As road traffic is becoming increasingly dense, new needs in terms of intelligent human-machine interaction are emerging for their control by human operators. One avenue of research consists in assisting them in their control task by an assistant agent. This paper presents a feasibility study in this field, involving interactions between humans and an assistant agent. For this purpose, a game theory-based model is proposed in order to be able to model a context-sensitive system for the cooperative realization of complex tasks. In this case, the participants of the game are human operators and an assisting agent interacting within the framework of the realization of a control task. Thus, each participants can choose an action between two possible ones (to cooperate or not). Then, the proposed utility functions allow to build the context-sensitive payoff matrix at each observation cycle of the human-machine interaction. To validate our model, we have implemented a simulated control situation; it concerns the regulation of traffic through intersections; this involves two human operators and an assistant agent. Thus, the assistant agent uses the game payoff matrix for its decision-making in using Nash equilibrium. This paper describes a feasibility study, focusing on an analysis of the results obtained during the execution of the simulation. Different research perspectives arise from this study in order to improve and generalize the proposed model.

This work contributes to the field of human-system interaction modeling through an artificial intelligence approach. It focuses on the cooperative realization of a complex task. For this purpose, we propose a human-agent interaction model based on game theory to describe the decision making between the human operator and the assistant agent. The proposed model is based on searching Nash equilibria for a repeated two-player game in which each player has a choice between two actions. In particular, the assistant agent knows how to calculate the equilibrium that depends on information coming from the context (human operator and work environment). This approach allows us to consider a context aware human-machine system. Then, the assistant agent knows how to optimize its intervention with regard to the human operator assisted by this agent during a complex task. For the validation of our model, we highlight the efficiency of the assistant agent using this principle by considering a road traffic simulator using Netlogo. An analysis of the simulation results is provided to illustrate the effectiveness of our approach.

The design of agents interacting with human beings is becoming a crucial problem in many real-life applications. Different methods have been proposed in the research areas of human–computer interaction (HCI) and multi-agent systems (MAS) to model teams of participants (agents and humans). It is then necessary to build models analyzing their decisions when interacting, while taking into account the specificities of these interactions. This paper, therefore, aimed to propose an explicit model of such interactions based on game theory, taking into account, not only environmental characteristics (e.g., criticality), but also human characteristics (e.g., workload and experience level) for the intervention (or not) of agents, to help the latter. Game theory is a well-known approach to studying such social interactions between different participants. Existing works on the construction of game matrices required different ad hoc descriptors, depending on the application studied. Moreover, they generally focused on the interactions between agents, without considering human beings in the analysis. We show that these descriptors can be classified into two categories, related to their effect on the interactions. The set of descriptors to use is thus based on an explicit combination of all interactions between agents and humans (a weighted sum of 2-player matrices). We propose a general model for the construction of game matrices based on any number of participants and descriptors. It is then possible to determine using Nash equilibria whether agents decide (or not) to intervene during the tasks concerned. The model is also evaluated through the determination of the gains obtained by the different participants. Finally, we illustrate and validate the proposed model using a typical scenario (involving two agents and two humans), while describing the corresponding equilibria.