Energy Management of a Semi-Autonomous Truck Using Blended Multiple Model Controller Based on Particle Swarm Optimization

Recently, electrification and automation of heavy-duty refuse trucks has gained attention both in the industry and academia due to new legislations announced by the European Union. During a typical drive cycle of a refuse truck, a significant variation in mass is observed as the vehicle covers its garbage collection points. However, proposed control strategies that exists in the literature are mainly based on a single controller with fixed gains structure, which limits optimum use of the on-board energy and precludes the vehicle to follow the desired drive cycle owing to the mass variation. In the framework of European Union-funded OBELICS and ESCALATE projects, a novel control strategy for a semi-autonomous refuse truck is developed to enhance its performance by optimizing the gas and brake pedal gains. This work stands out from classical approaches presenting a Particle Swarm Optimization (PSO) algorithm which calculates optimal controller gains and a Multiple Model Controller (MMC) which adapts the controller gains based on the vehicle mass. The main aim of this paper is to: i) optimize the controller parameters, ii) reduce the vehicle's total energy consumption and iii) reduce the speed tracking error. To this end, a cost function satisfying these objectives is formulated for both autonomous and manual driving modes. The vehicle performance is tested for the Eskisehir city drive cycle. Simulation results demonstrate that the proposed MMC strategy enhances the performance of the FORD refuse truck by 5.19% in autonomous driving mode and 0.534% in manual driving mode compared to traditional approaches .