INFORMATION SECURITY ASSESSMENT FOR MULTI-AGENT ROBOTIC SYSTEM UNDER THE INFORMATION IMPACT

Abstract

The paper deals with the features of information security guaranteeing of the multi-agent robotic system with self-organizing behavior. The main attention is paid to the possibility of implementing information security threats on the level of interaction between the individual elements. The definitions “information impact” and “disorganization” are introduced for multi-agent robotic system. As a criterion for the system safety state assessment, probability is selected of number of items available at time t for required task execution of multi-agent robotic system, not suffering from the effects of the information impact. A method for estimating the probability of the multi-agent robotic system being in a safe state is proposed. The method is based on mathematical apparatus of Markov chains. Its distinction is the usage of functional dependencies for intensity information impact. The method gives the possibility to identify required characteristics of the individual elements in the early stages of development. Graphs of probability for secure system state of group of elements at different intensities of information impact by intruder and intensities are given, characterizing software and hardware capabilities of element output from the unsafe condition. The system behavior is modeled in the dynamics for different functional dependencies of the information impact intensity. An example of group consisting of four identical elements staying in a safe condition and attacking by three disorganizing elements is shown. Technique of obtaining numerical values for the intensities of information impact at successive instants is revealed.

Keywords

• information security
• robotic systems
• self-organizing behavior
• information impact
• vulnerability