Moving Obstacle Avoidance: a Data-Driven Risk-Aware Approach

TL;DR

This paper introduces a risk-aware model predictive control framework for moving obstacle avoidance, predicting obstacle trajectories from noisy data and providing probabilistic safety guarantees for autonomous agents.

Contribution

It presents a novel structured approach combining bootstrapping and risk-aware MPC to improve moving obstacle prediction and avoidance under uncertainty.

Findings

Successful simulation of drone avoiding moving obstacles

Probabilistic guarantees on obstacle avoidance achieved

Effective handling of noisy trajectory measurements

Abstract

This paper proposes a new structured method for a moving agent to predict the paths of dynamically moving obstacles and avoid them using a risk-aware model predictive control (MPC) scheme. Given noisy measurements of the a priori unknown obstacle trajectory, a bootstrapping technique predicts a set of obstacle trajectories. The bootstrapped predictions are incorporated in the MPC optimization using a risk-aware methodology so as to provide probabilistic guarantees on obstacle avoidance. We validate our methods using simulations of a 3-dimensional multi-rotor drone that avoids various moving obstacles, such as a thrown ball and a frisbee with air drag.