Behavioral Intelligence-Based Cooperative Task Handling for Social Robots

The area of social robotics depicts the role of humanoid robots performing different tasks within human society. However, it raises a philosophical debate about how the growing potential of humanoid robots will benefit society. This paper highlights the challenges associated with social robots adopting human-like behavior. Furthermore, we introduce a cooperative multi-agent system approach designed to optimize task handling and reflect the organizational behavior of agents. The method introduces a dynamic cooperative model (DCM) to explore the steady behavior transitions among agents, serving both environmental contextual goals and individual goals. Initially, every agent sets up its cooperative workspace based on task execution similarity data obtained from a central planner approach. The DCM offers a balanced decision-making strategy to serve environmental goals. We utilized an efficient optimization-based approach to balance the distribution of effort cost flow among all agents, allowing every individual to share a similar amount of effort to improve the system's productivity. To serve individuals' goals, we introduce raw and greedy decision-making strategies. Individuals derive the concepts of raw and greedy decision-making from their cooperative instincts and greedy cooperative natures, respectively. Cooperative instinct values represent the degree of eagerness to participate in cooperation for each agent during their possible interactions. However, greedy decision-making serves the individuals' goal of attaining higher relief with a minimum amount of cooperation with other agents, and vice versa. We perform simulations in MATLAB, and the convergence in behavior of agents confirms this method's overall ability for collision-free task execution and smooth strategy transitions.