Co-supervisor: Hai Thanh Nguyen (Telenor)
This project investigates IoT from a specific perspective: How IoT can be used to evolve cooperative behavior and altruism, and to encourage people to be concerned with the so called “social welfare”, i.e., to increase the welfare and benefits (including selv, in long term) of all parties in the society. This project description actually involves not one project but several related projects.
Telenor’s IoT Infrastructure will be used as a basis.
Some situations need coordination of people’s behavior in order to find a solution that is both individually rational and maximizes social welfare and altruist.
To give an example, existing Smart Parking applications sense the empty parking slots and inform the new drivers entering the Parking place about these slots. Assume that there is one slot left and there are more than one drivers who want to park there. They would most probably manage to resolve the problem without explicit communication based on the courtesy rules. However, the solution then would not necessarily be optimal because there may be some special situations not taken into account such as one of the drivers may have a doctor appointment which is soon, etc.
If this is a full IoT system, then the cars would be able to communicate with each other and negotiate/argue/auction on behalf of their owners. This project will study the methods for multi-agent coordination in IoT that promotes social welfare. The car-agents may negotiate with the other cars in order to park on the slot, through giving information about their driver’s emergency: “I am in hurry, have a doctor appointment. Can I park here if you are not in hurry”. The drivers that shows cooperative/altruistic behaviour would be remembered by the system which will be taken into account in a future scenario.
We envison that internet of things will be an “internet of agents”, therefore we aim an agent-based solution. In such a scenario there are various issues to study, e.g.,:
- Main features of altruism? Reciprocity, imitation, etc? The mechanism of altruism itself.
- how the private info will be protected. Not everybody wants the others to know her/his doctor appointment, while some people don’t care. How to model the personal information sharing and privacy protection
- can the personal agent representing person learn privacy preferences of their owner.
- Trust: how the other drivers can believe in your (personal agent’s) saying you have an appointment. Trust mechanisms in IoT.