictive Control (NMPC).

Differentiable MPC for End-to-end Planning and Control

TLDR
The foundations for using Model Predictive Control as a differentiable policy class for reinforcement learning in continuous state and action spaces are presented and it is shown that the MPC policies are significantly more data-efficient than a generic neural network and that the method is superior to traditional system identification in a setting where the expert is unrealizable.

NeuroBEM: Hybrid Aerodynamic Quadrotor Model

TLDR
This work proposes a hybrid approach fusing first principles and learning to model quadrotors and their aerodynamic effects with unprecedented accuracy, outperforming existing models with 50% reduced prediction errors, and shows strong generalization capabilities beyond the training set.

A Lyapunov function for robust stability of moving horizon estimation

We provide a novel robust stability analysis for moving horizon estimation (MHE) using a Lyapunov function. Additionally, we introduce linear matrix inequalities (LMIs) to verify the necessary
...