Input-to-State Safety with Input Delay in Longitudinal Vehicle Control

TL;DR

This paper develops a control method for connected automated trucks that ensures safety despite input delays and model uncertainties, using control barrier functions and predictor feedback.

Contribution

It introduces a novel control approach that explicitly accounts for input delay and disturbances in longitudinal vehicle control, ensuring provably safe and collision-free operation.

Findings

The proposed controller outperforms delay-ignoring designs in simulations.

It guarantees safety despite significant input delays.

The method effectively handles model uncertainties.

Abstract

Safe longitudinal control is discussed for a connected automated truck traveling behind a preceding connected vehicle. A controller is proposed based on control barrier function theory and predictor feedback for provably safe, collision-free behavior by taking into account the significant response time of the truck as input delay and the uncertainty of its dynamical model as input disturbance. The benefits of the proposed controller compared to control designs that neglect the delay or treat the delay as disturbance are shown by numerical simulations.