UnmannedSim : urban and multi-agent navigation network-enabled drones simulator for path planning and localization research

Abstract

The possibility of deploying unmanned aerial vehicles (UAV) for smart city applications has sparked great interest in recent times. High-density cities like Hong Kong, Tokyo and New York have been moving forward to automation to lighten the load of the cities, as well as to provide a time saving, yet safe environment for their citizens. Simulating such UAV activities before practical implementation has therefore turned into a necessity. However, simulation of cities through traditional methods do not replicate the real-life errors in navigation caused by high-density tall buildings, variable wind speed, GNSS sensor measurement errors and other factors. In our research, we have built a new simulator named UnmannedSim which utilizes popular existing software packages such as AirSim and Unreal Engine while adding real-life errors to the simulator.