Implementation of Predictive Control Model and Obstacle Avoidance Fusion Sensor in Autonomous Vehicles

This research aims to develop an autonomous vehicle system based on swerve drive that is able to perform obstacle avoidance maneuvers in a precise, adaptive, and real-time manner. The system is designed by integrating three main components, a swerve drive mechanism to produce omni-directional movement, a fusion sensor (lidar, camera, and CMPS12 orientation sensor) to improve the accuracy of environmental perception, and a Predictive Control Model (MPC) as the main controller to set a stable and smooth motion trajectory. The prototype of the vehicle was designed using the Arduino Mega 2560 as the low-level controller and the ROS as the data processing center. The test was carried out in an indoor environment with basic movement scenarios, steering angle stability, object detection, and obstacle avoidance at various distances. The results showed that the swerve drive system was able to produce forward, lateral, rotation, and diagonal movements stably with an average steering angle error of 2–4 degrees. The fusion sensor shows consistent object detection performance, especially at a distance of ?80 cm. The initial implementation of the MPC successfully resulted in a smooth obstacle avoidance maneuver, although actuator response and parameter tuning still need to be optimized for near-range (<50 cm) obstacle scenarios. Thus, this study proves that the combination of swerve drive, sensor fusion and MPC is an effective approach to improve the navigation and safety capabilities of autonomous vehicles. The developed system can be the basis for further development towards autonomous vehicles that are more intelligent, responsive, and adaptive.