简介：Theeffectsofmeasurementnoiseareinvestigatedinthecontextofbipartiteconsensusofmulti-agentsystems.Inthesystemsetting,discrete-timedouble-integratordynamicsareassumedfortheagent,andmeasurementnoiseispresentfortheagentreceivingthestateinformationfromitsneighbors.Time-varyingstochasticbipartiteconsensusprotocolsaredesignedinordertolessentheharmfuleffectsofthenoise.Consequently,thestatetransitionmatrixoftheclosed-loopsystemisanalyzed,andsufficientandnecessaryconditionsfortheproposedprotocoltobeameansquarebipartiteconsensusprotocolaregivenwiththehelpoflineartransformationandalgebraicgraphtheory.Itisproventhatthesigneddigraphtobestructurallybalancedandhavingaspanningtreearetheweakestcommunicationassumptionsforensuringbipartiteconsensus.Inparticular,theproposedprotocolisameansquarebipartiteaverageconsensusoneifthesigneddigraphisalsoweightbalanced.