Occupancy sensing using ultra-wide band

Patent metadata

Field	Value
Publication number	US-11277166-B2
Application number	US-202016913271-A
Country	US
Kind code	B2
Filing date	Jun 26, 2020
Priority date	Jun 26, 2020
Publication date	Mar 15, 2022
Grant date	Mar 15, 2022

Abstract

Occupancy sensing using ultra-wideband (UWB) keyless infrastructure is provided. Channel impulse response (CIR) measurements are received from a plurality of UWB transceiver nodes arranged about a plurality of locations. A classification model it utilized to predict occupancy of each of the plurality of locations based on CIR tensors formed from the CIR measurements for each of the UWB transceiver nodes.

First claim

What is claimed is:

1. A method for occupancy sensing using ultra-wideband (UWB), comprising:
receiving channel impulse response (CIR) measurements from a plurality of UWB transceiver nodes arranged about a plurality of locations;
performing time alignment on the CIR measurements to time-align the CIR tensors across the UWB transceiver nodes; and
utilizing a classification model to predict occupancy of objects in each of the plurality of locations based on the CIR tensors as time-aligned, formed from the CIR measurements from each of the UWB transceiver nodes.

2. The method of claim 1 , further comprising:
periodically reassigning one of the plurality of UWB transceiver nodes to be a transmitter and the other of the plurality of UWB transceiver nodes to be receivers; and
collecting the CIR measurements of the one of the plurality of UWB transceiver nodes operating as a transmitter from the other of the UWB transceiver nodes operating as receivers.

3. The method of claim 1 , further comprising normalizing the CIR tensors before applying the CIR tensors to the classification model.

4. The method of claim 1 , wherein the classification model is a single-input multi-output classification model, and the CIR tensors of the plurality of UWB transceiver nodes are concatenated into a single 3D tensor as input to the single-input multi-output classification model.

5. The method of claim 1 , wherein the classification model is a multi-input multi-output classification model, each of the UWB transceiver nodes feeds data into a separate set of layers, and outputs of the layers for each of the UWB transceiver nodes are concatenated into a single layer for multi-task classification.

6. The method of claim 1 , wherein the classification model is a multi-input multi-output classification model using a multi-task mask, and further comprising:
utilizing the multi-task mask to identify multi-task attentions from the CIR tensors to produce multi-task weights for each seating location; and
weighting the outputs using the multi-task weights to produce outputs, as weighted, that account for spatial features correlated across the UWB transceiver nodes.

7. A system for occupancy sensing using ultra-wideband (UWB), comprising:
a computing device including a processor programmed to:
receive channel impulse response (CIR) measurements from a plurality of UWB transceiver nodes arranged about a plurality of locations;
perform time alignment on the CIR measurements to time-align the CIR tensors across the UWB transceiver nodes; and
utilize a classification model to predict occupancy of objects in each of the plurality of locations based on the CIR tensors as time-aligned formed from the CIR measurements from each of the UWB transceiver nodes.

8. The system of claim 7 , wherein the processor is further programmed to:
periodically reassign, per channel, one of the plurality of UWB transceiver nodes to be a transmitter and the other of the plurality of UWB transceiver nodes to be receivers; and
collect the CIR measurements of the one of the plurality of UWB transceiver nodes operating as a transmitter from the other of the UWB transceiver nodes operating as receivers.

9. The system of claim 7 , wherein the processor is further programmed to normalize the CIR tensors before applying the CIR tensors to the classification model.

10. The system of claim 7 , wherein the classification model is a single-input multi-output classification model, and the processor is further programmed to:
concatenate the CIR tensors of the plurality of UWB transceiver nodes into a single 3D tensor; and
input the 3D tensor to the single-input multi-output classification model.

11. The system of claim 7 , wherein the classification model is a multi-input multi-output classification model, and the processor is further programmed to:
feed data from each of the UWB transceiver nodes into a separate set of layers; and
concatenate outputs of the layers for each of the UWB transceiver nodes into a single layer for multi-task classification.

12. The system of claim 7 , wherein the classification model is a multi-input multi-output classification model using a multi-task mask, and the processor is further programmed to:
utilize the multi-task mask to identify multi-task attentions from the CIR tensors to produce multi-task weights for each seating location; and
weight the outputs using the multi-task weights to produce outputs, as weighted, that account for spatial features correlated across the UWB transceiver nodes.

13. A non-transitory computer-readable medium comprising instructions for occupancy sensing using ultra-wideband (UWB) that, when executed by a processor, cause the processor to:
receive channel impulse response (CIR) measurements from a plurality of UWB transceiver nodes arranged about a plurality of locations;
perform time alignment on the CIR measurements to time-align the CIR tensors across the UWB transceiver nodes; and
utilize a classification model to predict occupancy of objects in each of the plurality of locations based on the CIR tensors as time-aligned, formed from the CIR measurements from each of the UWB transceiver nodes.

14. The medium of claim 13 , further comprising instructions that, when executed by the processor, cause the processor to:
normalize the CIR tensors before applying the CIR tensors to the classification model; and
utilize the CIR tensors, as normalized and time-aligned, as input to the classification model.

15. The medium of claim 13 , further comprising instructions that, when executed by the processor, cause the processor to:
periodically reassign, per channel, one of the plurality of UWB transceiver nodes to be a transmitter and the other of the plurality of UWB transceiver nodes to be receivers; and
collect the CIR measurements of the one of the plurality of UWB transceiver nodes operating as a transmitter from the other of the UWB transceiver nodes operating as receivers.

16. The medium of claim 13 , wherein the classification model is a single-input multi-output classification model, and, further comprising instructions that, when executed by the processor, cause the processor to:
concatenate the CIR tensors of the plurality of UWB transceiver nodes into a single 3D tensor; and
input the 3D tensor to the single-input multi-output classification model.

17. The medium of claim 13 , wherein the classification model is a multi-input multi-output classification model, and further comprising instructions that, when executed by the processor, cause the processor to:
feed data from each of the UWB transceiver nodes into a separate set of layers; and
concatenate outputs of the layers for each of the UWB transceiver nodes into a single layer for multi-task classification.

18. The medium of claim 13 , wherein the classification model is a multi-input multi-output classification model using a multi-task mask, and further comprising instructions that, when executed by the processor, cause the processor to:
utilize the multi-task mask to identify multi-task attentions from the CIR tensors to produce multi-task weights for each seating location; and
weight the outputs using the multi-task weights to produce outputs, as weighted, that account for spatial features correlated across the UWB transceiver nodes.

Assignees

• Bosch Gmbh Robert

Inventors

• Zeng Yunze
• Ma Yongsen
• Jain Vivek

Classifications

• G06F18/24

  Classification techniques · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/048

  Activation functions · CPC title

• H04B1/719

  Interference-related aspects · CPC title

• H04B2001/6908

  using time hopping · CPC title