AutoNavi has announced the launch of the world's first map-based AI navigation agent (NaviAgent), marking a significant shift in navigation services from traditional travel tools to intelligent travel companions capable of thinking, predicting, and offering personalized experiences. This agent employs a full-chain intelligent model of "thinking - predicting - acting," building an intelligent closed loop through four modules: perception, planning, execution, and expression. It integrates multiple core technologies, including the TrafficVLM traffic perception model and AmapVoice emotional speech technology, achieving a transition from a "tool-oriented" mindset to a "partner-oriented" one.
Traditional navigation services typically follow pre-set routes mechanically. In contrast, the AutoNavi AI navigation agent acts like an experienced driver, capable of real-time road condition awareness, risk prediction, and proactive strategy adjustments. For example, it can predict upcoming road conditions, precisely calculate optimal speeds for passing through intersections to avoid congestion, sense user fatigue from late-night work commutes, proactively play soothing music, recommend the nearest parking lot, and communicate with users through natural conversational voice.
Technically, leveraging the TrafficVLM traffic perception and prediction capabilities, the AutoNavi AI navigation agent overcomes the limitations of traditional navigation's "local perception," upgrading from "single-view observation" to "global thinking." This technology uses a visual language model as its core, converting traffic information of different spatiotemporal scales into image frames. The model then infers the status and predicts trends, giving the navigation system a beyond-visual-range perception capability. In high-speed driving scenarios, the AI navigation agent can perceive lane-level traffic flow changes, dynamic events, and lane obstructions, automatically activating the navigation assistance function based on the user's route, recommending the globally optimal lane, and making lane changes smoother and more efficient. Furthermore, it can detect the dynamics of surrounding vehicles in real-time, issue lane-level safety warnings, predict driving risks in advance, and provide real-time warnings through voice broadcasts and navigation displays.
The AI navigation agent also constructs a three-dimensional interaction system encompassing emotion recognition, intent understanding, and emotional expression, extending services to the emotional value domain. "Xiao Gao Laoshi" (Little Gao Teacher), the embodied virtual interaction carrier, powered by AmapVoice technology and trained on hundreds of thousands of hours of voice data, has achieved a qualitative change from mechanical speech to emotional interaction, mastering various emotions and paralinguistic expressions. For example, when a user finishes a long work week and embarks on a weekend trip, the AI navigation agent can understand the travel intentions, route, and mood, proactively adjusting the interaction content to provide a caring and companionable journey.
Currently, the AutoNavi AI navigation agent is being rolled out to users nationwide. Its launch marks a shift in travel products from efficiency-first to an intelligent experience that prioritizes proactive service. With the continuous iteration of AI capabilities, future travel will not only ensure users "arrive safely" but also aim for a "pleasant journey."