Multi-Sensor Fusion-Based Intelligent Vehicle Automatic Tracking and Target Aiming System

Abstract

To address the limitations of conventional mobile robots in terms of insufficient tracking accuracy and dynamic aiming drift, this paper proposes a multi-sensor fusion-based intelligent vehicle system for automatic path tracking and target aiming. The system employs a dual-microcontroller architecture comprising the MSPM0G3507 and STM32F103RCT6, integrating grayscale sensors, Hall-effect encoders, a six-axis gyroscope, and the K230 AI vision module for perception and data acquisition. A grayscale-weighted deviation mathematical model is established, the discrete recursive formulas for cascade PID control are derived, and the Kalman filter state equations are formulated to achieve precise dual closed-loop modeling of the pan-tilt unit. Experimental results demonstrate that the vehicle completes stable multi-lap tracking with a per-lap timing error not exceeding 0.5 s, the pan-tilt unit achieves static target lock within 4 s, and multisensor fusion reduces the aiming root-mean-square error (RMSE) by approximately 63%, validating the feasibility of the proposed multi-sensor fusion scheme for mobile aiming robots.