Tactile Sensor Signal Processing via Cellular Neural Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing tactile sensor signal processing methods are slow, making it difficult to respond quickly to sudden forces and fail to properly handle sliding and torsion events, which is problematic in robotics and other fields requiring real-time force distribution evaluation.
Innovation Solution
A signal processing method that uses greyscale image coding of force maps generated from elementary sensors' data, processed by a cellular neural network to recognize and classify three-dimensional force events, including sliding and torsion, enabling faster intervention and control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a large number of elementary sensors are arranged along the tactile surface to achieve higher sensing reliability and capture force distributions, then measurement precision and reliability are improved, but processing time increases and productivity decreases
Solution Approach 1:
The tactile sensor array is divided into multiple sensor groups, where each group processes a specific region or aspect of tactile information independently. This segmentation allows parallel processing of force distributions across different regions, reducing overall processing time while maintaining comprehensive sensing coverage and reliability.
Solution Approach 2:
The patent transforms the processing approach by introducing a new dimension of analysis through force map generation and event classification. Instead of processing raw sensor signals sequentially, the system creates two-dimensional force maps and classifies events based on patterns in these maps, enabling faster parallel processing while maintaining sensing reliability.
2Measurement precision
If traditional signal processing methods are used to evaluate force distributions from multiple elementary sensors, then comprehensive force map evaluation is achieved, but response time increases making rapid decision-making difficult
Solution Approach 1:
The system performs preliminary processing by generating force maps from elementary sensor signals before full evaluation is needed. These force maps pre-organize the data from multiple sensors into a structured format that facilitates rapid event recognition and classification, reducing the time required for comprehensive force distribution evaluation when tactile events occur.
3Adaptability or versatility
If known signal processing methods are applied to tactile sensor data, then basic force sensing is achieved, but sliding and torsion events cannot be properly recognized and classified
Solution Approach 1:
The patent implements dynamic event classification that adapts to different types of tactile events including sliding and torsion. The system dynamically adjusts its analysis based on patterns detected in the force maps, enabling reliable recognition and classification of diverse events such as sliding, torsion, and pressing by identifying characteristic force distribution patterns for each event type.
Data Source
Figure 1a~2b
Figure 3a~4b
Figure 5a~6b
AI summary
The invention is a signal processing method for a tactile sensor, said tactile sensor having a tactile surface and elementary sensors (10) being suitable for three dimensional force sensing and being arranged along the tactile surface, wherein sensing is carried out by reading signals of the elementary sensors (10). For each elementary sensor (10), force components (Tx, Ty, Sn) measured along three spatial axes normal to each other are determined, three force maps are generated from the respective force components (Tx, Ty, Sn) along the same axis, the three force maps are evaluated, and then the results of the three evaluations are jointly evaluated. The invention also relates to a tactile sensor having a tactile surface and elementary sensors (10) suitable for three dimensional force sensing and arranged along the tactile surface. The inventive tactile sensor comprises sensor units consisting of four elementary sensors (10) arranged in the corner points of a square.