Embedded Bark Recognition for Accurate Stop-Barking Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing stop-barking devices for pets inaccurately identify barking sounds, leading to wrongful punishments and potential safety risks for pets due to their reliance on sound decibel levels alone.
Innovation Solution
An intelligent stop-barking method and device using a pet-worn embedded terminal with a built-in barking model and punishment module, employing sound collection, model comparison, and punishment triggering based on decibel values and machine learning models to accurately identify dog barking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional sound recognition systems are used that only detect decibel levels, then the device complexity is low, but the measurement precision of barking identification deteriorates leading to wrong determinations
Solution Approach 1:
The patent pre-trains a barking recognition model on a powerful computing platform (PC or cloud) before deploying it to the embedded terminal. This preliminary action transfers the complex computational workload to a more capable system, allowing the resource-constrained embedded device to perform accurate barking identification without requiring high computational power locally.
Solution Approach 2:
The patent introduces an intermediary processing layer that converts the trained barking recognition model into a format suitable for embedded deployment. This intermediary step involves model optimization and conversion techniques that bridge the gap between powerful training environments and resource-constrained execution environments, enabling accurate identification without direct connection to PC or cloud.
2Ease of operation
If the embedded terminal operates independently without PC or cloud connection, then the ease of operation improves, but the manufacturing precision of the barking model deteriorates
Solution Approach 1:
The barking recognition model is pre-trained and optimized on a powerful computing platform before being deployed to the embedded terminal. This preliminary action ensures high model accuracy is achieved during training, and the optimized model can then operate independently on the resource-constrained embedded device without requiring continuous PC or cloud connection.
Solution Approach 2:
The patent extracts the essential barking recognition functionality from the complex PC/cloud-based training system and consolidates it into a compact, optimized model that can run independently on the embedded terminal. This extraction process separates the heavy computational training phase from the lightweight inference phase, enabling independent operation with maintained accuracy.
3Device complexity
If resource-constrained embedded systems are used, then the device complexity is reduced, but the measurement precision of sound analysis deteriorates
Solution Approach 1:
The patent replaces complex mechanical computing resources with an optimized software model approach. Instead of relying on raw computational power for real-time sound analysis, the system uses a pre-trained neural network model that has been optimized for embedded deployment, substituting computational complexity with algorithmic efficiency.
Solution Approach 2:
The patent changes the parameters of the barking recognition model to suit embedded system constraints. This includes optimizing model architecture, reducing precision requirements where appropriate, and tuning hyperparameters to achieve the best balance between accuracy and computational resource usage on the embedded platform.
Data Source
AI summary
The present invention discloses an intelligent stop-barking method, device, and computer-readable storage medium. The method runs on an embedded terminal worn by a pet. The embedded terminal is built in with a preset pet barking model. The method includes: sound collection: obtaining a decibel value of ambient sound of the pet's surrounding environment; model comparison: if the decibel value of the sound is greater than a preset decibel value, then the sound collected is compared with the pet barking model and a comparison result is output; and output punishment: if the comparison result is greater than a preset punishment value, then the punishment module is triggered to stop the pet from barking. The technical solution, through machine learning models and signal processing techniques, enables high-accuracy dog barking sound recognition. It reduces reliance on powerful computing resources, and achieves stand-alone operation of data model, thereby improving feasibility and popularization.