Hybrid Speech Data Processing in Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network architectures for vehicles face challenges in providing reliable and efficient speech recognition processing due to mobility and network availability issues, leading to potential delays in processing.
Innovation Solution
A hybrid speech data processing method and system that digitizes speech inputs into packets, stores them temporarily, and transmits them via wireless voice and data communication channels to a speech recognition server, converting packets to text data when necessary to ensure effective processing and outputting recognition results through an interactive voice response system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If speech packets are transmitted using wireless voice communication channel to speech recognition server, then speech recognition processing can be performed, but network delays and processing delays may occur due to vehicle mobility and network availability issues
Solution Approach 1:
The system performs preliminary actions by storing speech packets locally in memory before transmission is complete or if transmission fails. This buffer storage mechanism ensures that speech data is preserved and can be processed later, preventing loss of speech input due to network delays or unavailability. The hybrid approach proactively handles potential transmission failures by having a fallback local processing path ready.
Solution Approach 2:
The patent introduces an intermediary hybrid processing architecture that includes both cloud-based speech recognition server and local interactive voice response system. This intermediary structure allows speech packets to be routed through multiple paths - first attempting cloud processing, then falling back to local processing if cloud access is delayed or unavailable. This mediator approach resolves the contradiction by providing alternative processing routes that maintain reliability while minimizing delays.
2Reliability
If hybrid speech data processing with dual communication channels is implemented, then processing reliability is improved, but system complexity increases
Solution Approach 1:
The system dynamically adapts its communication strategy based on real-time network conditions. It monitors the status of wireless voice and data communication channels, and automatically switches between cloud-based and local processing modes. This dynamic behavior allows the system to maintain simplicity in normal operation while providing complex fallback capabilities only when needed, thus improving reliability without permanently increasing operational complexity.
Solution Approach 2:
The patent applies local quality by implementing different processing capabilities at different locations in the system architecture. The vehicle contains local speech packet storage and local interactive voice response processing capability, while also having access to remote speech recognition server. This localized quality ensures that critical speech processing functions can operate independently if needed, reducing the complexity burden on the overall system while maintaining high reliability.
3Loss of time
If speech packets are stored in memory for predetermined time, then processing delays are prevented, but memory resources are consumed
Solution Approach 1:
The system implements partial action by storing only the necessary speech packets in memory for the predetermined time period required for potential fallback processing. It does not indefinitely store all speech data, but rather maintains a controlled buffer of recent speech packets that may need processing if cloud communication fails. This partial storage approach prevents excessive memory consumption while still providing sufficient buffer to handle typical network delays.
Data Source
AI summary
A computer-implemented method for hybrid speech data processing in a vehicle includes receiving a first speech input at an input device in the vehicle and digitizing the first speech input into packets. The method includes storing the packets at a memory for predetermined amount of time and transmitting the packets using a wireless voice communication channel to a speech recognition server. Upon detecting a speech recognition condition is met associated with the packets or the wireless voice communication channel, the method includes converting one or more of the packets stored at the memory into text data and transmitting the text data using a wireless data communication channel to an interactive voice response system. Further, the method includes receiving a speech recognition result from the interactive voice response system and outputting the speech recognition result in the vehicle using an output device.