Method and internet of things system of charging information recommendation for new energy vehicle in smart city

Patent metadata

Field	Value
Publication number	US-12475502-B2
Application number	US-202318450363-A
Country	US
Kind code	B2
Filing date	Aug 15, 2023
Priority date	Oct 14, 2022
Publication date	Nov 18, 2025
Grant date	Nov 18, 2025

Abstract

The present disclosure provides a method of charging information recommendation for a new energy vehicle in a smart city. This method is executed by a charging pile management platform. This method includes: based on the service platform, obtaining a charging request of a user by the user platform; based on the charging request, determining candidate charging piles; based on queuing information of the candidate charging piles and time information for going to the candidate charging piles, sorting the candidate charging piles to determine an object charging pile; processing the travel demand information, location information, and the environmental information of the object charging pile based on a recommendation model to determine first sub-recommendation information; based on a failure rate of the object charging pile, determining second sub-recommendation information; based on the first sub-recommendation information and the second sub-recommendation information, determining recommendation information; and feeding back the recommendation information to the user.

First claim

What is claimed is:

1 . A method of charging information recommendation for a new energy vehicle in a smart city, which is realized by an Internet of Things (IoT) system of charging information recommendation for a new energy vehicle in a smart city, the IoT system including a user platform, a service platform and a charging pile management platform;
the method being executed by the charging pile management platform, the method comprising: based on the service platform, obtaining a charging request of a user by the user platform, wherein the charging request includes travel demand information; based on the charging request, determining candidate charging piles; based on queuing information of the candidate charging piles and time information for going to the candidate charging piles, sorting the candidate charging piles to determine an object charging pile; processing the travel demand information and location information of the object charging pile, and the environmental information of the object charging pile based on a recommendation model to determine first sub-recommendation information, the recommendation model being a machine learning model; based on a failure rate of the object charging pile, determining second sub-recommendation information; determining a first weight of the first sub-recommendation information and a second weight of the second sub-recommendation information based on times of iterations of a knowledge map; wherein nodes of the knowledge map include charging piles, and for each of the nodes, a node feature of the node includes a failure rate of a charging pile corresponding to the node; an edge of the knowledge map is used to connect two nodes whose similarity is greater than a preset threshold; and an edge feature of the edge includes the similarity of the two nodes connected by the edge; based on the first weight of the first sub-recommendation information and the second weight of the second sub-recommendation information, determining the recommendation information; and based on the service platform, feeding back the recommendation information to the user through the user platform.

2 . The method of claim 1 , wherein the recommendation model includes multiple input layers and an output layer, and the multiple input layers and the output layer are obtained through joint training; training data includes sample historical travel demand information, sample location information of a historical object charging pile, and sample historical environmental information, a label is sample first sub-recommendation information corresponding to the training data; and the training includes:
inputting the sample historical travel demand information, the sample location information of the historical object charging pile, and the sample historical environmental information into one of the multiple input layers, respectively, to obtain a sample travel feature vector, a sample location feature vector, and an environmental feature vector; inputting the sample travel feature vector, the sample location feature vector, and the environmental feature vector into the output layer to obtain the first sub- recommendation information output by the output layer; and constructing a loss function based on the first sub-recommendation information output by the output layer and the sample first sub-recommendation information, and training and updating the multiple input layers and the one output layer based on the loss function.

3 . The method of claim 1 , wherein the charging pile management platform comprises a main management platform database and multiple management sub-platforms, the method further comprises: dividing the multiple management sub-platforms based on urban areas.

4 . The methods of claim 3 , wherein determining the candidate charging piles based on the charging request includes:
based on multiple management sub-platforms, determining the candidate charging piles, and transmitting the candidate charging piles to the main management platform database through the multiple management sub-platforms; based on the queuing information of the candidate charging piles and time information for going to the candidate charging piles, sorting the candidate charging piles to determine an object charging pile includes: by the main management platform database, sorting the candidate charging piles based on the queuing information of the candidate charging piles and the time information for going to the candidate charging piles, and determining the object charging pile.

5 . The method of claim 3 , wherein the IoT system also includes a sensor network platform and an object platform; the method further comprises:
obtaining relevant information of charging piles based on the object platform; transmitting the relevant information of the charging piles to the multiple management sub-platforms based on the sensor network platform, so that the multiple management sub-platforms determine the candidate charging piles based on the relevant information.

6 . The method of claim 5 , wherein the sensor network platform includes a general sensor network platform database and a plurality of sensor network sub-platforms, the method further comprises: storing and processing different data sent by the object platform respectively based on the plurality of sensor network sub-platforms.

7 . The method of claim 2 , wherein predicting a failure rate of the object charging pile includes:
based on the knowledge map, determining the failure rate of the object charging pile.

8 . The method of claim 1 , wherein based on queuing information of the candidate charging piles and time information for going to the candidate charging piles, sorting the candidate charging piles to determine an object charging pile includes:
obtaining first basic information of the candidate charging piles and a charging portrait of the user; based on at least one of the queuing information, time information, the first basic information, and the charging portrait, determining a sorting result through a sorting model, wherein the sorting model is a machine learning model; and based on the sorting result, determining the object charging pile.

9 . The method of claim 8 , wherein the charging portrait is a user feature vector which represents tendency of the user to select a charging pile; the method further comprises:
training a prediction model to obtain a feature extraction model, wherein the prediction model and the feature extraction model are machine learning models; extracting second basic information and historical charging data of the user based on the feature extraction model to obtain the user feature vector.

10 . The method of claim 9 , wherein the prediction model includes a feature extraction layer and a judgment layer;
the feature extraction layer outputs a first user feature vector based on second basic information and historical charging data of a first user; the feature extraction layer outputs a second user feature vector based on second basic information and historical charging data of a second user; and the judgment layer judges a similarity of charging piles selected by the first user and the second user based on the first user feature vector and the second user feature vector.

11 . The method according to claim 10 , wherein the training samples for training the prediction model include a first group of training samples and a second group of training samples;
the first group of training samples includes multiple pairs of first training data, each pair of first training data is second basic information and historical charging data corresponding to two users who initiate charging requests in s

Assignees

• Chengdu Qinchuan Iot Tech Co Ltd

Inventors

• Shao Zehua
• Zhou Junyan
• Liu Bin
• Liang Yongzeng
• Wu Yuefei

Classifications

• H04L67/12

  specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks · CPC title

• H04L67/52

  specially adapted for the location of the user terminal · CPC title

• G16Y10/40

  Transportation · CPC title

• Y02T90/16

  Information or communication technologies improving the operation of electric vehicles · CPC title

• Y02T90/12

  Electric charging stations · CPC title