In-vehicle and out-of-vehicle voice interaction method, system, device, and storage medium

Abstract

The application discloses an in-vehicle and out-vehicle voice interaction method, system, device and storage medium, relates to the technical field of intelligent cockpit and man-machine interaction, and comprises the following steps: collecting sensing data; identifying an interaction trigger type (in-vehicle active request, out-vehicle passive request or system active trigger) based on the sensing data and / or a vehicle event; determining a corresponding response permission policy according to the trigger type, and determining whether the system is allowed to respond based on the response permission policy; if allowed, calling a large language model to generate voice text, combining external sensing features to synthesize voice, and playing through an out-vehicle loudspeaker or an in-vehicle sound system. The scheme identifies three types of interaction intents, namely in-vehicle active request, out-vehicle passive request and system active trigger, calls a large language model to generate scene-adapted voice text, and combines external sensing features to perform personalized voice synthesis and play, thereby improving the response efficiency, naturalness of expression and object adaptability of in-vehicle and out-vehicle voice interaction.

Images