System and method for a computer based cooperative work system

Abstract

An agent mediated Computer Supported Cooperative Work (CSCW) system creates a sense of group work and at the same time keeps the privacy and maintains the security of each user. A multi-agent negotiation process is used in the system to reduce fractions among group members during the geographically distributed team work. A markup language, such as XML (extended Markup Language), is used in the system to encode communication messages. Event perception is an important task in agent mediated CSCW system which uses an eigen space to perform the event perception task. For the case where the number of devices is large, an eigen pyramid is constructed which can be used to discriminate different events.

1 - 15 . (canceled)  
   
   
       16 . A computer supported cooperative work (CSCW) method comprising the steps of: 
 dividing sensing devices associated with sensing the environment of a user into groups;    calculating coefficient vectors for each group to form a layer of an eigen space pyramid;    obtaining readings from the sensing devices according to the groups;    generate coefficients for data of each of the groups obtained by the readings; and    connecting all the coefficients together to form a next layer of the eigen space pyramid,    wherein the eigen space pyramid is used to perceive subsequent events from the sensing devices.    
   
   
       17 . The CSCW method according to  claim 16 , wherein the eigen space pyramid perceives subsequent events by matching readings from the sensing devices to sense the environment against learned models of all events.  
   
   
       18 . The CSCW method according to  claim 16 , wherein an average of the coefficient vectors for is calculated by:  
     
       
         
           
             
               
                 C 
                 -> 
               
               i 
             
             = 
             
               
                 
                   ∑ 
                   
                     j 
                     = 
                     1 
                   
                   
                     N 
                     exemplar 
                   
                 
                 ⁢ 
                 
                     
                 
                 ⁢ 
                 
                   c 
                   ij 
                 
               
               
                 N 
                 exemplar 
               
             
           
         
       
     
   
   
       19 . The CSCW method according to  claim 16 , further comprising: 
 for each of the groups, running training data and detecting their principal directions;    collecting the coefficient vectors with respect to the principal directions for each training exemplar;    inputting the coefficient vectors of each training exemplar as an input to the next level of the eigen space pyramid.    
   
   
       20 . The CSCW method according to  claim 19 , further comprising dividing each training vector into a group which generates N exemplar ×N events  coefficients, wherein a total length for the second level of input will be k×N exemplar ×N events , which is less than N total =k(n acceptable −1)+r, where r is a reading.  
   
   
       21 . The CSCW method according to  claim 20 , wherein: 
 if k×N exemplar ×N events >n acceptable , new data is input and the process is repeated to reduce an amount of data.    if k×N exemplar ×N events  is much less than n acceptable  and a further eigen coefficient extraction is meaningless, then k×N exemplar ×N events  are final coefficients of the training exemplar; and    if k×N exemplar ×N events  is less than n acceptable  and a further eigen coefficient extraction is meaningful, then another round of eigen coefficient extraction is performed and newly generated coefficients will be taken as final coefficients of the exemplar of the training events. and newly generated coefficients will be taken as final coefficients of the exemplar of the training events.    
   
   
       22 . The CSCW method according to  claim 21 , wherein: 
 after extracting the final coefficients of all exemplar of all events, the average of the final coefficients of all the exemplar with respect to a given event are taken as the model of the corresponding event.