How Phones and Towers Use AI to Improve Wireless Signal Quality

This patent describes how a mobile phone and a base station work together, using machine learning, to accurately measure and predict the quality of the wireless connection, making communication more reliable.

What it covers

This patent outlines a method where a mobile phone, called a User Equipment (UE), helps a base station improve its understanding of the wireless signal path, known as the channel. The UE first receives a 'first message' from the base station with instructions for sending a special 'uplink reference signal' (Claim 1). After transmitting this signal, the UE receives another special signal, a 'downlink reference signal,' from the base station. The UE then measures the quality of the downlink signal and sends this 'channel information' back to the base station. Crucially, the 'first message' also includes information about a machine learning model that the base station uses to learn and improve its channel measurements (Claim 1). For example, a phone might tell the network its capability to support this learning, or request to change its learning behavior based on how busy it is or how fast it's moving (Claims 3, 6).

What it doesn't cover

• —Does not cover channel estimation methods that do not involve a machine learning model for measuring the channel.
• —Does not cover scenarios where the mobile phone itself performs the primary machine learning for channel estimation, as the patent specifies 'the learning is performed by the base station' (Claim 1).
• —Does not cover systems where the mobile phone does not transmit an uplink reference signal based on configuration from the base station.
• —Does not cover systems where the mobile phone does not transmit channel information measured from a downlink reference signal.
• —Does not cover situations where the mobile phone does not communicate its learning support capabilities or desired learning modes to the base station.

The clever bit

The novelty lies in the specific, collaborative interaction between the UE and the base station, where the UE actively provides data and state information to help the base station train and use a machine learning model for better channel estimation. This offloads the heavy computational learning task to the base station while leveraging the UE's local context.

Why it matters

Accurate channel estimation is vital for efficient and reliable wireless communication, especially in complex environments like urban areas or with fast-moving users. By using machine learning, base stations can adapt more intelligently to changing conditions, leading to fewer dropped calls, faster data speeds, and better overall network performance. This technology is foundational for advanced wireless systems like 5G and future 6G networks, which demand extremely precise signal management.

Real-world examples

• 1.5G New Radio (NR) networks
• 2.Future 6G wireless communication systems
• 3.Advanced cellular base stations
• 4.Smartphones and IoT devices operating on modern cellular networks