RFIC Hardware Allocation Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for configuring radio frequency integrated circuits (RFICs) in User Equipment (UE) for Carrier Aggregation and Multiple-Input, Multiple-Output (MIMO) modes face challenges due to hardware constraints, where not all connections between local oscillators and mixers, or mixers and active ports, are available, making manual assignment complex and inefficient.
Innovation Solution
A machine learning-based method using a neural network is employed to automate the assignment of mixers and local oscillators to active ports by generating estimated quality values for state transitions, optimizing connections to maximize the likelihood of connecting all active ports, leveraging reinforcement learning and Monte-Carlo tree search to find optimal mixer and LO assignments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual assignment of mixers and local oscillators to active ports is performed, then hardware constraints can be satisfied, but the configuration process becomes complex and inefficient
Solution Approach 1:
The system employs a neural network that automatically performs the mixer and local oscillator assignment without human intervention. The neural network receives the hardware constraint specifications and autonomously determines the optimal connections, eliminating the need for manual configuration while satisfying all hardware limitations.
Solution Approach 2:
The patent replaces the manual mechanical configuration process with an automated computational system. Instead of physically connecting components based on manual analysis, the system uses neural network algorithms to compute optimal assignments, substituting human cognitive effort with machine learning-based automation.
2Quantity of substance
If the number of local oscillators is reduced to minimize hardware, then the number of possible connections decreases, but this may limit the ability to connect all active ports
Solution Approach 1:
The system dynamically adjusts the connection configuration parameters based on the available hardware resources. The neural network optimizes the assignment of mixers and local oscillators to active ports, finding valid configurations even when the number of local oscillators is minimized, thereby maintaining connection completeness despite reduced hardware quantity.
Solution Approach 2:
The patent performs feasibility checks and preliminary analysis before finalizing the configuration. The system evaluates whether a given number of local oscillators can satisfy all active port connections, and adjusts the assignment strategy in advance to ensure that connection completeness is achieved with the minimized hardware count.
3Adaptability or versatility
If more connections are made between local oscillators and mixers to handle Carrier Aggregation and MIMO modes, then the adaptability increases, but the hardware complexity and difficulty of configuration increase
Solution Approach 1:
The neural network-based configuration system provides a universal solution that handles multiple operational modes (Carrier Aggregation, MIMO, and their combinations) through a single automated platform. The system adapts to different mode requirements by generating appropriate connection assignments without requiring separate configuration procedures for each mode.
Solution Approach 2:
The system dynamically generates connection configurations based on the required operational mode. Instead of using a fixed static configuration, the neural network adapts the mixer and local oscillator assignments in real-time according to the active ports needed for Carrier Aggregation and MIMO operations, optimizing the connection topology for each specific mode requirement.
Data Source
AI summary
A system and method for configuring an RF network based on machine learning. In some embodiments, the method includes: receiving, by a first neural network, a first state and a first state transition, the first state including: one or more identifiers for available active ports, and a set of available connections between two or more circuit elements, each of the circuit elements being one of: (1) a first circuit type, (2) a second circuit type that operatively connects a circuit element of the first circuit type to one of the available active ports, and (3) the available active ports; and generating, by the first neural network, a first estimated quality value, for the first state transition.