Method and system for behavior-based authentication of a user

Abstract

A method for behaviour-based authentication of a current user to a mobile, portable communication system, is implemented using at least one sensor for capturing gross-motor measurement data, a gross-motor classification module, a processor, and an internal memory. Furthermore, a user is registered in the mobile, portable communication system. The sensor is designed to recognise the gross-motor measurement data of a gross-motor movement of the current user of the mobile, portable communication system and the gross-motor classification module is trained to capture a generic gross-motor movement pattern with the help of training data sets of a user cohort. In addition, the gross-motor classification module implements a machine-learning method. The gross-motor classification module is executed by the processor of the mobile, portable communication system.

1 . method for behaviour-based authentication of a current user to a mobile, portable communication system, which has at least one sensor for capturing gross-motor measurement data, a gross-motor classification module, a processor and an internal memory,
 wherein a user is registered in the mobile, portable communication system,   wherein the sensor is designed to capture the gross-motor measurement data of a gross-motor movement of the current user of the mobile portable communication system,   wherein the gross-motor classification module is trained to recognise a generic gross-motor movement pattern with the help of training data sets of a user cohort and implements a machine-learning method, wherein the gross-motor classification module is executed by the processor of the mobile portable communication system,   wherein the method comprises the following steps:   a) repeated execution of the following steps with use of the machine-learning method:
 i. capture of the gross-motor measurement data by the at least one sensor of the mobile, portable communication system, wherein the gross-motor measurement data are the movement data of the gross-motor movement of the current user, 
 ii. input of the gross-motor measurement data into the gross-motor classification module, 
 iii. generation of a first classification result by the gross-motor classification module, detailing whether the current user is the user registered in the mobile, portable communication system, 
 iv. storage of the first classification result in the memory of the mobile, portable communication system, and 
 v. training of the gross-motor classification module with the gross-motor measurement data of the current user in order to train the gross-motor classification module for a user-specific gross-motor movement pattern on the condition that, in accordance with the first classification result, the current user is the user registered in mobile, portable communication system, 
   b) access to the memory of the mobile, portable communication system in order to read out at least one of the stored first classification results from the memory,   c) evaluation of the at least one read-out first classification result in accordance with a specified checking criterion,   d) generation of an authentication signal if the checking criterion is met, wherein the authentication signal signals a successful authentication of the current user.   
     
     
         2 . The method according to  claim 1 , wherein the mobile, portable communication system comprises an untrained application behaviour classification module,
 wherein the application behaviour classification module is executed by the processor of the mobile, portable communication system,   wherein the method further comprises:   a) repeated execution of the following steps with use of the machine-learning method:
 i. capture of application data, 
 ii. input of the application data into the application behaviour classification module, 
 iii. generation of a second classification result by the application behaviour classification module, detailing whether the current user is the user registered in the mobile, portable communication system, 
 iv. storage of the second classification result in the memory of the mobile, portable communication system, 
 v. training of the application behaviour classification module with the application data of the current user in order to train the application behaviour classification module to a user-specific application behaviour pattern on the condition that, in accordance with the second classification result, the current user is the user registered in the system and/or on the condition that, in accordance with the first classification result, the current user is the user registered in the system, 
   b) access to the memory of the mobile, portable communication system in order to read out at least one of the stored second classification results from the memory,   wherein the second classification result is also included in the evaluation of the first classification result in accordance with the checking criterion.   
     
     
         3 . The method according to  claim 2 , wherein the application data can comprise:
 position data of the mobile, portable communication system, which are captured by a method for position determination by a sensor for determining the position of the mobile, portable communication system, and/or   application usage data of the current user, and/or   biometric data of the current user, which are captured by a sensor for capturing biometric data, and/or   connection data of the mobile, portable communication system to other devices and/or   calendar and/or time data of a clock implemented in the mobile, portable communication system or an external clock, the signal of which is received by a sensor of the mobile, portable communication system.   
     
     
         4 . The method according to  claim 1 , wherein the mobile, portable communication system comprises a fine-motor classification module,
 wherein the fine-motor classification module is configured for classification of fine-motor measurement data and is trained for recognition of a fine-motor movement of a registered user, wherein the fine-motor classification module is executed by the processor of the mobile, portable communication system,   wherein the method further comprises:
 repeated execution of the following steps with use of the machine-learning method:
 capture of the fine-motor measurement data, 
 input of the fine-motor measurement data into the fine-motor classification module, 
 generation of a third classification result by the fine-motor classification module, detailing whether the current user is the user registered in the mobile, portable communication system, 
 storage of the third classification result in the memory of the mobile, portable communication system, 
 training of the fine-motor classification module with the fine-motor measurement data of the current user in order to train the fine-motor classification module to a user-specific fine-motor movement pattern on the condition that, in accordance with the third classification result, the current user is the user registered in the system and/or on the condition that, in accordance with the first classification result, the current user is the user registered in the system, 
 
 access to the memory of the mobile, portable communication system in order to read out at least one of the stored third classification results from the memory, 
   wherein the third classification result is also included in the evaluation of the first classification result in accordance with the checking criterion.   
     
     
         5 . The method according to  claim 1 , wherein at least one first pattern in the form of a first pattern function and at least one first comparison data set are stored in the memory of the mobile, portable communication system,
 wherein the first comparison data set comprises a plurality of the gross-motor measurement data, wherein at least one first comparison parameter is calculated from the plurality of the gross-motor measurement data of the first comparison data set,   wherein the gross-motor classification module performs the following steps when the gross-motor measurement data are input:
 a) comparison of the captured gross-motor measurement data with the at least one first pattern function, 
 b) assignment of the gross-motor measurement data to the first pattern assigned to the first pattern function and attainment of at least one first classification parameter corresponding to the first pattern, if the gross-motor measurement data can be assigned to the at least one first pattern, 
 c) calculation of a confidence value for each first classification parameter by a comparison of the at least one first classification parameter with the relevant first comparison parameter of the first comparison data set, and 
 d) generation of the first classification result from the first confidence values of the first classification parameters, 
   and wherein the step of training comprises an addition of the captured gross-motor measurement data to the first comparison data set.   
     
     
         6 . The method according to  claim 2 , wherein at least one second pattern in the form of a second pattern function and at least one second comparison data set are stored in the memory of the mobile, portable communication system,
 wherein the second comparison data set comprises a plurality of the application data, wherein at least one second comparison parameter is calculated from the plurality of the application data of the second comparison data set, wherein the application behaviour classification module performs the following steps when the application data are input:
 a) comparison of the captured application data with the at least one second pattern function, 
 b) assignment of the application data to the second pattern assigned to the second pattern function and attainment of at least one second classification parameter corresponding to the second pattern, if the application data can be assigned to the at least one second pattern, 
 c) calculation of a confidence value for each second classification parameter by a comparison of the second classification parameter with the relevant second comparison parameter of the second comparison data set, and 
 d) generation of the second classification result from the second confidence values of the second classification parameters, 
   and wherein the step of training comprises an addition of the captured application data to the second comparison data set.   
     
     
         7 . The method according to  claim 4 , wherein at least one third pattern in the form of a third pattern function and at least one third comparison data set are stored in the memory of the mobile, portable communication system, wherein the third comparison data set comprises values for at least one third comparison parameter, wherein the fine-motor classification module performs the following steps when the fine-motor measurement data are input:
 a) comparison of the captured fine-motor measurement data with the at least one third pattern function,   b) assignment of the fine-motor measurement data to the third pattern assigned to the third pattern function and attainment of at least one third classification parameter corresponding to the third pattern, if the fine-motor measurement data can be assigned to the at least one third pattern,   c) calculation of a confidence value for each third classification parameter by a comparison of the third classification parameter with the relevant third comparison parameter of the third comparison data set, and   d) generation of the third classification result from the third confidence values of the third classification parameters,   
       and wherein the step of training comprises an addition of the captured fine-motor measurement data to the third comparison data set. 
     
     
         8 . The method according to  claim 1 , wherein the checking criterion is met if:
 at least one of the first and/or second and/or third classification results exceeds/exceed a threshold value specified by the checking criterion and/or   a maximum age of the first and/or second and/or third classification results specified by the checking criterion is not exceeded and/or   a minimum number of first and/or second and/or third classification results exceeding the threshold value is present.   
     
     
         9 . The method according to  claim 6 , wherein the gross-motor measurement data which are part of the first comparison data set and which are older than a defined time are removed from the first comparison data set and are deleted from the memory of the mobile, portable communication system, and the application data, which are part of the second comparison data set and which are older than the defined time are removed from the second comparison data set and are deleted from the memory of the mobile, portable communication system, and the fine-motor measurement data which are part of the third comparison data set and which are older than the defined time are removed from the third comparison data set and are deleted from the memory of the mobile, portable communication system. 
     
     
         10 . The method according to  claim 4 , wherein, at the time of evaluation of the first, second and/or third classification result, the second and/or third classification result is included only if, in the previous step, the gross-motor classification module recognised a gross-motor movement of the current user in the gross-motor measurement data. 
     
     
         11 . The method according to  claim 1 , wherein the current user of the mobile, portable communication system must authenticate himself to the mobile, portable communication system following an initial commissioning. 
     
     
         12 . The method according to  claim 1 , wherein the mobile, portable communication system sends a signal to the current user after the initial commissioning, which signal contains a request for personalisation of the mobile, portable communication system by generation of the at least one first and/or second and/or third comparison data set by the current user. 
     
     
         13 . The method according to  claim 1 , wherein the machine-learning method is a random forest algorithm. 
     
     
         14 . A system for behaviour-based authentication of a current user to a mobile, portable communication system, which has at least one sensor for capturing gross-motor measurement data, a gross-motor classification module, a processor and an internal memory,
 wherein a user is registered in the mobile, portable communication system,   wherein the sensor is designed to capture the gross-motor measurement data of a gross-motor movement of the current user of the mobile portable communication system,   wherein the gross-motor classification module is trained to recognise a generic gross-motor movement pattern with the help of training data sets of a user cohort and implements a machine-learning method, wherein the gross-motor classification module is executed by the processor of the mobile portable communication system,   wherein the system for behaviour-based authentication of the current user to the mobile, portable communication system performs the following method steps:
 a) repeated execution of the following steps with use of the machine-learning method:
 i. capture of the gross-motor measurement data by the at least one sensor of the mobile, portable communication system, wherein the gross-motor measurement data are the movement data of the gross-motor movement of the current user, 
 ii. input of the gross-motor measurement data into the gross-motor classification module, 
 iii. generation of a first classification result by the gross-motor classification module, detailing whether the current user is the user registered in the mobile, portable communication system, 
 iv. storage of the first classification result in the memory of the mobile, portable communication system, and 
 v. training of the gross-motor classification module with the gross-motor measurement data of the current user in order to train the gross-motor classification module for a user-specific gross-motor movement pattern on the condition that, in accordance with the first classification result, the current user is the user registered in mobile, portable communication system, 
 
 b) access to the memory of the mobile, portable communication system in order to read out at least one of the stored first classification results from the memory, 
 c) evaluation of the at least one read-out first classification result in accordance with a specified checking criterion, 
 d) generation of an authentication signal if the checking criterion is met, wherein the authentication signal signals a successful authentication of the current user. 
   
     
     
         15 . The system according to  claim 14 , wherein the mobile, portable communication system comprises an untrained application behaviour classification module,
 wherein the application behaviour classification module is executed by the processor of the mobile, portable communication system,   wherein the system for behaviour-based authentication of the current user to the mobile, portable communication system performs the following method steps:
 a) repeated execution of the following steps with use of the machine-learning method:
 i. capture of application data, 
 ii. input of the application data into the application behaviour classification module, 
 iii. generation of a second classification result by the application behaviour classification module, detailing whether the current user is the user registered in the mobile, portable communication system, 
 iv. storage of the second classification result in the memory of the mobile, portable communication system, 
 v. training of the application behaviour classification module with the application data of the current user in order to train the application behaviour classification module to a user-specific application behaviour pattern on the condition that, in accordance with the second classification result, the current user is the user registered in the system and/or on the condition that, in accordance with the first classification result, the current user is the user registered in the system, 
 
 b) access to the memory of the mobile, portable communication system in order to read out at least one of the stored second classification results from the memory, 
   wherein the second classification result is also included in the evaluation of the first classification result in accordance with the checking criterion.   
     
     
         16 . The system according to  claim 15 , wherein the application data can comprise:
 position data of the mobile, portable communication system, which are captured by a method for position determination by a sensor for determining the position of the mobile, portable communication system, and/or   application usage data of the current user, and/or   biometric data of the current user, which are captured by a sensor for capturing biometric data, and/or   connection data of the mobile, portable communication system to other devices and/or   calendar and/or time data of a clock implemented in the mobile, portable communication system or an external clock, the signal of which is received by a sensor of the mobile, portable communication system.   
     
     
         17 . The system according to  claim 14 , wherein the mobile, portable communication system comprises a fine-motor classification module,
 wherein the fine-motor classification module is configured to classify fine-motor measurement data and is trained to recognise a fine-motor movement of a registered user, wherein the fine-motor classification module is executed by the processor of the mobile, portable communication system,   wherein the system for behaviour-based authentication of the current user to the mobile, portable communication system performs the following method steps:
 repeated execution of the following steps with use of the machine-learning method:
 capture of the fine-motor measurement data, 
 input of the fine-motor measurement data into the fine-motor classification module, 
 generation of a third classification result by the fine-motor classification module, detailing whether the current user is the user registered in the mobile, portable communication system, 
 storage of the third classification result in the memory of the mobile, portable communication system, 
 training of the fine-motor classification module with the fine-motor measurement data of the current user in order to train the fine-motor classification module to a user-specific fine-motor movement pattern on the condition that, in accordance with the third classification result, the current user is the user registered in the system and/or on the condition that, in accordance with the first classification result, the current user is the user registered in the system, 
 
 access to the memory of the mobile, portable communication system in order to read out at least one of the stored third classification results from the memory, 
   wherein the third classification result is also included in the evaluation of the first classification result in accordance with the checking criterion.   
     
     
         18 . The system according to  claim 14 , wherein at least one first pattern in the form of a first pattern function and at least one first comparison data set are stored in the memory of the mobile, portable communication system,
 wherein the first comparison data set comprises a plurality of the gross-motor measurement data, wherein at least one first comparison parameter is calculated from the plurality of the gross-motor measurement data of the first comparison data set,   wherein the gross-motor classification module performs the following steps when the gross-motor measurement data are input:
 a) comparison of the captured gross-motor measurement data with the at least one first pattern function, 
 b) assignment of the gross-motor measurement data to the first pattern assigned to the first pattern function and attainment of at least one first classification parameter corresponding to the first pattern, if the gross-motor measurement data can be assigned to the at least one first pattern, 
 c) calculation of a confidence value for each first classification parameter by a comparison of the at least one first classification parameter with the relevant first comparison parameter of the first comparison data set, and 
 d) generation of the first classification result from the first confidence values of the first classification parameters, 
   and wherein the step of training comprises an addition of the captured gross-motor measurement data to the first comparison data set.   
     
     
         19 . The system according to  claim 15 , wherein at least one second pattern in the form of a second pattern function and at least one second comparison data set are stored in the memory of the mobile, portable communication system,
 wherein the second comparison data set comprises a plurality of the application data, wherein at least one second comparison parameter is calculated from the plurality of the application data of the second comparison data set,   wherein the application behaviour classification module performs the following steps when the application data are input:
 a) comparison of the captured application data with the at least one second pattern function, 
 b) assignment of the application data to the second pattern assigned to the second pattern function and attainment of at least one second classification parameter corresponding to the second pattern, if the application data can be assigned to the at least one second pattern, 
 c) calculation of a confidence value for each second classification parameter by a comparison of the second classification parameter with the relevant second comparison parameter of the second comparison data set, and 
 d) generation of the second classification result from the second confidence values of the second classification parameters, 
   and wherein the step of training comprises an addition of the captured application data to the second comparison data set.   
     
     
         20 . The system according to  claim 17 , wherein at least one third pattern in the form of a third pattern function and at least one third comparison data set are stored in the memory of the mobile, portable communication system,
 wherein the third comparison data set comprises values for at least one third comparison parameter,   wherein the fine-motor classification module performs the following steps when the fine-motor measurement data are input:
 a) comparison of the captured fine-motor measurement data with the at least one third pattern function, 
 b) assignment of the fine-motor measurement data to the third pattern assigned to the third pattern function and attainment of at least one third classification parameter corresponding to the third pattern, if the fine-motor measurement data can be assigned to the at least one third pattern, 
 c) calculation of a confidence value for each third classification parameter by a comparison of the third classification parameter with the relevant third comparison parameter of the third comparison data set, and 
 d) generation of the third classification result from the third confidence values of the third classification parameters, 
   and wherein the step of training comprises an addition of the captured fine-motor measurement data to the third comparison data set.