Research on Dog Bark Recognition Algorithm Based on Mel-Frequency Cepstral Coefficients and Lightweight Neural Network

Lin Haipeng — Mon, 24 Aug 2026 00:00:00 +0800

Dog barks serve as key bioacoustic signals for pet monitoring and urban noise early warning. Conventional recognition methods suffer severe feature distortion in noisy environments, while deep learning models carry excessive parameters and slow edge inference. This work proposes an algorithm combining spectral entropy-based adaptive denoising MFCC and ECA lightweight attention network. It fuses static and temporal differential cepstral features and uses audio augmentation to mitigate small-sample overfitting. Tests on ESC50, UrbanSound8K and self-built dog bark datasets show 4.3% higher accuracy than standard MFCC at 0 dB SNR. The model has only 2.07 M parameters and 0.31 G computation load, cutting inference latency by 62.8% versus CNN-LSTM, with an average F1 of 0.912 (5–20 dB SNR). Edge tests validate its real-time performance for embedded bioacoustic detection.

Modular Architecture Design of Multi-Scenario Simulation Test Platform for Vehicular Wireless Signals

Zhao Yanbo — Tue, 25 Aug 2026 00:00:00 +0800

Vehicular wireless communication underpins vehicle interaction, vehicle-road coordination and autonomous driving, whose signal transmission performance directly affects driving safety. To solve the drawbacks of traditional vehicle wireless test platforms including tight coupling, low scenario reusability, insufficient simulation accuracy and poor multi-condition compatibility, this paper proposes a layered decoupled, dynamically reconfigurable modular multi-scenario vehicle wireless simulation test architecture. First, we analyze wireless transmission characteristics of typical driving scenarios (highway, urban, tunnel, extreme weather) and build a multi-parameter coupled channel model covering path loss, multipath fading and Doppler time-varying features. Second, six core modules (scenario scheduling, channel simulation, interference emulation, data acquisition, closed-loop control, etc.) are decoupled, with standardized interfaces developed to realize pluggable reconfiguration of scenarios, parameters and hardware. Finally, a hardware-in-the-loop platform is built for multi-scenario comparative tests. Experimental results show the architecture’s average simulation error ≤3.2% and maximum error ≤4.3%. Compared with traditional platforms, scenario adaptation efficiency rises by 46.2%, hardware reuse rate increases from 38.5% to 100%, and scenario switching completes within 10 ms. It supports mainstream vehicular wireless standards (5G, V2X, UWB). This design effectively fixes the defects of conventional test systems and provides reliable engineering support for high-precision testing and algorithm iteration of 5G/6G IoV.

Journal of Academic Research and Advances

Research on Dog Bark Recognition Algorithm Based on Mel-Frequency Cepstral Coefficients and Lightweight Neural Network

Modular Architecture Design of Multi-Scenario Simulation Test Platform for Vehicular Wireless Signals