# US 2026/0095867 A1 · Methods and systems for low power wide area network localization

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0095867 A1
- Document ID: US20260095867A1
- Status: Published application
- Kind: A1
- Application number: 19/254,595
- Filed: 2025-06-30
- Published: 2026-04-02
- Earliest claimed: 2019-07-10
- Assignees: Swisscom AG
- Applicants: Swisscom AG
- Inventors: Ahmed Saadallah
- Web page: https://patentgrep.com/patent/US20260095867A1

## CPC

- H04W 52/0245: {according to signal strength}
- G01S 5/02: using radio waves (using satellite radio beacon systems for determining position G01S19/00)
- G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/063: using electronic means
- G06N 3/09: Supervised learning
- H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72)
- G01S: RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Methods and systems are provided for low power wide area network localization. Signals transmitted between a mobile device and one or more network elements of a wireless network may be obtained while the mobile device is moving within a reception area of the wireless network. The one or more network elements have known locations. The obtained signals may be evaluated using a model that is configured to use input values relating to a signal quality parameter of electromagnetic signals. The evaluating may include deriving for each obtained signal, one or more values representative of the signal quality parameter of the signal, applying the values as input signals into the model, and generating or adjusting, based on applying of the values into the model, values representative of a location of the mobile device and a location dependent measure of uncertainty in the location corresponding to each generated location value.

## Claims

### Claim at position 1

1-22. (canceled)

### Claim 23

A method comprising:
obtaining, while a mobile device is moving within a reception area of a wireless network, one or more signals transmitted between the mobile device and one or more network elements of the wireless network, wherein the one or more network elements have known locations; and
evaluating the one or more signals, using a model that is configured to use input values relating to a signal quality parameter of electromagnetic signals, wherein the evaluating comprises:
deriving for each signal of the one or more signals, one or more values representative of the signal quality parameter of the signal;
applying the values as input signals into the model; and
generating or adjusting, based on applying of the values into the model, values representative of a location of the mobile device and a location dependent measure of uncertainty in the location corresponding to each generated location value.

### Claim 24 (depends on 23)

The method of claim 23, further comprising, using the model, determining the influence of prior input values or prior locations during generation of a current values representative of the location of the mobile device.

### Claim 25 (depends on 23)

The method of claim 23, wherein the model comprises a trained neural network model having a sequential encoder and decoder stage.

### Claim 26 (depends on 25)

The method of claim 25, wherein the encoder comprises a recurrent neural network (RNN), long short-term memory (LSTM), or an encoder stage of a transformer.

### Claim 27 (depends on 23)

The method of claim 23, further comprising, using the model, treating location values in the model as continuous.

### Claim 28 (depends on 23)

The method of claim 23, further comprising using a temporal sequence of a plurality of subsequent input signals as a simultaneous input to the model, and wherein an input layer of the model comprises a plurality of parallel units configured for receiving the temporal sequence.

### Claim 29 (depends on 23)

The method of claim 23, wherein the model comprises one or more elements configured for determining weights representative of at least one characteristics associated with one or more other locations or other input signals representative of locations to generated values representative of actual location.

### Claim 30 (depends on 29)

The method of claim 29, further comprising using or applying output of the one or more elements as input into a decoder stage of the model.

### Claim 31 (depends on 23)

The method of claim 23, further comprising, using the model, encoding the input signals based on positions within a temporal sequence of input signals.

### Claim 32 (depends on 23)

The method of claim 23, wherein determining location dependent uncertainty comprises testing a trained version of a neural network model using training data not used during training of the model with a multitude of different dropout conditions.

### Claim 33 (depends on 32)

The method of claim 32, wherein dropouts comprise horizontal dropouts between parallel units of the neural network.

### Claim 34 (depends on 32)

The method of claim 32, wherein dropouts comprise horizontal dropouts between parallel units of the neural network and vertical dropouts between vertically stacked layers of the neural network model.

### Claim 35

A system comprising:
one or more circuits configured to:
obtain, while a mobile device is moving within a reception area of a wireless network, one or more signals transmitted between the mobile device and one or more network elements of the wireless network, wherein the one or more network elements have known locations; and
evaluate the one or more signals, using a model that is configured to use input values relating to a signal quality parameter of electromagnetic signals, wherein the evaluating comprises:
deriving for each signal of the one or more signals, one or more values representative of the signal quality parameter of the signal;
applying the values as input signals into the model; and
generating or adjusting, based on applying of the values into the model, values representative of a location of the mobile device and a location dependent measure of uncertainty in the location corresponding to each generated location value.

### Claim 36 (depends on 35)

The system of claim 35, wherein the one or more circuits are configured to, using the model, determine the influence of prior input values or prior locations during generation of a current values representative of the location of the mobile device.

### Claim 37 (depends on 35)

The system of claim 35, wherein the one or more circuits are configured to, using the model, treat location values in the model as continuous.

### Claim 38 (depends on 35)

The system of claim 35, wherein the one or more circuits are configured to use a temporal sequence of a plurality of subsequent input signals as a simultaneous input to the model, and wherein an input layer of the model comprises a plurality of parallel units configured for receiving the temporal sequence.

### Claim 39 (depends on 35)

The system of claim 35, wherein the one or more circuits are configured to, using the model, encoding the input signals based on positions within a temporal sequence of input signals.

### Claim 40 (depends on 35)

The system of claim 35, wherein the one or more circuits are configured to, using the one or more elements in the model, determine weights representative of at least one characteristics associated with one or more other locations or other input signals representative of locations to generated values representative of actual location.

### Claim 41 (depends on 40)

The system of claim 40, wherein the one or more circuits are configured to use or apply output of the one or more elements as input into a decoder stage of the model.

### Claim 42 (depends on 35)

The system of claim 35, wherein the one or more circuits are configured to, when determining location dependent uncertainty, test a trained version of a neural network model using training data not used during training of the model with a multitude of different dropout conditions.
