A Reliable Behavioral Model: Optimizing Energy Consumption and Object Clustering Quality by Naïve Robots

Related Works

Among clustering scenarios, studied in literature, there are those with both physical and simulated experiments (Gauci, Chen, Li, Dodd, & Groß, 2014) or only simulated experiments (Barfoot & D’eleuterio, 2005; Hartmann, 2005; Vorobyev, Vardy, & Banzhaf, 2014). These scenarios differ also in terms of environment size, number of robots and objects, perception capacities, and possible behaviors of used robots. Hartmann (2005) presents a scenario, where agents-like-ants have to perform a clustering task in a grid of square cells, with a sensing range of 8 cells and 6 possible behaviors. While Barfoot & D’eleuterio (2005) develop a model of 60 objects to be clustered by 30 agents-like-robots that have a local perception of 5 neighbor-cells and 2 behaviors. Vorobyev et al. (2014) use number of objects in short and long detection areas as inputs to a neural network which gives one of the 3 behaviors as output. Through physical robots’ experiments, Gauciet et al. (2014) prove that a clustering task can be accomplished using simple memory-less and compute-less robots equipped with only one frontal sensor and 3 possible behaviors.