Proximity Sensor Signal Resampling for Hand Movement Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in recognizing dynamic hand movements due to varying vector sizes and asynchronous data from proximity sensors, making it difficult to apply shape-recognition algorithms effectively.
Innovation Solution
A method that resamples signals from proximity sensors to generate a set of date-stamped samples with regular time intervals and synchronized data across sensors, allowing for the use of simpler classification algorithms like support vector machines for movement recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If dynamic time warping is used for hand movement recognition, then variable-sized vectors from sensors can be processed, but the algorithm complexity increases significantly
Solution Approach 1:
The patent applies preliminary action by pre-processing sensor data to generate fixed-size feature vectors before classification. The system extracts features (distance, coverage, convergence time) and resamples them to create standardized input vectors, eliminating the need for complex dynamic time warping algorithms while maintaining ability to handle variable-duration hand movements
Solution Approach 2:
The patent changes the parameter representation from variable-sized raw sensor vectors to fixed-size processed vectors. By transforming the input format through feature extraction and resampling, the system converts adaptable variable-sized inputs into standardized fixed-sized outputs that can be processed by simpler classification algorithms
2Measurement precision
If multiple proximity sensors are used for hand movement recognition, then coverage and accuracy improve, but data synchronization becomes difficult
Solution Approach 1:
The system performs preliminary synchronization by establishing a common time reference before processing sensor data. It defines a reference time based on the first sensor's measurement and adjusts timestamps from subsequent sensors relative to this reference, enabling coherent integration of asynchronous sensor data without complex real-time synchronization
Solution Approach 2:
The patent introduces an intermediary time reference framework that mediates between asynchronous sensor measurements. By using a common reference time and calculating time differences relative to this reference, the system enables straightforward temporal alignment of data from multiple sensors without requiring complex coordination protocols
3Speed
If sensor data is sampled at irregular intervals, then the system can respond to dynamic hand movements, but the data cannot be processed by traditional classification algorithms
Solution Approach 1:
The system performs preliminary resampling to transform irregularly timed sensor measurements into uniformly spaced data points. By creating a standardized time grid and interpolating or selecting appropriate samples, it converts the irregular sampling pattern into a regular pattern that is compatible with traditional classification algorithms while preserving the dynamic response capability
Solution Approach 2:
The patent changes the temporal parameter structure from irregular sampling intervals to regular sampling intervals. Through resampling and time-grid alignment, it transforms the time-domain characteristics of the data to match the input requirements of classification algorithms, enabling use of simpler algorithms like support vector machines
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
This approach enables accurate recognition of hand movements by converting variable-sized vectors into fixed-size vectors, facilitating the use of simpler algorithms and improving the synchronization of data from multiple sensors.
Implementation Method 1
These sensors emit an infrared beam in the direction of an object and measure the time-of-flight (TOF technology) of this beam, in other words the time taken between its emission and its reception by the sensor after reflection on the object.
Implementation Method 2
These sensors emit an infrared beam in the direction of an object
Data Source
AI summary
The method for processing signals originating for example from several proximity sensors for the recognition of a movement of an object, comprises first respective samplings of the said signals delivered by the sensors so as to obtain a first set of first date-stamped samples, the generation, from the first set of first date-stamped samples, of new sampling times comprising a start of movement time, an end of movement time, and times regularly spaced between the start of movement time and the end of movement time, a re-sampling of the signal delivered by each sensor between the start of movement time and the end of movement time at the said new sampling times using the first samples, in such a manner as to generate a second set of second date-stamped samples, and a processing of the said second set of date-stamped samples by a movement recognition algorithm.