System to determine item weight using data from weight sensors

Patent metadata

Field	Value
Publication number	US-11468400-B1
Application number	US-201815938992-A
Country	US
Kind code	B1
Filing date	Mar 28, 2018
Priority date	Mar 28, 2018
Publication date	Oct 11, 2022
Grant date	Oct 11, 2022

Abstract

One or more load cells measure the weight of items at a fixture. Weight changes occur as items are picked from or placed to the fixture and may be used to determine when the item was picked or placed, quantity and so forth. Individual weights for a type of item may vary. A set of data comprising weight changes associated with interactions involving a single one of a particular type of item is gathered. These may be weight changes due to picks, places, or both. A model, such as a probability distribution, may be created that relates a particular weight of that type of item to a probability. The model may then be used to process other weight changes and attempt to determine what type of item was involved in an interaction.

First claim

What is claimed is:

1. A system comprising:
a platform to stow one or more types of items;
at least one weight sensor to measure a weight on the platform; and
one or more processors to execute instructions to:
determine, based on first output from the at least one weight sensor, a first plurality of event data, wherein each event data of the first plurality of event data is indicative of a weight change at a particular area of the platform;
determine a first confidence value indicative of a likelihood that individual ones of the first plurality of event data are representative of a pick or place of a single item;
determine a first set of event data comprising:
the event data having the first confidence value that is greater than a first threshold value; and
the weight changes for the each event data in the first plurality of event data that is deemed to be associated with a first type of item designated as being stowed in the particular area;

determine, using the first set of event data, a first probability distribution;
determine, based on second output from the at least one weight sensor, second event data indicative of a second weight change at the particular area of the platform;
determine, based on the first probability distribution, a first probability that the second weight change involves the first type of item;
determine the first probability exceeds a second threshold value;
determine, based on the second weight change, a change in quantity of the first type of item supported by the platform; and
perform an action, responsive to the determination of the change in quantity of the first type of item at the platform, the action comprising one or more of:
billing an account for the change in quantity of the first type of item, or
generating an order to restock the first type of item at the platform.

2. The system of claim 1 , wherein the first probability distribution is a convex combination of the weight changes in the first set of event data.

3. The system of claim 1 , the first set of event data further comprising a plurality of a most recent k instances of event data, wherein k is an integer value greater than zero.

4. A method comprising:
acquiring, from one or more weight sensors of a fixture, a plurality of weight measurements associated with a first type of item stowed at a particular area of the fixture;
determining a plurality of detected weight changes for the first type of item based on the plurality of weight measurements;
determining a first confidence value indicative of a likelihood that individual ones of the plurality of detected weight changes are representative of a pick or place of a single item;
determining a first set of detected weight changes, the first set of detected weight changes comprising the detected weight changes having the first confidence value greater than a first threshold value;
determining, based on the first set of detected weight changes, a first item weight model for the first type of item at the particular area of the fixture, the first item weight model comprising a distribution of the first set of detected weight changes and corresponding confidence values indicative of a likelihood that the first set of detected weight changes are associated with a pick or place of the first type of item at the particular area of the fixture;
determining a weight change at the fixture;
determining, based on the first item weight model and the weight change, a second confidence value representative of a likelihood that the weight change involves the first type of item at the particular area of the fixture;
determining the second confidence value exceeds a second threshold value;
determining a change in quantity of the first type of item at the particular area of the fixture; and
responsive to the change in quantity of the first type of item at the particular area of the fixture, performing an action comprising one or more of:
billing an account for the change in quantity of the first type of item, or
generating an order to restock the first type of item at the fixture.

5. The method of claim 4 , wherein the determining the change in quantity of the first type of item at the particular area of the fixture comprises:
accessing a first weight measurement and a second weight measurement obtained by the one or more weight sensors of the fixture;
determining the weight change based on a difference between the first weight measurement and the second weight measurement; and
wherein the determining the change in quantity of the first type of item at the particular area of the fixture is based on the second confidence value.

6. The method of claim 4 , wherein the plurality of weight measurements is further associated with one or more of a pick or place of one or more of the first type of item at the particular area of the fixture.

7. The method of claim 4 , wherein the plurality of detected weight changes consists of a predetermined number of detected weight changes.

8. The method of claim 4 , further comprising:
determining a timestamp of each detected weight change of the plurality of detected weight changes; and
wherein the determining the first item weight model for the first type of item at the particular area of the fixture comprises:
applying a first factor to each of the detected weight changes of the first set of detected weight changes, wherein the first factor is based on the timestamp such that the first factor is smaller for detected weight changes having an older timestamp than for detected weight changes having a younger timestamp.

9. The method of claim 4 , wherein the first item weight model for the first type of item at the particular area of the fixture comprises a probability distribution that is a convex combination of the first set of detected weight changes, such that coefficients of the probability distribution are non-negative and sum to 1.

10. The method of claim 4 , wherein the first item weight model for the first type of item at the particular area of the fixture comprises a weighted histogram of the first set of detected weight changes as weighted based on capture time and further wherein each bin of the weighted histogram specifies a weight range.

11. The method of claim 4 , further wherein the first item weight model for the first type of item at the particular area of the fixture produces output data indicative of an integer quantity of items of the first type of item at the particular area of the fixture.

12. A method comprising:
acquiring, from one or more weight sensors of a fixture, a first plurality of weight measurements associated with a first type of item stowed at a particular area of the fixture;
determining, based on the first plurality of weight measurements, a first plurality of detected weight changes indicative of one or more of picks or placements of the first type of item;
determining a first confidence value indicative of a likelihood that individual ones of the first plurality of detected weight changes are representative of a pick or place of a single item;
determining a first set of detected weight changes, the first set of detected weight changes comprising detected weight changes having the first confidence value greater than a first threshold value;
determining, based on the first set of detected weight changes, a first item weight model for the first type of item at the particular area of the fixture, the first item weight model representing a probability distribution of the first set of detected weight changes and corresponding probability values indicative of a likelihood that the first set of detected weight changes are representative of a p

Assignees

• Amazon Tech Inc

Inventors

• Kumar Dilip
• Bo Liefeng
• Chacko Nikhil
• Crandall Robert
• Desai Nishitkumar Ashokkumar
• Eledath Jayakrishnan Kumar
• Gopal Gopi Prashanth
• Medioni Gerard Guy
• Munger Paul Eugene
• Srivastava Kushagra

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06Q10/087 Primary

  Inventory or stock management, e.g. order filling, procurement or balancing against orders · CPC title

• G01G19/42

  for counting by weighing (G01G19/387 takes precedence) · CPC title

• G01G19/40

  with provisions for indicating, recording, or computing price or other quantities dependent on the weight (indicating means for weighing apparatus G01G23/18; recording means for weighing apparatus G01G23/18; computers in general G06) · CPC title

• G01G19/387

  for combinatorial weighing, i.e. selecting a combination of articles whose total weight or number is closest to a desired value · CPC title