Wireless Charging Object Classification Using ML Beacon Reflection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional wireless charging systems face challenges in accurately distinguishing between valid wireless power receivers and foreign objects, leading to potential damage to objects with magnetic strips or RFID chips and requiring precise positioning, which can result in frustrating user experiences and equipment damage.
Innovation Solution
The implementation of machine learning-based systems that use trained classifiers to detect and classify objects by analyzing reflected power signals from test transmission signals, allowing for the determination of authorized wireless power receivers and avoidance of foreign objects, even when the receiver has no power, without the need for sophisticated sensors or data communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional wireless charging systems transmit power without accurate object classification, then power delivery can begin immediately, but foreign objects with magnetic strips or RFID chips may be damaged
Solution Approach 1:
The system performs preliminary classification of objects using machine learning models before initiating power delivery. Electrical measurements are collected and analyzed to determine whether an object is a valid wireless power receiver or a foreign object that should not be charged, preventing damage before it occurs
Solution Approach 2:
The system continuously monitors electrical measurements during the charging process and uses machine learning classifiers to provide real-time feedback on object classification. This allows the system to adjust power delivery based on the classified object type, ensuring safe operation
2Loss of energy
If conventional wireless charging systems require precise positioning, then power transfer efficiency can be optimized, but user experience deteriorates due to frustration from inability to locate the exact charging position
Solution Approach 1:
The system uses multiple antenna zones that can be dynamically activated based on where a receiver is detected. The classification system works across multiple zones, allowing the charging pad to adapt to different object positions while maintaining efficient power transfer through selective zone activation
Solution Approach 2:
The machine learning classification system provides universal object detection and classification capability across the entire charging surface. Multiple antenna zones with classification circuits enable the system to handle objects at various positions uniformly, eliminating the need for precise positioning
3Device complexity
If conventional wireless charging systems use simple detection methods, then device complexity is reduced, but the ability to distinguish between valid receivers and foreign objects deteriorates
Solution Approach 1:
The system introduces machine learning classifiers as intermediary components between the electrical measurement circuits and the power delivery control. These classifiers process electrical measurements and provide refined classification decisions, improving accuracy without requiring complex hardware modifications
Solution Approach 2:
The system replaces complex hardware-based detection mechanisms with software-based machine learning classification. By using trained classifiers that analyze electrical measurements, the system achieves high classification accuracy without adding sophisticated sensors or complex mechanical detection systems
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 safe and efficient wireless power transmission by accurately identifying and avoiding foreign objects, preventing damage and ensuring authorized receivers receive power, while optimizing power transfer and user experience.
Implementation Method 1
a wireless-power-transmitting antenna is transmitting different power beacons
Implementation Method 2
analyzing reflected power signals from test transmission signals
Data Source
AI summary
An example method is provided for detecting and classifying foreign objects, performed at a computer system having one or more processors and memory storing one or more programs configured for execution by the one or more processors. The method includes obtaining a plurality of electrical measurements while a wireless-power-transmitting antenna is transmitting different power beacons. The method also includes forming a feature vector according to the plurality of electrical measurements. The method further includes detecting a presence of one or more foreign objects prior to transmitting wireless power to one or more wireless power receivers by inputting the feature vector to trained one or more classifiers, wherein each classifier is a machine-learning model trained to detect foreign objects distinct from the one or more wireless power receivers.