IP Library Granted Patent US 12676789
Granted Patent B2
US 12676789 · App. 18/941,966 · Granted Jul 7, 2026

Self-adaptive health monitoring systems including networks of tensor networks

Inventors: David Singer (Ann Arbor, MI); Alexander Manohar (Ann Arbor, MI); Connor Arrigan (Ann Arbor, MI)
Assignee: The Regents of The University of Michigan
H04L41/08H04L43/10H04L67/12
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12676789
App. No.
18/941,966
Granted
Jul 7, 2026
Kind
B2
Abstract

A computer system includes memory hardware configured to store computer-executable instructions, and a network of multiple tensor networks, each tensor network including multiple nodes. The computer system includes processor hardware configured to execute the computer-executable instructions to transmit a ping request from a first tensor network of the multiple tensor networks to a second tensor network of the multiple tensor networks, contract multiple nodes of the second tensor network in response to the ping request, to generate a probability distribution indicative of a state information of the second tensor network, transmit the probability distribution from the second tensor network to the first tensor network, and update the first tensor network to connect the probability distribution with at least one of multiple nodes of the first tensor network.

Claims (52)

1 . A computer system comprising:

memory hardware configured to store computer-executable instructions, and a network of multiple tensor networks, each tensor network including multiple nodes; and

processor hardware configured to execute the computer-executable instructions to:

transmit a ping request from a first tensor network of the multiple tensor networks to a second tensor network of the multiple tensor networks;

contract multiple nodes of the second tensor network in response to the ping request, to generate a probability distribution indicative of a state information of the second tensor network;

transmit the probability distribution from the second tensor network to the first tensor network; and

update the first tensor network to connect the probability distribution with at least one of multiple nodes of the first tensor network.

2 . The computer system of claim 1 , wherein the processor hardware is configured to contract the multiple nodes of the first tensor network and the probability distribution, to generate a contraction result indicative of a state information of the first tensor network based on contextual information of the second tensor network.

3 . The computer system of claim 2 , wherein the processor hardware is configured to:

compare the contraction result to a specified nominal range associated with the first tensor network; and

identify a failure condition in response to the contraction result being outside of the specified nominal range associated with the first tensor network.

4 . The computer system of claim 1 , wherein:

the first tensor network includes at least one first physical component node and multiple sensor nodes connected with the at least one first physical component node; and

the second tensor network includes at least one second physical component node and multiple sensor nodes connected with the at least one second physical component node.

5 . The computer system of claim 4 , wherein at least one anchor is connected with at least one of the multiple sensor nodes of the first tensor network.

6 . The computer system of claim 4 , wherein at least one external leg is connected with at least one of the multiple sensor nodes of the first tensor network.

7 . The computer system of claim 4 , wherein:

the first physical component node is a pump component; and

the second physical component node is a pipe component.

8 . The computer system of claim 7 , wherein:

the multiple nodes of the first tensor network include a pump current sensor node and a pump pressure sensor node; and

the multiple nodes of the second tensor network include a pipe pressure sensor node and a pipe flowrate sensor node.

9 . The computer system of claim 8 , wherein:

the processor hardware is configured to identify a failure condition according to the probability distribution; and

the failure condition includes at least one of a leak condition and a clog condition.

10 . The computer system of claim 8 , wherein:

the pump component and the pipe component are components of an autonomous ship vessel; and

the network of tensor networks is configured to facilitate self-adaptive health monitoring for the autonomous ship vessel.

11 . A method for executing a network of tensor networks, the method comprising:

transmitting a ping request from a first tensor network to a second tensor network, wherein the first tensor network and the second tensor network belong to a network of multiple tensor networks, and each of the multiple tensor networks includes multiple nodes;

contracting multiple nodes of the second tensor network in response to the ping request, to generate a probability distribution indicative of a state information of the second tensor network;

transmitting the probability distribution from the second tensor network to the first tensor network; and

updating the first tensor network to connect the probability distribution with at least one of multiple nodes of the first tensor network.

12 . The method of claim 11 , further comprising contracting the multiple nodes of the first tensor network and the probability distribution, to generate a contraction result indicative of a state information of the first tensor network based on contextual information of the second tensor network.

13 . The method of claim 12 , further comprising:

comparing the contraction result to a specified nominal range associated with the first tensor network; and

identifying a failure condition in response to the contraction result being outside of the specified nominal range associated with the first tensor network.

14 . The method of claim 11 , wherein:

the first tensor network includes at least one first physical component node and multiple sensor nodes connected with the at least one first physical component node; and

the second tensor network includes at least one second physical component node and multiple sensor nodes connected with the at least one second physical component node.

15 . The method of claim 14 , wherein at least one anchor is connected with at least one of the multiple sensor nodes of the first tensor network.

16 . The method of claim 14 , wherein at least one external leg is connected with at least one of the multiple sensor nodes of the first tensor network.

17 . The method of claim 14 , wherein:

the first physical component node is a pump component; and

the second physical component node is a pipe component.

18 . The method of claim 17 , wherein:

the multiple nodes of the first tensor network include a pump current sensor node and a pump pressure sensor node; and

the multiple nodes of the second tensor network include a pipe pressure sensor node and a pipe flowrate sensor node.

19 . The method of claim 18 , further comprising identifying a failure condition according to the probability distribution, wherein the failure condition includes at least one of a leak condition and a clog condition.

20 . The method of claim 18 , wherein:

the pump component and the pipe component are components of an autonomous ship vessel; and

the network of tensor networks is configured to facilitate self-adaptive health monitoring for the autonomous ship vessel.