Ensemble hill climbing optimization in adaptive cruise control for safe automated vehicle transportation

Road traffic crashes are currently one of the leading causes of death. Many accidents occur when the attention of a driver is diverted owing to fatigue or external factors. Adaptive cruise control (ACC) is a driver assistance system for safe vehicle transportation by combining a safe following distance with speed regulation. However, existing ACC systems fail to control the braking system when the preceding vehicle applies full brakes that can cause a jerk or collision. In addition, current ACC systems fail to consider road elevation and spatiotemporal problems that cause tracking errors. The existing issues result in high fuel consumption, discomfort during travel, and high error rates. To overcome these issues, a “Combinatorial Repeated Local Search and Simulated Annealing-Based Hill Climbing Optimization” (CRLSSA-HCO) technique is introduced for ACC with a predictive controller. In this technique, two models are introduced, namely the repeated local search-based hill climbing optimization (RLS-HCO) model, and the simulated annealing-based hill climbing optimization (SA-HCO) model. The RLS-HCO model performs local search optimization to avoid vehicle jerks and collisions by maintaining a safe distance between vehicles on a horizontal road. The RLS-HCO model operates in a spacing control mode and a speed control mode to maintain the safe distance. The input and control variables are optimized by performing a repeated local search, thereby reducing the fuel consumption and error rate. The SA-HCO model performs a global search optimization for avoiding traffic and addresses the spatiotemporal problems on hilly roads. The SA-HCO model optimizes the engine speed, engine torque, and gear ratio by using simulated annealing for reducing the fuel consumption and error rate. An experimental evaluation of the CRLSSA-HCO technique is carried out using performance metrics such as fuel consumption, distance error, and speed error, which are compared to those of state-of-the-art studies.