US 12,413,369 B2
Wireless communication device, RU selection device, and method
Dheeraj Kotagiri, Tokyo (JP)
Assigned to NEC CORPORATION, Tokyo (JP)
Appl. No. 17/801,447
Filed by NEC Corporation, Tokyo (JP)
PCT Filed Feb. 27, 2020, PCT No. PCT/JP2020/008124
§ 371(c)(1), (2) Date Aug. 22, 2022,
PCT Pub. No. WO2021/171510, PCT Pub. Date Sep. 2, 2021.
Prior Publication US 2023/0114359 A1, Apr. 13, 2023
Int. Cl. H04L 5/00 (2006.01); H04W 72/02 (2009.01)
CPC H04L 5/0058 (2013.01) [H04W 72/02 (2013.01)] 13 Claims
OG exemplary drawing
 
1. A wireless communication device comprising:
a resource unit (RU) selection module, configured to determine an RU used to transmit data from the wireless communication device to an access point;
wherein the RU selection module comprises:
at least one memory storing instructions; and
at least one processor configured to execute the instructions to:
generate device state information based on carrier sensing data indicating whether or not energy is detected in each RU used in wireless communication between the wireless communication device and an access point and acknowledgement signals received from the access point;
apply the device state information to a convolutional neural network (CNN) including one or more convolutional layers extracting features representative of an interference between the wireless communication device and other wireless communication device in a wireless communication channel and fully connected layers to extract features from the device state information representative of interactions of a wireless communication station with other wireless communication stations, and to generate scores for each RU representative of possibility of successful transmission based on the extracted features;
compute a reward for providing feedback on prediction accuracy of the generated scores;
determine an RU used to transmit data based on the generated scores;
store experiences including a set of the device state information, an index of RU in which packet is transmitted, updated device state information post packet transmission, and the reward in a memory unit;
sample batches of experiences from the memory unit;
compute a first score for each RU by applying the device state information included in the sampled batch to the CNN;
compute a second score for each RU by applying the updated device state information to the CNN;
compute a difference between the first score and the second score as an error; and
update the CNN based on the computed error.