CN103914069A - Multi-robot cooperation locating method capable of allowing cognitive errors - Google Patents
Multi-robot cooperation locating method capable of allowing cognitive errors Download PDFInfo
- Publication number
- CN103914069A CN103914069A CN201410095912.1A CN201410095912A CN103914069A CN 103914069 A CN103914069 A CN 103914069A CN 201410095912 A CN201410095912 A CN 201410095912A CN 103914069 A CN103914069 A CN 103914069A
- Authority
- CN
- China
- Prior art keywords
- robot
- particle
- machine people
- distribution
- weights
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Manipulator (AREA)
Abstract
The invention discloses a multi-robot cooperation locating method capable of allowing cognitive errors. The method comprises the following steps that firstly, one sending robot Ri selects particles needing to be exchanged according to weights of the particles; secondly, the sending robot Ri sends particle sets based on relative positioning calculation to a neighbor robot Rj; thirdly, the sending robot Ri receives the particle sets based the relative positioning calculation from the neighbor robot Rj; fourthly, the sending robot Ri updates the particle sets of itself; fifthly, the sending robot Ri calculates the weights of particle-concentration particles and conducts normalization processing; sixthly, the sending robot Ri calculates posterior probability distribution of the robot pose according to the particle-concentration particles and the weights of the particle-concentration particles. Even if the neighbor robot is cognized mistakenly, definite locating accuracy can be guaranteed due to the strategy that each robot keeps a certain number of particles in the particle sets rather than exchanging all the particles.
Description
Technical field
The present invention relates to a kind of localization method, particularly relate to a kind of multi-robot Cooperation localization method that can tolerate cognitive mistake.
Background technology
Due to the limitation of single robot observation, its positioning precision is relatively low.Cooperate and utilize neighbours robot relative position information to improve the positioning precision of oneself by multiple mobile robot.The co-positioned of multirobot has been subject to more and more researchists' attention in recent years.At any time, individual machine people only can be to the part observation of environment, and it is observedly usually subject to extraneous interference and produces stochastic error to environmental information.Robot is by merging the observation information of other robot, and then improves accuracy to environmental observation and comprehensive.And robot can, by merging the relative positioning information of neighbours robot, locate thereby realize more accurately.
In multi-robot system, multirobot location can be applied simple single robot localization method and be realized.But in some cases, robot equipment sensor and communication facilities can be regarded mobile road sign as other robot, and can know the relative position of neighbours robot by sensor.Thereby robot can utilize cooperation to improve the positioning precision of oneself.Utilize Kalman filtering can realize the co-positioned between multirobot.Because of Kalman filter demands noise Gaussian distributed, and particle filter does not generally have this requirement, therefore it is widely used in multi-robot Cooperation positioning field.
The literature search of prior art is found, the multi-robot Cooperation location based on particle filter all supposes that cognition is correct to other robot in robot.Fox etc. 2000 are at Autonomous Robots(autonomous robot magazine) deliver the multi-robot Cooperation localization method based on particle filter proposing in the literary composition of " Collaborative multi-robot localization " (multi-Robot Cooperative location) on magazine, in their method, robot one finds other robot, utilizes the positional information of another robot to realize co-positioned.Gasparri etc. are at meeting 2008Workshop on Formal Models and Methods for Multi-Robot Systems, the multirobot particle filter co-positioned of an A fast conjunctive resampling particle filter for collaborative multi-robot localization(quick associating resampling of delivering in (about the model of multi-robot system and the international conference of method)) Cooperative Localization Method that proposes in literary composition, in their method, but the weights of particle have exceeded threshold value, between robot, just exchange particle, to realize location more accurately.In above localization method, all do not consider the impact on location on the cognition mistake of other robot.
Summary of the invention
The technical matters existing for the existing multi-robot Cooperation location algorithm based on particle filter, the present invention proposes a kind of multi-robot Cooperation localization method that can tolerate cognitive mistake.
The present invention solves above-mentioned technical matters by following technical proposals: a kind of multi-robot Cooperation localization method that can tolerate cognitive mistake, it is characterized in that, and it comprises the following steps:
Step 1, a certain distribution of machine people Ri need to exchange particle according to the Weight selected of particle;
Step 2, a certain distribution of machine people Ri sends the Rj of reception robot to neighbours according to the particle collection of relative positioning calculating;
Step 3, a certain distribution of machine people Ri receives the particle collection calculating according to relative positioning from receiving the Rj of robot;
Step 4, a certain distribution of machine people Ri upgrades the particle collection of oneself;
Step 5, a certain distribution of machine people Ri calculates the weights of particle set particle, and does normalized;
Step 6, a certain distribution of machine people Ri calculates the posterior probability distribution of robot pose according to the particle of particle set and weights thereof.
Preferably, described step 1 comprises the following steps: step 11, distribution of machine people Ri has M particle
represent the posterior probability distribution p (x of robot pose
i,t| u
i,t, z
i,t), 20<M<100000, wherein x
i,t=(x
i, y
j, θ
i), u
i,tfor controlled quentity controlled variable, z
i,tfor observed quantity;
represent the prior probability distribution p (x of robot pose
i,t| u
i,t, z
i, t-1), wherein, particle
weights be
?
for the observation model of robot; Step 12, distribution of machine people Ri calculates and receives the relative pose of the Rj of robot with respect to robot oneself, and it is expressed as s
ij=(x
ij, y
ij, θ
ij); Distribution of machine people Ri is according to the weights of particle
from particle collection
the individual particle of middle selection (1-p0) M/ (N-1) represents probability distribution p (x
i,t| u
i,t, z
i,t), wherein 0<p0<1, N-1 is the number of distribution of machine people Ri neighbours robot; For the particle of each selection
have
particle
set
represent probability distribution p (x
j,t| u
i,t, z
i,t, s
ij, t).
Preferably, a certain distribution of machine people Ri of described step 2 sends particle assembly
give neighbours' the Rj of reception robot.
Preferably, a certain distribution of machine people Ri of described step 3 receives from receiving the Rj of robot particle assembly
p*M particle is retained in the particle assembly of distribution of machine people Ri oneself and is represented as Qi, and P is the probability of selected exchange particle, 0<P<1.
Preferably, the particle assembly of a certain distribution of machine people Ri renewal oneself of described step 4 is:
Preferably, an a certain distribution of machine people Ri calculating grain subset X for described step 5 '
i,tthe weights of middle particle, and do normalized.
Preferably, a certain distribution of machine people Ri of described step 6 according to grain subset X '
i,tand weights calculate the posterior probability distribution of robot pose.
Positive progressive effect of the present invention is: it is evaluated twice that the present invention makes to exchange the weights of particle, is once sent out robot assessment, another received robot assessment, thus improve the positioning precision of robot.Tolerance to cognitive wrong: even if occur the cognitive mistake of other robot, the particle of robot by retaining certain number is in the particle assembly of oneself, rather than the strategy that exchanges whole particles has guaranteed certain location accuracy.Modularization: the mode of operation of each robot is identical, has merged oneself locating information and relative positioning information, and fusion results is sent to neighbours robot.
Embodiment
Below in conjunction with specific embodiment, the present invention is described in detail.Following examples will contribute to those skilled in the art further to understand the present invention, but not limit in any form the present invention.It should be pointed out that to those skilled in the art, without departing from the inventive concept of the premise, can also make some distortion and improvement.These all belong to protection scope of the present invention.
With two robot cooperated orientate as example the present invention will be described in detail, the R1 of robot and the R2 of robot are in the scope of relative positioning, wherein R1 is distribution of machine people, R2 for receive robot:
Step 1: distribution of machine people R1 calculates the particle that needs exchange, and concrete comprises following calculating and communicate by letter:
1) distribution of machine people R1 has 150 particles
represent the posterior probability distribution p (x of robot pose
1, t| u
1, t, z
1, t), wherein x
1, t=(x
1, y
1, θ
i), u
1, tfor controlled quentity controlled variable, z
1, tfor observed quantity;
represent the prior probability distribution p (x of robot pose
1, t| u
1, t, z
1, t-1), wherein, particle
weights be
?
for the observation model of robot; Distribution of machine people R1 calculates and receives the relative pose of the R2 of robot with respect to robot oneself, and it is expressed as s
12=(x
12, y
12, θ
12).
2) distribution of machine people R1 is according to the weights of particle
from particle collection
the individual particle of middle selection (1-p0) M/ (N-1) represents probability distribution p (x
1, t| u
1, t, z
1, t), wherein M=150, p0=0.15, N-1=1 is the number of distribution of machine people R1 neighbours robot.For the particle of each selection
particle
set
represent probability distribution p (x
2, t| u
1, t, z
1,t,s
12, t).
Step 2: distribution of machine people R1 sends particle assembly
give neighbours' the R2 of reception robot.
Step 3: distribution of machine people R1 receives from receiving the R2 of robot particle assembly
22 particles of P*M=0.15*150 ≈ are retained in the particle assembly of distribution of machine people R1 oneself and are represented as Q1.P is the probability of selected exchange particle, 0<P<1, and its value is determined by experience.
Step 4: distribution of machine people R1 upgrade oneself particle assembly:
Step 5: a distribution of machine people R1 calculating grain subset X '
1, tthe weights of middle particle, and do normalized.
Step 6: distribution of machine people R1 according to grain subset X '
1, tand weights calculate the posterior probability distribution of robot pose.
The present invention is directed to the robot that disposes locating device, can intercom mutually, robot realizes the co-positioned based on particle filter by merging neighbours robot relative observation.Robot calculates the particle that needs exchange according to the weights of particle, retain certain number in the particle assembly of oneself, receives the particle calculating from neighbours robot simultaneously.Even if there is the cognitive mistake to neighbours robot, robot is by retaining certain number in the particle assembly of oneself, rather than the strategy that exchanges whole particles has guaranteed certain location accuracy.
Above specific embodiments of the invention are described.It will be appreciated that, the present invention is not limited to above-mentioned specific implementations, and those skilled in the art can make various distortion or modification within the scope of the claims, and this does not affect flesh and blood of the present invention.
Claims (7)
1. can tolerate a cognitive wrong multi-robot Cooperation localization method, it is characterized in that, it comprises the following steps:
Step 1, a certain distribution of machine people Ri need to exchange particle according to the Weight selected of particle;
Step 2, a certain distribution of machine people Ri sends the Rj of reception robot to neighbours according to the particle collection of relative positioning calculating;
Step 3, a certain distribution of machine people Ri receives the particle collection calculating according to relative positioning from receiving the Rj of robot;
Step 4, a certain distribution of machine people Ri upgrades the particle collection of oneself;
Step 5, a certain distribution of machine people Ri calculates the weights of particle set particle, and does normalized;
Step 6, a certain distribution of machine people Ri calculates the posterior probability distribution of robot pose according to the particle of particle set and weights thereof.
2. the multi-robot Cooperation localization method that can tolerate cognitive mistake as claimed in claim 1, is characterized in that, described step 1 comprises the following steps: step 11, distribution of machine people Ri has M particle
represent the posterior probability distribution p (x of robot pose
i,t| u
i,t, z
i,t), 20<M<100000, wherein x
i,t=(x
i, y
j, θ
i), u
i,tfor controlled quentity controlled variable, z
i,tfor observed quantity;
represent the prior probability distribution p (x of robot pose
i,t| u
i,t, z
i, t-1), wherein, particle
weights be
?
for the observation model of robot; Step 12, distribution of machine people Ri calculates and receives the relative pose of the Rj of robot with respect to robot oneself, and it is expressed as s
ij=(x
ij, y
ij, θ
ij); Distribution of machine people Ri is according to the weights of particle
from particle collection
the individual particle of middle selection (1-p0) M/ (N-1) represents probability distribution p (x
i,t| u
i,t, z
i,t), wherein 0<p0<1, N-1 is the number of distribution of machine people Ri neighbours robot; For the particle of each selection
have
particle
set
represent probability distribution p (x
j,t| u
i,t, z
i,t, s
ij, t).
3. the multi-robot Cooperation localization method that can tolerate cognitive mistake as claimed in claim 1, is characterized in that, a certain distribution of machine people Ri of described step 2 sends particle assembly
give neighbours' the Rj of reception robot.
4. the multi-robot Cooperation localization method that can tolerate cognitive mistake as claimed in claim 1, is characterized in that, a certain distribution of machine people Ri of described step 3 receives from receiving the Rj of robot particle assembly
p*M particle is retained in the particle assembly of distribution of machine people Ri oneself and is represented as Qi, and P is the probability of selected exchange particle, 0<P<1.
5. as claimed in claim 4 can the cognitive wrong multi-robot Cooperation localization method of tolerance, it is characterized in that, the particle assembly that a certain distribution of machine people Ri of described step 4 upgrades oneself:
6. as claimed in claim 1 can the cognitive wrong multi-robot Cooperation localization method of tolerance, it is characterized in that, a certain distribution of machine people Ri of described step 5 calculate a grain subset X '
i,tthe weights of middle particle, and do normalized.
7. as claimed in claim 1 can the cognitive wrong multi-robot Cooperation localization method of tolerance, it is characterized in that, a certain distribution of machine people Ri of described step 6 according to grain subset X '
i,tand weights calculate the posterior probability distribution of robot pose.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410095912.1A CN103914069A (en) | 2014-03-14 | 2014-03-14 | Multi-robot cooperation locating method capable of allowing cognitive errors |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410095912.1A CN103914069A (en) | 2014-03-14 | 2014-03-14 | Multi-robot cooperation locating method capable of allowing cognitive errors |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103914069A true CN103914069A (en) | 2014-07-09 |
Family
ID=51039829
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410095912.1A Pending CN103914069A (en) | 2014-03-14 | 2014-03-14 | Multi-robot cooperation locating method capable of allowing cognitive errors |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103914069A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020110359A1 (en) * | 2018-11-28 | 2020-06-04 | Mitsubishi Electric Corporation | System and method for estimating pose of robot, robot, and storage medium |
| WO2021065122A1 (en) * | 2019-10-02 | 2021-04-08 | Mitsubishi Electric Corporation | System and method for probabilistic multi-robot slam |
| US11448511B2 (en) * | 2020-01-30 | 2022-09-20 | Mitsubishi Electric Research Laboratories, Inc. | System and method for probabilistic multi-robot positioning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07271415A (en) * | 1994-03-30 | 1995-10-20 | Nec Corp | Cooperative robot control method |
| JPH1083208A (en) * | 1996-09-05 | 1998-03-31 | Matsushita Electric Ind Co Ltd | Arithmetic mechanism for inter-robot relative position |
| CN102103815A (en) * | 2009-12-17 | 2011-06-22 | 上海电机学院 | Method and device for positioning particles of mobile robot |
| CN102707675A (en) * | 2012-04-23 | 2012-10-03 | 深圳市中科鸥鹏智能科技有限公司 | Swarm-robot controller, swarm-robot control method and controller terminal |
| CN102830701A (en) * | 2012-08-30 | 2012-12-19 | 北京航空航天大学 | Coordination control method for system with multiple mobile robots |
| CN203275971U (en) * | 2013-04-27 | 2013-11-06 | 湖南科技大学 | Outdoor ground swarm-robot control system |
| CN203414589U (en) * | 2013-06-20 | 2014-01-29 | 沈阳工业大学 | Mutual positioning and confirmation device between robots |
-
2014
- 2014-03-14 CN CN201410095912.1A patent/CN103914069A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07271415A (en) * | 1994-03-30 | 1995-10-20 | Nec Corp | Cooperative robot control method |
| JPH1083208A (en) * | 1996-09-05 | 1998-03-31 | Matsushita Electric Ind Co Ltd | Arithmetic mechanism for inter-robot relative position |
| CN102103815A (en) * | 2009-12-17 | 2011-06-22 | 上海电机学院 | Method and device for positioning particles of mobile robot |
| CN102707675A (en) * | 2012-04-23 | 2012-10-03 | 深圳市中科鸥鹏智能科技有限公司 | Swarm-robot controller, swarm-robot control method and controller terminal |
| CN102830701A (en) * | 2012-08-30 | 2012-12-19 | 北京航空航天大学 | Coordination control method for system with multiple mobile robots |
| CN203275971U (en) * | 2013-04-27 | 2013-11-06 | 湖南科技大学 | Outdoor ground swarm-robot control system |
| CN203414589U (en) * | 2013-06-20 | 2014-01-29 | 沈阳工业大学 | Mutual positioning and confirmation device between robots |
Non-Patent Citations (3)
| Title |
|---|
| 刘利枚等: "粒子群优化的多机器人协作定位方法", 《中南大学学报（自然科学版）》 * |
| 杨玉君等: "自主微小型移动机器人的协作学习", 《计算机工程》 * |
| 钱堃等: "同时机器人定位与人体跟踪的多源感知协作粒子滤波方法", 《机器人》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020110359A1 (en) * | 2018-11-28 | 2020-06-04 | Mitsubishi Electric Corporation | System and method for estimating pose of robot, robot, and storage medium |
| JP2022502791A (en) * | 2018-11-28 | 2022-01-11 | 三菱電機株式会社 | Systems and methods for estimating robot posture, robots, and storage media |
| US11331801B2 (en) * | 2018-11-28 | 2022-05-17 | Mitsubishi Electric Research Laboratories, Inc. | System and method for probabilistic multi-robot positioning |
| JP7166446B2 (en) | 2018-11-28 | 2022-11-07 | 三菱電機株式会社 | System and method for estimating pose of robot, robot, and storage medium |
| WO2021065122A1 (en) * | 2019-10-02 | 2021-04-08 | Mitsubishi Electric Corporation | System and method for probabilistic multi-robot slam |
| US11347238B2 (en) | 2019-10-02 | 2022-05-31 | Mitsubishi Electric Research Laboratories, Inc. | System and method for probabilistic multi-robot SLAM |
| US11448511B2 (en) * | 2020-01-30 | 2022-09-20 | Mitsubishi Electric Research Laboratories, Inc. | System and method for probabilistic multi-robot positioning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108803349B (en) | Optimal consistency control method and system for nonlinear multi-agent system | |
| Pak et al. | Distributed hybrid particle/FIR filtering for mitigating NLOS effects in TOA-based localization using wireless sensor networks | |
| CN112070807B (en) | Multi-target tracking method and electronic device | |
| CN110579738B (en) | Moving target direction angle obtaining method and terminal equipment | |
| Zuo et al. | Posterior CRLB based sensor selection for target tracking in sensor networks | |
| CN103914069A (en) | Multi-robot cooperation locating method capable of allowing cognitive errors | |
| CN104090265A (en) | Positioning method and device | |
| Chen et al. | Information filter with speed detection for indoor Bluetooth positioning | |
| CN105338619A (en) | Positioning method and positioning device | |
| CN104066052B (en) | A kind of processing method and processing device of location data | |
| CN104598621A (en) | Track compression method based on sliding window | |
| CN107728124A (en) | A kind of more radar dynamic regulating methods and device based on comentropy | |
| CN102480784A (en) | Method and system for evaluating fingerprint positioning error | |
| Aryankia et al. | Neural network-based formation control with target tracking for second-order nonlinear multiagent systems | |
| CN103593856A (en) | Method and system for tracking single target | |
| Jia et al. | Mobile robot simultaneous localization and mapping based on a monocular camera | |
| CN104331078A (en) | Multi-robot cooperative localization method based on position mapping algorithm | |
| CN106570311B (en) | Track continuity assessment method and system under complex condition | |
| CN106296650B (en) | A kind of laser point cloud method for registering and device | |
| CN107241746B (en) | Method and device for balancing residual energy of sensor nodes in sensor network | |
| RU2629745C1 (en) | Mobile combat control center | |
| Moreira et al. | Where@ UM-Dependable organic radio maps | |
| CN102640538A (en) | Method and apparatus for selecting a receiver | |
| CN110554604A (en) | multi-agent synchronous control method, equipment and storage equipment | |
| Özkucur et al. | Localization with non-unique landmark observations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140709 |