Differentially Private Aperiodic Sampled-Data Consensus for Intelligent Interconnected Heterogeneous Vehicular Platoons

Abstract

In this article, the vehicular platoons under aperiodic sampled-data information exchange between connected neighbors are implemented with an average output consensus while achieving differential privacy protection. Initially, the complex vehicle dynamics system is simplified into a heterogeneous linear system interconnected through a communication graph. Subsequently, a distributed hybrid controller is deployed, specifically tailored for handling the intermittent sampled-data information, and is augmented with a dynamic noise generator. This framework restricts information exchange to a predefined neighborhood set of vehicles. To ensure differential privacy, design and incorporate random noise adhering to a Laplace distribution, where the decay index and control gain are adjustable parameters corresponding to the desired privacy level and system accuracy, respectively. This noise injection is guided by a differential privacy noise utilization algorithm. The controller design is then formulated and solved using a combination of the pole placement method and the generalized inverse concept, enabling effective networked control. Lastly, simulation examples are provided to rigorously validate the proposed theoretical framework, demonstrating its efficacy in maintaining platoon coherence while preserving the privacy of individual vehicle data.