The aim of this project was to design software which could be used to control autonomous guided vehicles (AGVs) through a domain without colliding with any obstacles which might be present. The domains modelled included both mobile and stationary obstacles. This was in order to more realistically represent environments such as Automated Warehouses, in which multiple, independently controlled, AGVs travelled while avoiding collisions with both stationary obstacles such as pallets and mobile obstacles such as other AGVs and humans. Several different existing techniques were reviewed in detail for suitability including Lee's algorithm, Generalised Voronai Diagrams and Neural Networks. Those techniques which showed suitability for navigation were implemented. Extensions and modifications were developed to improve the performance of these techniques and implemented. Sample domains were built, and the performance of the textbook algorithms compared against the performance of the newly developed variants. Benchmarks were tabulated and analysed.