Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Meta-Learning Augmented MPC for Disturbance-Aware Motion Planning and Control of Quadrotors

Dženan Lapandić, Fengze Xie, Christos K. Verginis, Soon-Jo Chung, Dimos V. Dimarogonas, Bo Wahlberg

TL;DR

A disturbance-aware motion planning and control framework for autonomous aerial flights composed of a disturbance-aware motion planner and a tracking controller that ensures safety bounds on the quadrotor’s behavior near obstacles with respect to the motion plan.

Abstract

A major challenge in autonomous flights is unknown disturbances, which can jeopardize safety and lead to collisions, especially in obstacle-rich environments. This paper presents a disturbance-aware motion planning and control framework designed for autonomous aerial flights. The framework is composed of two key components: a disturbance-aware motion planner and a tracking controller. The disturbance-aware motion planner consists of a predictive control scheme and a learned model of disturbances that is adapted online. The tracking controller is designed using contraction control methods to provide safety bounds on the quadrotor behaviour in the vicinity of the obstacles with respect to the disturbance-aware motion plan. Finally, the algorithm is tested in simulation scenarios with a quadrotor facing strong crosswind and ground-induced disturbances.

Meta-Learning Augmented MPC for Disturbance-Aware Motion Planning and Control of Quadrotors

TL;DR

A disturbance-aware motion planning and control framework for autonomous aerial flights composed of a disturbance-aware motion planner and a tracking controller that ensures safety bounds on the quadrotor’s behavior near obstacles with respect to the motion plan.

Abstract

A major challenge in autonomous flights is unknown disturbances, which can jeopardize safety and lead to collisions, especially in obstacle-rich environments. This paper presents a disturbance-aware motion planning and control framework designed for autonomous aerial flights. The framework is composed of two key components: a disturbance-aware motion planner and a tracking controller. The disturbance-aware motion planner consists of a predictive control scheme and a learned model of disturbances that is adapted online. The tracking controller is designed using contraction control methods to provide safety bounds on the quadrotor behaviour in the vicinity of the obstacles with respect to the disturbance-aware motion plan. Finally, the algorithm is tested in simulation scenarios with a quadrotor facing strong crosswind and ground-induced disturbances.
Paper Structure (11 sections, 3 theorems, 28 equations, 5 figures)

This paper contains 11 sections, 3 theorems, 28 equations, 5 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Notation
  4. Problem Formulation
  5. Meta-learning Augmented MPC
  6. Neural Network Model of Disturbances
  7. Contraction-Based Adaptive Controller (CBAC)
  8. Chance-constrained Upper Bound
  9. Optimal Control Problem as MPC
  10. Results
  11. Conclusion

Key Result

Theorem 1

Suppose there exists the contraction metric $M(x,x_d)\succ0$ and $M(x,x_d)=W^{-1}(x,x_d)$ obtained by solving Problem prob:W_convex_optimization for a given value of $\alpha\in(0,\infty)$ and that $\sup \| d'(x,x_d,u_d) \| \leq \bar{d}$. Suppose further that the system is controlled by the following where $e=x-x_d$, $K=R(x,x_d)^{-1}B(x)^TM(x,x_d)$, $\varphi=\varphi(x,x_d,x_r)$, $\phi = \phi(x,x_r)

Figures (5)

  • Figure 1: Block diagram of the proposed disturbance-aware motion planning and control algorithm.
  • Figure 2: Nominal MPC, RMSE: 0.29720
  • Figure 3: MPC+ML-CBAC, RMSE: 0.22470
  • Figure 4: Our algorithm, RMSE: 0.17714
  • Figure 6: (left) Nominal MPC fails to compensate, leaving a $25$ cm error; (right) our algorithm compensates and lands.

Theorems & Definitions (8)

  • Remark 1
  • Theorem 1
  • proof
  • Remark 2
  • Corollary 1
  • Remark 3
  • Corollary 2
  • Remark 4