Determining item mortality based on tracked environmental conditions during transit

Patent metadata

Field	Value
Publication number	US-12182755-B2
Application number	US-202318330078-A
Country	US
Kind code	B2
Filing date	Jun 6, 2023
Priority date	Aug 3, 2018
Publication date	Dec 31, 2024
Grant date	Dec 31, 2024

Abstract

Item mortality, such as failure rates and expected remaining lifespan, may be determined based on tracked environmental conditions during transit, such as vibration, temperature, and other conditions. For complex items, multiple different internal components may have varying degrees of susceptibility to damage from unfavorable conditions, with each of the components able to individually cause a failure of the entire item. Thus, the specific set of internal components may drive item mortality prediction. Sensors may be located in shipping containers that share data collected throughout the supply chain, using a blockchain, to increase confidence in the integrity of the data. The use of mortality models generated with anonymized component or process data may enable manufacturers to protect trade secrets, such as the specific components or manufacturing processes used, even while leveraging the manufacturer's detailed knowledge of an item.

First claim

What is claimed is:

1. A system for determining item mortality based on tracked environmental conditions during transit, the system comprising:
an outer transport container housing an inner transport container configured to carry a plurality of items in transit;
a first sensor mounted to the outer transport container and configured to gather first sensor data of a first environmental condition experienced by the plurality of items during transit;
a second sensor mounted to the inner transport container and configured to gather second sensor data of a second environmental condition experienced by the plurality of items during transit and different from the first environmental condition;
a processor; and
a computer-readable medium storing instructions that are operative when executed by the processor to:
receive the first sensor data from the first sensor;
receive the second sensor data from the second sensor;
create a neural net training data set by pairing historical data from the first and second sensor data and historical mortality data associated with each of a plurality of components of an item of the plurality of items;
train a neural net using the neural net training data;
generate a tiered mortality model for the item of the plurality of items using the neural net;
generate mortality data for the item by comparing, using the generated tiered mortality model, the first sensor data to a first damage threshold and the second sensor data to a second damage threshold to determine whether a first component of the plurality of components of the item has degraded or failed and whether a second component of the item in the plurality of items has degraded or failed;
based at least on the generated mortality data for the item in the plurality of items and disposition rules, determine an expected remaining lifespan for the item in the plurality of items; and
based at least on the determined expected remaining lifespan for the item in the plurality of items, generate a disposition recommendation for the item in the plurality of items.

2. The system of claim 1 wherein:
receiving at least one of the first sensor data or the second sensor data comprises receiving sensor data from a blockchain; and
the tiered mortality model is a composite model that is tiered according to a probabilistic failure rate of the plurality of components of the item.

3. The system of claim 1 wherein:
the system further comprises a third transport container housing the outer transport container; and
the first sensor data and the second sensor data each comprise measurements from at least one sensor selected from a list comprising: temperature, humidity, air pressure, condensation, salinity, CO 2 level, time, shock, and vibration.

4. The system of claim 1 wherein the disposition rules specify exposure conditions or a threshold of the mortality indicating acceptance, acceptance but at a reduced price, deferral to a secondary inventory market, or rejection of the item of the plurality of items.

5. The system of claim 1 , further comprising a container that stores the plurality of items during transit, wherein the container includes a lock unit, the lock unit being blockchain-enabled and configured to generate a blockchain entry upon an event involving the lock unit.

6. The system of claim 1 wherein the historical mortality data comprises:
historical failure rates and descriptions, historical repair rates and descriptions, and historical warranty claims.

7. The system of claim 1 wherein the instructions are further operative to:
generate the tiered mortality model to predict a failure rate and an expected remaining lifespan of the first component of the item and a failure rate and an expected remaining lifespan of the second component of the item using the neural net,
wherein the tiered mortality model is based at least on a combination of degradation of the first component of the item and degradation of the second component of the item.

8. The system of claim 7 wherein the generated mortality data is probabilistic and indicates a percentage chance of failure for the item in the plurality of items.

9. The system of claim 7 wherein the instructions are further operative to:
further train the neural net trained with the received first sensor data and the received second sensor data.

10. A method for determining item mortality by a processor based on tracked environmental conditions during transit, the method comprising:
receiving, by the processor, first sensor data from a first sensor for a first environmental condition experienced by a plurality of items during transit, the first sensor mounted to an outer transport container;
receiving, by the processor, second sensor data for a second environmental condition experienced by the plurality of items during transit and different from the first environmental condition, the second sensor mounted to an inner transport container housed within the outer transport container;
creating, by the processor, a neural net training data set by pairing historical data from the first and second sensor data and historical mortality data associated with each of a plurality of components of an item of the plurality of items;
training, by the processor, a neural net using the neural net training data;

generating, by the processor, a tiered mortality model for the item of the plurality of items using the neural net;
generating, by the processor, mortality data for the item by comparing, using the generated tiered mortality model, the first sensor data to a first damage threshold and the second sensor data to a second damage threshold to determine whether a first component of the item in the plurality of items has degraded or failed and whether a second component of the item in the plurality of items has degraded or failed;
based at least on the generated mortality data for the item in the plurality of items and disposition rules, determining, by the processor, an expected remaining lifespan for the item in the plurality of items; and
based at least on the determined expected remaining lifespan for the item in the plurality of items, generating, by the processor, a disposition recommendation for the item in the plurality of items.

11. The method of claim 10 , wherein receiving at least one of the first sensor data or the second sensor data further comprises receiving sensor data from a blockchain.

12. The method of claim 10 , wherein the first sensor data and the second sensor data each comprise measurements from at least one sensor selected from a list comprising: temperature, humidity, air pressure, condensation, salinity, CO 2 level, time, shock, and vibration.

13. The method of claim 10 , wherein the disposition rules specify exposure conditions or a threshold of the mortality indicating acceptance, acceptance but at a reduced price, deferral to a secondary inventory market, or rejection of the item of the plurality of items.

14. The method of claim 10 , further comprising storing the plurality of items in a container during transit, wherein the container includes a lock unit, the lock unit being blockchain-enabled and configured to generate a blockchain entry upon an event involving the lock unit.

15. The method of claim 10 , wherein the historical mortality data comprises:
historical failure rates and descriptions, historical repair rates and descriptions, and historical warranty claims.

16. The method of claim 10 , further comprising:
generating the tiered mortality model to predict a failure rate and an expected remaining lifespan of the fir

Assignees

• Walmart Apollo Llc

Inventors

• Jurich Jr Joseph
• Schuhardt David Justin

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0499

  Feedforward networks · CPC title

• B65D88/00

  Large containers · CPC title

• H04W4/38

  for collecting sensor information · CPC title

• G06Q30/0206

  Price or cost determination based on market factors · CPC title