Service management

Patent metadata

Field	Value
Publication number	US-11455591-B2
Application number	US-201916515070-A
Country	US
Kind code	B2
Filing date	Jul 18, 2019
Priority date	Jul 18, 2019
Publication date	Sep 27, 2022
Grant date	Sep 27, 2022

Abstract

Method and system are provided for customer table service management. The method includes receiving sensor load data over time from a customer table. The method analyzes the sensor load data during a waiting time between a time of one or more customers arriving at the table and a time of consumables being served to the table to learn background noise data of the one or more customers. The method further analyzes the sensor load data during a dining time after the time of consumables being served to the table to detect one or more events that require a service action, wherein analyzing the sensor load data during the dining time removes the learnt background noise data to distinguish sensor load data changes relating to consumption of the consumables on the table. The method outputs event detection notifications to prompt the required service action.

First claim

What is claimed is:

1. A computer-implemented method comprising:
improving point of sale devices based, at least in in part, on predicted loads of a customer table, wherein improving point of sale devices comprises:
receiving sensor load data over time from the customer table;
analyzing the sensor load data during a waiting time between a time of one or more customers arriving at the table and a time of consumables being served to the table to learn background noise data of the one or more customers, wherein analyzing the sensor load data during the waiting time includes using the sensor load data as a training set for a machine learning algorithm to identify background noise data for one or more customers;
learning background noise data by distributing the one or more customers over time in a K-means clustering algorithm;
identifying respective zones of influence on the customer table to each of the one or more customers;
removing anomalous behavior that affects a predicted time in which consumables served to the table reaches completion based on its' potential influence on at least a respective zone of the identified zones associated with a respective customer; and
generating a visual interface display that is shown prior to a predicted time in which consumables served to the table reaches completion.

2. The method as claimed in claim 1 , wherein receiving sensor load data includes receiving data relating to a total load on the table, a load distribution on the table, and impulse data in the form of a sudden change in load on the table.

3. The method as claimed in claim 1 , wherein analyzing the sensor load data during the waiting time includes determining a zone of influence of a customer on the table by using a clustering algorithm to cluster the sensor load data by customer and identifying background noise data for each customer.

4. The method as claimed in claim 3 , wherein analyzing the sensor load data during the dining time analyzes the sensor load data for each customer during the dining time and removes the learnt background noise data for each customer.

5. The method as claimed in claim 1 , wherein analyzing the sensor load data during the dining time to detect one or more events that require a service action includes detecting an addition of consumable load data and monitoring the consumable load data until a plateau indicating a required service action.

6. The method as claimed in claim 1 , including modeling multiple events that require a service action in the form of a pattern of sensor load data over time and wherein analyzing the sensor load data during the dining time to detect one or more events that require a service action compares the sensor load data to the modeled events.

7. The method as claimed in claim 6 , wherein the analyzing of the sensor load data during the dining time is carried out by a machine learning algorithm using prior labelling of behavior of sensor load data over time based on past customer behavior.

8. The method as claimed in claim 6 , wherein the analyzing of the sensor load data during the dining time is carried out by a threshold algorithm to determine a threshold similarity for the comparison of sensor load data to the modeled events.

9. The method as claimed in claim 1 , including receiving details of the consumables ordered by the one or more customers and comparing known loads of the consumables to the sensor load data to determine that all consumables have been provided to the table.

10. The method as claimed in claim 1 , including prioritizing detected events for multiple tables based on service parameters.

11. A computer system comprising:
a processor and a memory configured to provide computer program instructions to the processor to execute the function of components of a service analyzing system including:
a sensor data receiving component for receiving sensor load data over time from a table in a food establishment;
a waiting time analyzing component for analyzing the sensor load data during a waiting time between a time of one or more customers arriving at the table and a time of consumables being served to the table to learn background noise data of the one or more customers;
a dining time analyzing component for analyzing the sensor load data during a dining time after the time of consumables being served to the table including an event detecting component for detecting one or more events that require a service action and a background component for analyzing the sensor load data during the dining time by removing the learnt background noise data to distinguish sensor load data changes relating to consumption of the consumables on the table, wherein analyzing the sensor load data during the waiting time includes using the sensor load data as a training set for a machine learning algorithm to identify background noise data for one or more customers;
a dining time analyzing component for:
learning background noise data by distributing the one or more customers over time in a K-means clustering algorithm,
identifying respective zones of influence on the customer table to each of the one or more customers,
removing anomalous behavior that affects a predicted time in which consumables served to the table reaches completion based on its' potential influence on at least a respective zone of the identified zones associated with a respective customer, and

an output component for generating a visual interface display that is shown prior to a predicted time in which consumables served to the table reaches completion.

12. The system as claimed in claim 11 , including a sensor system provided at a table and in data communication with the service analyzing system, wherein the sensor system includes one or more load sensors and one or more impulse sensors providing data relating to a total load on the table, a load distribution on the table, and impulse data in the form of a sudden change in load on the table.

13. The system as claimed in claim 11 , wherein the waiting time analyzing component includes a customer zone component for determining a zone of influence of a customer on the table by using a clustering algorithm to cluster the sensor load data by customer and identifying background noise data for each customer.

14. The system as claimed in claim 11 , wherein the event detecting component includes detecting an addition of consumable load data and monitoring the consumable load data until a plateau indicating a required service action.

15. The system as claimed in claim 11 , including an event modeling component for modeling multiple events that require a service action in the form of a pattern of sensor load data over time and wherein the dining time analyzing component compares the sensor load data to the modeled events.

16. The system as claimed in claim 15 , wherein the dining time analyzing component uses a machine learning algorithm or threshold algorithm using prior labelling of behavior of sensor load data over time based on past customer behavior.

17. The system as claimed in claim 11 , including a central console including a visual display for displaying event detection notifications provided by the output component.

18. A computer program product, the computer program product comprising:
a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:
improve point of sale devices based, at least in in part, on predicted loads of a customer table, wherein improving point of sale devices comp

Assignees

• IBM

Inventors

• Lee James
• Woodhouse Jack Anthony
• Rozes Heloise
• Kloepfer Dominik
• Jansen Grace

Classifications

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06Q10/06395

  Quality analysis or management · CPC title

• G01G19/4144

  for controlling weight of goods in commercial establishments, e.g. supermarket, P.O.S. systems · CPC title