17848283. MACHINE LEARNING (ML)-BASED DYNAMIC DEMODULATOR SELECTION simplified abstract (QUALCOMM Incorporated)
Contents
MACHINE LEARNING (ML)-BASED DYNAMIC DEMODULATOR SELECTION
Organization Name
Inventor(s)
Shay Landis of Hod Hasharon (IL)
Alexei Yurievitch Gorokhov of San Diego CA (US)
Hari Sankar of San Diego CA (US)
Peer Berger of Hod Hasharon (IL)
MACHINE LEARNING (ML)-BASED DYNAMIC DEMODULATOR SELECTION - A simplified explanation of the abstract
This abstract first appeared for US patent application 17848283 titled 'MACHINE LEARNING (ML)-BASED DYNAMIC DEMODULATOR SELECTION
Simplified Explanation
The patent application describes a method for wireless communication using a receiving device and an artificial neural network. Here are the key points:
- The method involves predicting the least complex demodulator that will achieve a specific goal for each data block in a set of data blocks.
- The prediction is based on the expected features of the data block that will be received by the receiving device.
- A set of demodulators with different levels of complexity is available, and the method dynamically selects the least complex demodulator based on the features of the data block.
- The selected demodulator is then used to demodulate the data block.
Potential applications of this technology:
- Wireless communication systems
- Mobile devices
- Internet of Things (IoT) devices
- Satellite communication
- Radio communication
Problems solved by this technology:
- Efficient use of resources by selecting the least complex demodulator for each data block
- Improved performance and accuracy in wireless communication
- Adaptability to different types of data blocks and communication scenarios
Benefits of this technology:
- Enhanced communication efficiency and reliability
- Reduced power consumption
- Improved data transmission rates
- Increased flexibility in adapting to changing communication conditions
Original Abstract Submitted
A method for wireless communication by a receiving device includes predicting with an artificial neural network, at each data block of a set of data blocks, a least complex demodulator that will achieve a goal. The predicting is based on features of a data block expected to be received at the receiving device. The method also includes dynamically selecting the least complex demodulator, from a set of demodulators with different levels of complexity, based on the features of the data block expected to be received. The method further includes demodulating the data block with the selected demodulator for the data block.