Customer Case - Pachira information technology (Beijing) Co., Ltd.

AI+Car

Relying on Pachira's self-developed AI capability platform to provide in-vehicle voice services for car front and rear installations

Location：HomeSolutionsSmart Cockpit SolutionAI+CarCustomer Case

Industry Pain Points

行业痛点

Car machine itself function operation keys

Poor human-computer interaction experience

Lack of exclusive companionship

Unable to complete personalized OTA upgrade

Solution Products

行业痛点

Car assistant
Multimode interaction
End-to-cloud integration

Solution Description

In order to achieve more friendly human-computer interaction between vehicle and machine, Pachira uses its own AI technology to create a new generation of vehicle assistant. Provide a unified portal for natural language interaction. Realize intelligent service ecology such as intelligent knowledge service, intelligent data query and intelligent iot operation to improve customer service perception in all aspects. Conduct big data analysis in the background, assist knowledge operation and personalized OTA upgrade.

Application scenarios

All types of front and rear loading models

Solution Description

Through deep integration of voice, buttons, touch, text, image (gesture, face, expression, etc.), vehicle and machine state, the output of voice, GUI, image, video, vehicle control unit and so on can realize any time, any function, any application can interact through voice

Application scenarios

In-car personalized Settings, active transaction early warning, accurate identification of customer status, etc

Solution Description

The advantages of cloud accuracy and fast terminal response are combined to determine state information intelligently by combining confidence and response delay. Context synchronization strategy; Full coverage of abnormal conditions.

Application scenarios

All kinds of car machine

Advantage

行业痛点

Multiple rounds of context interaction

Task switching is performed based on the context of the task relationship. Error correction and interaction in multiple rounds of dialogue

Full duplex interaction

A single wake-up call allows for continuous interaction across domains. Supports real-time conversation interruption and dynamic VAD. The machine refuses to recognize relevant speech from non-users

Sound source localization

Through sound source positioning technology, provide two-tone area/four-tone area scheme, support the main driver/co-driver mode.

One-shot

The integration of 'wakeword + speech semantic recognition' can realize zero interval, zero delay and seamless connection between wakeword and voice manipulation, abandon the traditional form of ask and answer, and reduce the steps of user voice manipulation