Task partitioning for foraging robot swarms based on penalty and reward

This thesis is concerned with foraging robots that are retrieving items to a destination using odometry for navigation in enclosed environments, and their susceptibility to dead-reckoning noise. Such noise causes the location of targets recorded by the robots to appear to change over time, thus reducing the ability of the robots to return to the same location. Previous work on task partitioning was attempted in an effort to decrease this error and increase the rate of item collection by making the robots travel shorter distances. \par
%
Dynamic Partitioning Strategy (DPS) is introduced and explored in this thesis which adjusts the travelling distance from the items location to a collection point as the robots locate the items, through the use of a penalty and reward mechanism. Robots adapt according to their dead-reckoning error rates, where the probability of finding items is related to the ratio between the penalty and the reward parameters. \par
%
In addition, the diversity in the degrees of error within the members of a robot swarm and the performance repercussion in task partitioning foraging tasks is explored. This is achieved by following an experimental framework composed of three stages: emulation, simulation and hardware. An emulation is generated from an ensemble of machine learning techniques. The emulator allows to perform enriched analyses of simulations of the swarm from a global perspective in a relatively low time compared with experiments in simulations and hardware. Experiments with simulation and hardware provide the contribution of each robot in the swarm to the task.