IP Library › Granted Patent US 12,638,543
Granted Patent B2
US 12,638,543 · App. 18/117,181 · Granted May 26, 2026

General and robust distributed linear filtering and prediction with optimal gain

Inventor: Subhro Das (Cambridge, MA)
Assignee: International Business Machines Corporation
G01S5/0294G01S5/021
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 12,638,543
App. No.
18/117,181
Granted
May 26, 2026
Kind
B2
Abstract

In one aspect of the invention, there is a computer-implemented method including: detecting, by a processor set of a first sensor agent, sensor data from one or more sensors comprised in the first sensor agent; determining, by the processor set, an own series of estimates, based on the sensor data; transmitting, by the processor set, the own series of estimates; receiving, by the processor set, at least one additional series of estimates from additional sensor agents; restoring, by the processor set, in response to detecting that a second sensor agent of the additional sensor agents has become disconnected and then re-connected, the transmitting of the series of estimates to the second sensor agent and the receiving of the series of estimates from the second sensor agent; and outputting, by the processor set, based on the own series of estimates and the additional series of estimates, a series of consensus estimates.

Claims (55)

1 . A method, comprising:

detecting, by a processor set of a first sensor agent in a distributed processing system for making predictions about a random dynamical system, sensor data from one or more sensors comprised in the first sensor agent;

determining, by the processor set, an own series of estimates associated with the random dynamical system using a distributed estimation algorithm, based on the sensor data;

transmitting, by the processor set, the own series of estimates;

receiving, by the processor set, at least one additional series of estimates associated with the random dynamical system from one or more additional sensor agents in the distributed processing system;

restoring, by the processor set, in response to detecting that a second sensor agent of the one or more additional sensor agents has become disconnected and then re-connected, the transmitting of the own series of estimates to the second sensor agent and the receiving of the at least one additional series of estimates from the second sensor agent;

re-positioning, by the processor set, the first sensor agent in response to detecting that the second sensor agent of the one or more additional sensor agents has become disconnected; and

outputting, by the processor set, based on the own series of estimates and the additional series of estimates, a series of consensus estimates of the random dynamical system.

2 . The method of claim 1 , further comprising relaying a transmission of the at least one series of estimates from the one or more additional sensor agents.

3 . The method of claim 1 , wherein the re-positioning the first sensor agent comprises re-positioning the first sensor agent toward a last detected position of the second sensor agent.

4 . The method of claim 3 , wherein the re-positioning the first sensor agent further comprises re-positioning the first sensor agent while remaining within connectivity of at least one other sensor agent of the one or more additional sensor agents.

5 . The method of claim 1 , further comprising, in response to the detecting that the second sensor agent of the one or more additional sensor agents has become re-connected, re-positioning the first sensor agent.

6 . The method of claim 5 , wherein the re-positioning the first sensor agent comprises re-positioning the first sensor agent away from a detected present position of the second sensor agent.

7 . The method of claim 5 , wherein the re-positioning the first sensor agent comprises re-positioning the first sensor agent to a position that increases a local homogeneity of distribution of the first sensor agent, the second sensor agent, and the one or more additional sensor agents.

8 . The method of claim 1 , wherein the own series of estimates comprises a time series of filter estimates of observed states, based on the sensor data, and a time series of prediction estimates of predicted future states, based on the sensor data and on predictive modeling.

9 . The method of claim 8 , wherein the determining the own series of estimates comprises performing a distributed estimation algorithm to generate prediction updates and filtering updates.

10 . The method of claim 9 , wherein the determining the own series of estimates further comprises performing error analysis, wherein performing error analysis comprises:

determining a predictor error and a filter error for the first sensor agent;

determining local innovations and local innovation noise for the first sensor agent; and

performing recursive updates of the evolution of covariances, wherein a spectral radius of a dynamics matrix of error is less than 1,

wherein the first sensor agent and the one or more additional sensor agents form a distributed estimator that converges to consensus estimates,

wherein the consensus estimates have bounded mean-squared error (MSE) and the filter error is asymptotically stable.

11 . The method of claim 1 , wherein determining the own series of estimates further comprises performing an optimal gain design, wherein performing the optimal gain design comprises applying Gauss-Markov theorem, thereby generating gain matrices.

12 . The method of claim 11 , wherein the performing an optimal gain design further comprises:

using an optimal distributed filter that performs optimal fusion of estimates under unknown correlations by a Semidefinite Programming (SDP) relaxation, which is pre-computed and stored on the first sensor agent; and

determining a recursive iteration of a filter error covariance matrix which comprises a distributed version of a discrete algebraic Riccati equation for the distributed estimation algorithm.

13 . The method of claim 1 , further comprising using the outputted series of consensus estimates for at least one application selected from among the group of: multi-agent control, positioning, navigation, state estimation in electrical power grid infrastructure, spatio-temporal environment monitoring, spatio-temporal field monitoring, connected vehicular network for traffic balancing and collision avoidance, wildlife monitoring, and collaborative object tracking.

14 . A computer program product comprising one or more computer readable storage media having program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to:

detect sensor data from one or more sensors comprised in a first sensor agent in a distributed processing system for making predictions about a random dynamical system;

determine an own series of estimates associated with the random dynamical system using a distributed estimation algorithm, based on the sensor data;

transmit the own series of estimates;

receive at least one additional series of estimates associated with the random dynamical system from one or more additional sensor agents in the distributed processing system;

restore, in response to detecting that a second sensor agent of the one or more additional sensor agents has become disconnected and then re-connected, the transmitting of the own series of estimates to the second sensor agent and the receiving of the at least one additional series of estimates from the second sensor agent;

re-position the first sensor agent, in response to detecting that the second sensor agent of the one or more additional sensor agents has become disconnected, wherein the re-positioning the first sensor agent comprises re-positioning the first sensor agent toward a last detected position of the second sensor agent, while remaining within connectivity of at least one other sensor agent of the one or more additional sensor agents; and

output, based on the own series of estimates and the additional series of estimates, a series of consensus estimates of the random dynamical system.

15 . The computer program product of claim 14 , wherein the program instructions are further executable to:

relay a transmission of the at least one series of estimates from the one or more additional sensor agents;

determine, based on the own series of estimates and the additional time series of estimates, a series of consensus estimates.

16 . The computer program product of claim 14 , wherein the program instructions are further executable to:

in response to the detecting that the second sensor agent of the one or more additional sensor agents has become re-connected, re-positioning the first sensor agent, wherein the re-positioning the first sensor agent comprises re-positioning the first sensor agent away from a detected present position of the second sensor agent, to a position that increases a local homogeneity of distribution of the first sensor agent, the second sensor agent, and the one or more additional sensor agents.

17 . A system comprising:

a processor set, one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable to:

detect sensor data from one or more sensors comprised in a first sensor agent in a distributed processing system for making predictions about a random dynamical system;

determine an own series of estimates associated with the random dynamical system using a distributed estimation algorithm, based on the sensor data;

transmit the own series of estimates;

receive at least one additional series of estimates associated with the random dynamical system from one or more additional sensor agents in the distributed processing system;

restore, in response to detecting that a second sensor agent of the one or more additional sensor agents has become disconnected and then re-connected, the transmitting of the series of own estimates to the second sensor agent and the receiving of the at least one additional series of estimates from the second sensor agent;

re-position the first sensor agent, in response to detecting that the second sensor agent of the one or more additional sensor agents has become disconnected; and

output, based on the own series of estimates and the additional series of estimates, a series of consensus estimates of the random dynamical system.

18 . The system of claim 17 , wherein the program instructions are further executable to:

relay a transmission of the at least one series of estimates from the one or more additional sensor agents;

determine, based on the own series of estimates and the additional time series of estimates, a series of consensus estimates.

19 . The system of claim 17 , wherein the program instructions are further executable to:

in response to the detecting that the second sensor agent of the one or more additional sensor agents has become re-connected, re-positioning the first sensor agent, wherein the re-positioning the first sensor agent comprises re-positioning the first sensor agent away from a detected present position of the second sensor agent, to a position that increases a local homogeneity of distribution of the first sensor agent, the second sensor agent, and the one or more additional sensor agents.

20 . The system of claim 17 , wherein the re-positioning the first sensor agent comprises re-positioning the first sensor agent toward a last detected position of the second sensor agent, while remaining within connectivity of at least one other sensor agent of the one or more additional sensor agents.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Mar 3, 2023
From: DAS, SABHRO
To: INTERNATIONAL BUSINESS MACHINES CORPORATION
Reel/Frame 062877/0719 →
Continuity (1)
Related Publication 20240302486A1 · Sep 12, 2024
References Cited (51)
US 9633315B2 · Chapelle et al. · 2017 [cited by applicant]
US 9984337B2 · Kadav et al. · 2018 [cited by applicant]
US 11252260B2 · Xie et al. · 2022 [cited by applicant]
US 11455573B2 · Das · 2022 [cited by applicant]
US 20210092571A1 · Shibuya · 2021 [cited by examiner]
US 20210160042A1 · Sugata · 2021 [cited by examiner]
US 20220085975A1 · Manamohan et al. · 2022 [cited by applicant]
US 20220101480A1 · Kalamkar et al. · 2022 [cited by applicant]
CN 113010305 · 2021 [cited by applicant]
WO 201758348 · 2017 [cited by applicant]
WO WO2023097022A1 · 2023 [cited by examiner]
Das et al., “Consensus+Innovations Distributed Kalman Filter With Optimized Gains”, IEEE Transactions On Signal Processing, vol. 65, No. 2, Jan. 15, 2017; pp. 467-481. [cited by applicant]
McMahan et al., “Federated Learning: Collaborative Machine Learning without Centralized Training Data”, https://ai.googleblog.com/2017/04/federated-learning-collaborative.html, Apr. 6, 2017; 6 Pages. [cited by applicant]
Das, Grace Period Disclosure: “On Observability and Optimal Gain Design for Distributed Linear Filtering and Prediction”, https://arxiv.org/pdf/2203.03521.pdf, Mar. 7, 2022; 8 Pages. [cited by applicant]
Das, “On Observability and Optimal Gain Design for Distributed Linear Filtering and Prediction”, Proceedings of EUSIPCO 2022; Downloaded Jan. 9, 2023; pp. 1846-1850. [cited by applicant]
Carli et al., “Distributed Kalman filtering based on consensus strategies”, IEEE Journal on Selected Areas in Communications, Apr. 25, 2008, pp. 622-633. [cited by applicant]
Casbeer et al., “Distributed information filtering using consensus filters”, 2009 American Control Conference, Jun. 10-12, 2009, pp. 1882-1887. [cited by applicant]
Cattivelli et al., “Diffusion Strategies for Distributed Kalman Filtering and Smoothing”, IEEE Transactions on Automatic Control, Feb. 17, 2010, pp. 2069-2084. [cited by applicant]
Chee-Yee Chong, “Forty years of distributed estimation: A review of noteworthy developments”, 2017 Sensor Data Fusion: Trends, Solutions, Applications (SDF), Oct. 2017, 10 pages. [cited by applicant]
Das et al., “Distributed Estimation of Dynamic Fields Over Multi-agent Networks”, arXiv:1701.02710v1 [cs.IT], Jan. 10, 2017, 2 pages. [cited by applicant]
Das et al., “Distributed Kalman filtering and network tracking capacity”, 2013 Asilomar Conference on Signals, Systems and Computers, Nov. 2013. [cited by applicant]
Das et al., “Distributed Kalman Filtering With Dynamic Observations Consensus”, IEEE Transactions on Signal Processing, Apr. 17, 2015, pp. 4458-4473. [cited by applicant]
Das et al., “Distributed Kalman filtering”, 21st European Signal Processing Conference (EUSIPCO 2013), Sep. 2013, 5 pages. [cited by applicant]
Das et al., “Distributed linear estimationof dynamic random fileds”, 2013 51st Annual Allerton Conference on Communication, Control, and Computing (Allerton), Oct. 2013, pp. 1120-1125. [cited by applicant]
Das et al., “Distributed linear filtering and prediction of time-varying random fields”, Carnegie Mellon University, Thesis, Jun. 23, 2016, 132 pages. [cited by applicant]
Das et al., “Distributed state estimation in multi-agent networks”, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, May 2013, pp. 4246-4250. [cited by applicant]
Dimakis et al., “Gossip Algorithms for Distributed Signal Processing”, arXiv:1003.5309v1 [cs.DC], Mar. 27, 2010, 29 pages. [cited by applicant]
Ding et al., “A Survey on Model-Based Distributed Control and Filtering for Industrial Cyber-Physical Systems”, IEEE Transactions on Industrial Informatics, Mar. 15, 2019, pp. 2483-2499. [cited by applicant]
Doostmohammadian et al., “On the Genericity Properties in Distributed Estimation: Topology Design and Sensor Placement”, IEEE Journal of Selected Topics in Signal Processing, Feb. 8, 2013, pp. 195-204. [cited by applicant]
Fan RK Chung, “Lectures on Spectral Graph Theory”, https://mathweb.ucsd.edu/˜fan/research/cbms.pdf, 1997, 25 pages. [cited by applicant]
Ghai et al., “A Regret Minimization Approach to Multi-Agent Control”, arXiv:2201.13288v3 [math.OC], Feb. 25, 2022, 13 pages. [cited by applicant]
He et al., “Distributed Design of Robust Kalman Filters Over Corrupted Channels”, IEEE Transactions on Signal Processing, Apr. 5, 2021, pp. 2422-2434. [cited by applicant]
Howard et al., “An Optimal Kalman-Consensus Filter for Distributed Implementation Over a Dynamic Communication Network”, IEEE Access , May 3, 2021, pp. 66696-66706. [cited by applicant]
Kar et al., “Gossip and distributed Kalman filtering: Weak consensus under weak detectability”, IEEE Transactions on Signal Processing, Apr. 1, 2011, pp. 1766-1784. [cited by applicant]
Khan et al., “Coordinated networked estimation strategies using structured systems theory”, 2011 50th IEEE Conference on Decision and Control and European Control Conference, Dec. 2011, pp. 2112-2117. [cited by applicant]
Khan et al., “Distributing the Kalman Filter for Large-Scale Systems”, IEEE Transactions on Signal Processing, arXiv:0708.0242v2 [cs.IT], Feb. 25, 2008, 30 pages. [cited by applicant]
Khan et al., “Networked estimation under information constraints”, arXiv:1111.4580v1 [cs.IT], Nov. 19, 2011, 32 pages. [cited by applicant]
Li et al., “Cooperative Field Prediction and Smoothing via Covariance Intersection”, IEEE Transactions on Signal Processing, Jan. 8, 2021, pp. 797-808. [cited by applicant]
Li et al., “Distributed Kalman Filtering Over Massive Data Sets: Analysis Through Large Deviations of Random Riccati Equations”, IEEE Transactions on Information Theory, Jan. 9, 2015, pp. 1351-1372. [cited by applicant]
Park et al., “An augmented observer for the distributed estimation problem for LTI systems”, 2012 American Control Conference (ACC), Jun. 2012, pp. 6775-6780. [cited by applicant]
R. E. Kalman, “New results in linear filtering and prediction theory”, https://asmedigitalcollection.asme.org/fluidsengineering/article-abstract/83/1/95/426820/New-Results-in-Linear-Filtering-and-Prediction?redirectedFr… [cited by applicant]
Reza Olfati-Saber, “Distributed Kalman Filter with Embedded Consensus Filters”, Proceedings of the 44th IEEE Conference on Decision and Control, Dec. 2005, 6 pages. [cited by applicant]
Reza Olfati-Saber, “Distributed Kalman filtering for sensor networks”, 2007 46th IEEE Conference on Decision and Control, Dec. 2007, pp. 5492-5498. [cited by applicant]
Reza Olfati-Saber, “Kalman-Consensus Filter : Optimality, stability, and performance”, Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, Dec. 20… [cited by applicant]
Ribeiro et al., “Kalman Filtering in Wireless Sensor Networks”, IEEE Control Systems Magazine, Mar. 25, 2010, pp. 66-86. [cited by applicant]
Rudolph Emil Kalman, “A New Approach to Linear Filtering and Prediction Problems”, https://asmedigitalcollection.asme.org/fluidsengineering/article-abstract/82/1/35/397706/A-New-Approach-to-Linear-Filtering-and-Predicti… [cited by applicant]
Savas et al., “On distributed linear filtering with noisy communication”, 2017 American Control Conference (ACC), May 2017, pp. 2699-2704. [cited by applicant]
Schizas, “Consensus in ad hoc WSNs with noisy links—Part II: Distributed estimation and smoothing of random signals”, IEEE Transactions on Signal Processing, Mar. 14, 2008, pp. 1650-1666. [cited by applicant]
Sebastian et al., “All-in-one: Certifiable Optimal Distributed Kalman Filter under Unknown Correlations”, arXiv:2105.15061v3 [eess.SY], Jun. 13, 2022, 7 pages. [cited by applicant]
Yan et al., “Feedback-based target localization in underwater sensor networks: A multisensor fusion approach”, IEEE Transactions on Signal and Information Processing over Network, Aug. 19, 2018, pp. 168-180. [cited by applicant]
Zhang et al., “Distributed filtering for switched linear systems with sensor networks in presence of packet dropouts and quantization”, IEEE Transactions on Circuits and Systems I: Regular Papers, May 6, 2017, pp. 2783-… [cited by applicant]