Abstract: With the rapid development of new technologies such as autonomous driving and intelligent cockpit in automobile, voice interaction has emerged as one of the most widely adopted modes of interaction in vehicle cabins. It not only makes human-vehicle interaction more natural and convenient but also reduces accident rates and enhances driving safety. Given the variety of voice interaction products available in the market, it is crucial to have a feasible and reliable test-evaluation method to assess whether these products, once integrated into vehicle infotainment systems, meet the practical demands of in-driving interaction. Based on user research, this paper designs test scenarios and evaluation metrics from both subjective and objective perspectives, considering both driving safety and user experience. Six vehicle models are evaluated for their voice interaction systems. The research findings indicate that the proposed method is executable and capable of evaluating the merits and demerits of in-vehicle voice interaction systems, providing valuable insights for system functional development and design, benchmark tests, as well as optimization and iteration processes.

Key words: voice interaction system; real vehicle test; subjective evaluation; objective evaluation

摘要： 随着汽车自动驾驶和智能座舱等新技术的快速发展，语音交互成为汽车座舱最广泛应 用的交互方式之一，不仅让人车交互变得更加自然和便利，还能够降低事故率，提高驾驶安 全系数。市面上形形色色的语音交互产品，整合到车机后是否满足实际行车过程交互的使用 要求，需要一套可行可靠的测试评价方法。文章基于用户调研情况，考虑驾驶安全性和用户 体验性，从主、客观两个方面设计了测试场景和评价指标，对 6 款车型的语音交互系统进行 测评。研究结果表明，文中所设计的方法具有可执行性并且可以评价车载语音交互系统的优 劣，对于系统功能开发设计、对标测试以及优化迭代有重要帮助。

关键词: 语音交互系统；实车测试；主观评价；客观评价