Gait Training Load Sensor Using CNN for Precise Pressure Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing gait training systems face challenges in accurately detecting load distribution, leading to inaccurate timing in walking cycles and suboptimal driving control, as viscoelastic bodies deform not only at the applied load site but also in surrounding regions, making precise load detection difficult.
Innovation Solution
A gait training system incorporating a load distribution sensor with a two-dimensional array of sensors and an inference device using convolutional neural networks (CNNs) for precise load estimation, which includes a viscoelastic sheet for deformation-based load detection and a machine learning model for accurate pressure calculation, integrated with a robot leg and treadmill for enhanced control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a viscoelastic body is used to detect load distribution, then the sensor can detect pressure through deformation, but the viscoelastic body deforms also at surrounding regions making it difficult to detect load with high accuracy
Solution Approach 1:
The patent divides the detection area into multiple discrete sensor elements arranged in a two-dimensional array. Each sensor element independently measures deformation at its specific location, allowing the system to distinguish between deformation caused by actual load application versus surrounding region deformation. This segmentation enables precise identification of the load application site by analyzing the spatial distribution of deformation across the array.
Solution Approach 2:
The patent transitions from a single-point or linear deformation measurement to a two-dimensional array of sensors. By adding the spatial dimension and analyzing the distribution pattern of deformation across multiple points, the system can accurately identify the load application site even when surrounding regions deform. The two-dimensional data structure allows for sophisticated analysis of deformation patterns to distinguish true load locations from surrounding deformation effects.
2Reliability
If load distribution is detected with a sheet-shaped viscoelastic body, then the system can infer walking cycle timing, but the deformation spreads to surrounding regions causing inaccurate timing inference
Solution Approach 1:
The two-dimensional array of sensor elements provides discrete measurement points that can independently track deformation at each location. This segmentation allows the system to precisely identify which specific sensors detect deformation, thereby accurately determining the timing of load application and corresponding walking cycle events without interference from surrounding region deformation.
Solution Approach 2:
The system uses the deformation data from the two-dimensional sensor array to provide feedback on load distribution patterns. By continuously monitoring which sensors detect deformation and analyzing the spatial-temporal patterns, the system can accurately infer walking cycle timing and adjust its interpretation based on the observed deformation distribution, improving the reliability of timing inference.
3Ease of operation
If driving control is performed based on load distribution from a viscoelastic sheet, then the system can control robot leg timing, but the surrounding deformation reduces control accuracy
Solution Approach 1:
The two-dimensional sensor array provides segmented, location-specific deformation data that enables precise identification of load application sites. This granular information allows the driving control system to accurately determine when and where load is applied, leading to more appropriate timing control of the robot leg without being adversely affected by surrounding region deformation.
Solution Approach 2:
By utilizing two-dimensional spatial distribution data from the sensor array, the system gains additional dimensional information about load patterns. This spatial context enables the control algorithm to distinguish between actual load application and surrounding deformation, improving the accuracy of driving control decisions and timing synchronization.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system achieves high accuracy in load detection and walking phase determination, enabling more appropriate driving control and effective gait training by distinguishing between actual and surrounding deformation regions, thus improving rehabilitation outcomes.
Implementation Method 1
a load distribution sensor having a plurality of sensors arranged in a two-dimensional array to detect a distribution of a load applied from a user, and an inference device that has a CNN for performing a convolutional calculation process using two-dimensional data based on sensor outputs of the sensors as an input
Data Source
AI summary
A gait training system according to an embodiment includes a load distribution sensor having a plurality of sensors arranged in a two-dimensional array to detect a distribution of a load applied from a user, and an inference device that has a CNN for performing a convolutional calculation process using two-dimensional data based on sensor outputs of the sensors as an input, and that estimates a value of the load received from the user.