Resilient Cruise Control of Heterogeneous Platoons Against Byzantine Attacks: Theory and Experiment

This article studies the problem of the resilient cruise control in heterogeneous vehicle platoons against
$f$
-local Byzantine attacks (BAs). Agents under BAs become traitors of the swarm, who try to mislead its neighbors while adopting wrong inputs. Thus, BAs are extremely challenging to be suppressed. This study introduces a novel hierarchical protocol characterized by a virtual twin layer (TL), motivated by the rationale of digital twin. This protocol separates the defense scheme against
$f$
-local BAs into two parts: one defense scheme against Byzantine edge attacks (BEAs) via the TL and another scheme against Byzantine node attacks (BNAs) via the cyber–physical layer (CPL). The TL employs a trusted-edge strategy, enhancing the network resilience by incorporating a minimal fraction of the key edges. It is rigorously proven that a TL topology meeting strong
$(2f+1)$
-robustness is sufficient for achieving distributed resilient estimation against BEAs. On the CPL, a series of decentralized chattering-free controllers is proposed, guaranteeing the resilient cruise tracking of heterogeneous platoons against exponentially unbounded BNAs. Besides, these controllers can achieve uniformly ultimately bounded convergence. The theoretical results’ effectiveness and practicality are validated through a numerical simulation example and an unmanned ground vehicle experiment involving heterogeneous platoons against
$f$
-local BAs.