Radar-Based Object Classification for Obscured Items
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for object classification, particularly in ubiquitous computing and tangible user interfaces, face challenges with non-destructive and non-disruptive approaches that can accurately identify materials, shapes, colors, orientations, and positions of objects without illumination, especially when objects are obscured or composed of composite materials.
Innovation Solution
A radar-based classification method that uses a radar unit to receive and classify radar signals reflected from objects, employing machine learning classifiers like random forest or deep neural networks to distinguish between different materials, shapes, colors, and orientations, even when objects are packaged or obscured, and can operate in proximity to the object without the need for illumination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If radar-based classification is used to identify obscured or packaged objects without illumination, then object recognition capability is improved, but device complexity increases
Solution Approach 1:
The patent replaces optical sensing mechanisms (cameras, illuminators) with radar-based electromagnetic sensing. The radar unit transmits electromagnetic signals that penetrate packaging materials and obscured surfaces, reflecting off the target object to provide classification data without requiring line-of-sight visibility or external illumination.
Solution Approach 2:
The patent introduces a radar unit as an intermediary sensing device between the computing device and the target object. This radar intermediary emits and receives electromagnetic signals that can pass through packaging materials, effectively mediating the detection process when direct optical sensing fails.
2Productivity
If non-destructive sensing methods are used to classify objects in real-time, then productivity is improved, but measurement precision may deteriorate
Solution Approach 1:
The patent performs preliminary classification by analyzing radar signal characteristics (reflection patterns, absorption coefficients, time-of-flight) to identify material properties before any physical interaction with the object. This preliminary electromagnetic sensing enables real-time identification without requiring destructive sampling or prolonged measurement periods.
Solution Approach 2:
The patent analyzes multiple radar signal parameters including reflection intensity, absorption coefficients, time-of-flight, and frequency characteristics to classify materials. By monitoring changes in these electromagnetic parameters as signals interact with different materials, the system achieves precise real-time identification without physical contact or destruction of the object.
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
Enables accurate, non-destructive, and non-disruptive classification of objects, allowing for recognition of materials, shapes, and orientations in real-time, even when objects are not visible or are obscured, using a portable and versatile radar system integrated into devices like mobile phones or smartwatches.
Implementation Method 1
Radar uses an emission of electromagnetic radio waves, for example with a frequency within 1 GHz-300 GHz, which is then reflected back from an object and received by a detector. The time of flight may be used to calculate the distance to an object, and using the Doppler shift the velocity of the object may also be measured.
Implementation Method 2
The time of flight may be used to calculate the distance to an object, and using the Doppler shift the velocity of the object may also be measured.
Implementation Method 3
Properties which may affect the received radar intensity from an object may include, for example, the absorption and scattering properties of a material of the object at the wavelengths used
Implementation Method 4
Properties which may affect the received radar intensity from an object may include, for example, the absorption and scattering properties of a material of the object at the wavelengths used
Implementation Method 5
The received signal may have contributions from reflection from the bottom surface of the object, scattering from the internal structure of the object, and reflection from the rear surface of the material.
Data Source
AI summary
A classification method comprises positioning an object and a radar unit in proximity to each other; receiving by the radar unit radar signals reflected from the object; and classifying the object, wherein the classifying is based on the radar signals and/or at least one feature extracted from the radar signals, and the classifying of the object comprises determining classification data for the object.