Combined ANFIS method with FA, PSO, and ICA as Steering Control Optimization on Electric Car

Abstract

Adaptive Neuro-Fuzzy Inference System (ANFIS) is an Artificial Intelligence (AI) called Artificial Neural Network (ANN) based on Takagi-Sugeno Fuzzy Inference System (FIS). ANFIS integrates neural networks and Fuzzy Logic principles, has the ability to take advantage of both within a single framework. ANFIS can be used to control and optimize the system automatically. Optimization of steer movement is necessary for steer movement with the vehicle. Faulty movement leads the car will generate errors position of the car on the vehicle line. Several studies have been developed in fully automated steer by cabling systems, including those devoted to input paths, using GPS technology and trajectories. The steering system in this study used PID and ANFIS controllers, which were tuned with some Artificial Intelligence (AI). AI is expected to assist in accelerating and optimizing the control process. In this research, we will develop a completely “Automatic Steer by Wire System” model using 10 Degree Of Freedom (DOF) consisting of 3-DOF Vehicle Handling Model and 7-DOF Vehicle Ride Model. The method already used is the PID that is tuned using AI. This research combines ANFIS method with Firefly Algorithm (FA), Particle Swarm Optimization (PSO) and Imperialist Competitive Algorithm (ICA). Then compare the result with the FA-PID, PSO-PID, ICA-PID, FA-PID-ANFIS, PSO-PID-ANFIS, and ICA-PID-ANFIS. Furthermore, this method is used to design a real electric car. At a standard speed of 13.8 km/h, the ICA-PID-ANFIS method has the smallest error of 0.005070 m. If the speed of the vehicle is changed is able to maintain stay on track only up to speed 62.1 km/h. But the ICA-PID method can keep it on track to speeds of 69.0 km/h. The results of this simulation is a simulation of the ideal conditions of the vehicle. This research is used to determine the best method that can adjust the actual trajectory. So it can be used as reference for real vehicles.