Agent motivational and repulsion forces optimization in agent basic movement model

Human will have their own preferences when they are walking such as direction in their movement, destination and walking speed. These preferences and walking patterns can be affected by several factors including psychological behavior of evading other people or obstacle to avoid any collision by exerting certain repulsion forces. To demonstrate plausible human movement pattern, agent simulation model as a tool has been developed in this research studies where the agent will represent the human with the physical characteristics of body radius and Field of View (FoV) to navigate their surroundings. The modification in the proposed agent model was done by utilizing the physical force approach which results in a simpler mathematical representation compared to the original model. In the proposed model, two parameters, namely, Kad and Kro, are used to represent the agent motivational force to move towards its target destination and the obstacle repulsion, respectively. Moreover, heuristics and non-linear parameters were used in the proposed model to provide more realistic agent motion by maintaining their safety boundary from obstacles. To increase the accuracy of the modified model, the Particle Swarm Optimization (PSO) algorithm is used to find the optimal settings for the parameters of the agent basic movement model. The PSO has achieved superior results by optimizing the model parameters to provide safe distance trajectories between agents and obstacles.