This paper presents a personalized human-machine shared driving (HMSD) system aimed at aiding drivers in obstacle avoidance path planning and tracking based on driving styles. Initially, a driver-in-the-loop experimental bench is utilized to collect the driver characteristic data. Then, the collected data undergoes normalization and clustering to quantify driving styles. Furthermore, a personalized path planning approach is developed to enhance driving safety, vehicle stability, and traffic efficiency. Finally, the proposed system is validated using the driver-in-the-loop experimental setup and a questionnaire survey. Results show the system's capability to provide personalized assistance, reduce driving load, and make driver-vehicle interaction more reliable and smoother, garnering high satisfaction. Notably, the study reveals significant individual differences in the perception of the system's effectiveness and trustworthiness. Aggressive drivers, confident in their driving abilities, display the lowest level of trust in the system. Conversely, cautious drivers perceive the system as reliable assistance, leading to the highest level of trust. Moderates maintain a modest level of trust, seeking a balance. These findings provide valuable insights for developing intelligent vehicle human-machine interaction systems, suggesting that customization to individual driving styles can significantly improve system acceptance and effectiveness.