Progressive Smoothing for Motion Planning in Real-Time NMPC

Rudolf Reiter,Katrin Baumgärtner,Rien Quirynen,Moritz Diehl

2024-03-04

Abstract:Nonlinear model predictive control (NMPC) is a popular strategy for solving motion planning problems, including obstacle avoidance constraints, in autonomous driving applications. Non-smooth obstacle shapes, such as rectangles, introduce additional local minima in the underlying optimization problem. Smooth over-approximations, e.g., ellipsoidal shapes, limit the performance due to their conservativeness. We propose to vary the smoothness and the related over-approximation by a homotopy. Instead of varying the smoothness in consecutive sequential quadratic programming iterations, we use formulations that decrease the smooth over-approximation from the end towards the beginning of the prediction horizon. Thus, the real-time iterations algorithm is applicable to the proposed NMPC formulation. Different formulations are compared in simulation experiments and shown to successfully improve performance indicators without increasing the computation time.

Systems and Control,Optimization and Control

What problem does this paper attempt to address?

### Problems Addressed by the Paper The paper aims to address issues encountered in Nonlinear Model Predictive Control (NMPC) for real-time applications, particularly in obstacle avoidance for motion planning of autonomous vehicles. Specifically: 1. **Smoothing of Obstacle Shapes**: - Non-smooth obstacle shapes (e.g., rectangular) introduce additional local minima in the optimization problem. - Smooth approximations like ellipses, while performing better, are overly conservative and impact performance. 2. **Method for Varying Smoothness**: - A method is proposed to gradually change the smoothness through homotopy transformation to reduce uncertainty at the end of the prediction horizon and increase flexibility, thereby improving overall performance. 3. **Algorithm Performance Improvement**: - In simulation experiments, different formulation methods are compared, and the new method is shown to improve performance metrics without increasing computation time. The core contribution of the paper is the proposal of a new method for smoothing obstacle shapes (referred to as Progressive Smoothing) and demonstrating its effectiveness in obstacle avoidance through numerical analysis. Additionally, two alternative smoothing methods (LogSumExp and Boltzmann) are introduced and their performance is evaluated in closed-loop simulations.

Progressive Smoothing for Motion Planning in Real-Time NMPC

Autonomous Vehicle Motion Control Using NMPC Based on Orthogonal Collocation Method: Integration Stability Study

Accurate Pseudospectral Optimization of Nonlinear Model Predictive Control for High-Performance Motion Planning

Nonlinear Model Predictive Path Tracking Control for Autonomous Vehicles Based on Orthogonal Collocation Method

Balancing Accuracy and Efficiency: Fast Motion Planning Based on Nonlinear Model Predictive Control

Real-time Motion Planning and Trajectory Tracking in Complex Environments Based on Bézier Curves and Nonlinear MPC Controller

ASAP-MPC: An Asynchronous Update Scheme for Online Motion Planning with Nonlinear Model Predictive Control

Predictive Motion Planning for Autonomous Vehicles with Geometric Constraints Via Convex Optimization

Model Predictive Optimization for Imitation Learning from Demonstrations.

A Non-Linear Model Predictive Controller with Obstacle Avoidance for a Space Robot

R$^2$NMPC: A Real-Time Reduced Robustified Nonlinear Model Predictive Control with Ellipsoidal Uncertainty Sets for Autonomous Vehicle Motion Control

Nonlinear MPC of Quadrotors for Path Planning and Dynamic Obstacle Avoidance

Real-time Motion Planning Based on MPC with Obstacle Constraint Convexification for Autonomous Ground Vehicles

Progressive Learning for Physics-informed Neural Motion Planning

Progressive Barrier Tightened Fast Nonlinear Model Predictive Control for High-Performance Path Tracking

Efficient Nonlinear Model Predictive Path Tracking Control for Autonomous Vehicle: Investigating the Effects of Vehicle Dynamics Stiffness

Efficient and integration stable nonlinear model predictive controller for autonomous vehicles based on the stabilized explicit integration method

Mixed-Integer MPC-Based Motion Planning Using Hybrid Zonotopes with Tight Relaxations

Custom Non-Linear Model Predictive Control for Obstacle Avoidance in Indoor and Outdoor Environments

Model Predictive Control with Environment Adaptation for Legged Locomotion

Aerial navigation in obstructed environments with embedded nonlinear model predictive control