Browse Prior Art Database

Decentralized coordination of autonomous agents Disclosure Number: IPCOM000246198D
Publication Date: 2016-May-16
Document File: 2 page(s) / 102K

Publishing Venue

The Prior Art Database


Disclosed is a method to perform secured actions in the context of the Internet of Things, where 'things' are represented here by 'agents'. The actions are executed by an autonomous agent when an environmental condition based on trigger agent states becomes true.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 2

Decentralized coordination of autonomous agents

We present here an method based on connected agents, that allows to perform an action based on agent states. Agents are connected devices (a 'thing' in the 'Internet of Things'). There are two kinds of agent:

Action Agent , capable to perform an action.

Trigger Agent , that maintain a state. Trigger agents can be "static" (a clock, a sensor, ...) or they can be connected personal devices associated to real persons (typically: a smartphone, a digital key, …).

    The method requires also a condition based on trigger agent state, and when this condition becomes trues, the action agent performs its action.

    For example: the action agent can be the controller of a bank safe mechanism. Its action consists in opening the safe door only when:

• Several expected persons stays in front of the door (e.g. bank director, security officer, etc.) .

• Each person is authenticated.

• Each person declares its intention to see the safe unlocked, by performing a "secret action".

• No negative environmental event (alarm, intrusion, etc.) cancels the decision to open the safe.

• Expected positive events must occurs (e.g. date constraint)

The "secret action" is local to each trigger agent. For example, it can be a PIN code

typed in a special application of a smartphone, combined with a gesture like holding the smartphone vertically. A Trigger Agent does not know the secret action of the other agents. The action agent does not need to know each secret action. It will gather trigger agent states and decides perform the action.

     All agents are autonomous. They don't need to depend on a central server. They share the network and then they all communicate only with the action agent.

    All Trigger Agents are identified by a code, named ID. Trigger Agents status are identified by another code, named ST. These codes are like password, that is a number, a pass phrase, a One Time Password (OTP), etc. These codes should be encrypted when stored or sent over the network, unless a mechanism based on OTP is used because it ensures the codes are not re-playable.

The method consists in 2 phases: Learn/Maintenance and Check .

   1. Learn / Maintenance Phase
Each Trigger Agent is learn to switch to 'OK to perform' state. This is the secret action. Agent and its state are represented by a couple of codes (ID, ST). For Static Trigger Agent, a sensor or an algorithm (for a date matchi...