Predictive adaptive cruise control system based on a vehicle-cloud dual closed-loop control architecture

Abstract

Conventional Adaptive Cruise Control (ACC) systems often neglect long-distance road slope information in energy-saving strategies, making it challenging to achieve coordinated optimization of safety and fuel efficiency under complex road conditions. To address this limitation, a Predictive Adaptive Cruise Control (PACC) algorithm is proposed based on a vehicle-cloud dual closed-loop architecture, enabled by the extended perception and computational capabilities of a Cloud Control System (CCS). The algorithm adopts a hierarchical dual-loop control architecture. On the cloud side, it utilizes waypoint-triggered dynamic programming combined with long-distance slope data to generate economic speed profiles. On the vehicle side, Gaussian Process Regression predicts preceding vehicle speed and integrates it with the planned speed sequence for real-time following speed optimization. The cloud establishes a minute-level planning macro-loop, while the vehicle implements a second-level speed optimization micro-loop and millisecond-level real-time control, enhancing system stability and ensuring safe and energy-efficient operation of commercial vehicles. Simulation results demonstrate that, under various cruising speed conditions, the proposed PACC algorithm significantly improves following performance, reduces braking actions, and achieves 3% to 10% fuel savings compared to conventional ACC systems, while ensuring driving safety.