LSTM Motion Classification Using Insole Pressure Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current VR motion recognition technologies face challenges in mapping real-world motion to virtual environments naturally and intuitively, especially in complex and changing real scenes, due to the need for external devices and constraints on user movement.
Innovation Solution
A motion behavior pattern classification method using a motion control sensor, specifically pressure sensors in insoles, to capture and classify motion data through an LSTM neural network, allowing for intuitive and natural interaction by predicting motion behavior patterns without artificial feature extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion capture systems use external devices (specialized cameras and tracking suits), then motion mapping accuracy is improved, but device complexity and production overhead increase
Solution Approach 1:
The patent extracts the motion sensing function from complex external devices and concentrates it into a single wearable sensor unit. This sensor unit captures motion data directly from the user's body, eliminating the need for multiple specialized cameras and tracking suits while maintaining measurement precision through direct physical sensing.
Solution Approach 2:
The patent introduces an intermediary processing system that bridges the simple wearable sensor and the virtual environment. This intermediary system uses machine learning models to interpret raw sensor data and generate natural motion mappings, resolving the contradiction by adding intelligence rather than physical complexity.
2Loss of time
If motion recognition uses traditional sensor-based methods, then setup speed is improved, but adaptability to complex real scenes deteriorates
Solution Approach 1:
The patent implements dynamic adaptability through machine learning models that continuously learn from and adapt to complex real-world scenes. The system evolves its motion recognition capabilities based on environmental feedback, allowing fast setup without sacrificing adaptability to changing scenarios.
Solution Approach 2:
The patent changes the parameters of motion recognition by using deep learning models with adjustable parameters that can be tuned for different environments. This allows the same sensor-based system to adapt to various complex scenes by modifying computational parameters rather than physical configuration.
3Loss of information
If motion classification uses manual feature extraction, then interpretability is improved, but classification accuracy and speed deteriorate
Solution Approach 1:
The patent replaces manual feature extraction mechanisms with automated deep learning-based feature learning. Neural networks automatically discover relevant features from raw sensor data, eliminating the need for hand-crafted features while improving both classification accuracy and processing speed through efficient computational architectures.
Data Source
AI summary
A motion behavior pattern classification method, system and device relating to the human motion recognition field. The method includes: S1, determining a candidate motion behavior pattern which includes a motion behavior pattern to be classified; S2, acquiring time series of behavior data of the candidate motion behavior pattern through a motion control sensor; S3, establishing an LSTM motion behavior pattern classification model through the time series; S4, predicting the motion behavior pattern to be classified through the LSTM motion pattern classification model, comparing the prediction result obtained using the sequence within the T time with the prediction result obtained using the sequence within the T+Δt time using the iterative process to obtain the final prediction result, wherein T1≤T≤T2, T1>0, T2>T1, and the incremental step is set to be Δt, Δt>0. The technical solution of the present invention may improve accuracy while reducing the latency of the motion behavior pattern classification.