Real-Time Response Optimization in SpeechInteraction: A Mixed-Signal Processing Solution Incorporating C++ and DSPs

Aiming at the high real-time requirements of mobile devices, smart home and medical terminals, this paper proposes a mixed-signal processing scheme with C++ and DSP synergy, and designs an end-to-end framework to reduce the delay from feature extraction to instruction parsing. Experimental results show that the response time under typical instructions is reduced by about 34% compared with existing schemes. The scheme follows the secure coding specification and adopts a modular sandboxing mechanism to prevent signal injection attacks, providing a reference paradigm for high real-time and high-security scenarios such as in-vehicle voice and remote control.