In future military conflicts, urban environments will increasingly be among the key venues. "Complex events" within complex structures of urban environments and their levels (subsurface, surface and supersurface) place special demands on armed forces and task forces due to dangers when advancing, especially from "invisible" opponents such as snipers, roadside bombs, etc.
In such situations, systems based on ground robots (UGV) and drones (UAS) - hereinafter together referred to as "agents" - can also be used in a swarm for reconnaissance in unclear terrain, provide significant contributions to the situation picture and help to minimize their own failures and collateral damage. Such systems should navigate and act as autonomously as possible, without additional effort and without special technical knowledge of the deployed team. In this way, the soldiers can concentrate on their essential tasks.
In actual systems, a GNSS signal is used for navigation, which in urban environments in general and in the subsurface area in particular is often weak, (actively and/or passively) disturbed or not available at all.
NIKE-SwarmNav therefore aims to develop components which allow a swarm of UGVs and UAS’ to navigate and operate in areas without a GNSS signal. This essentially requires three main components that are closely linked: 1.) state estimation (location and position determination) using onboard sensors and information from neighboring agents (cooperative state estimation), 2.) swarm coordination algorithms allowing for the mission to run self-organized and robust based on the state estimation and 3.) a robust swarm communication method, which enables the cooperative state estimation and dynamic task distribution for a self-organized behavior of the agents.
Particularly innovative here are the interplay of the onboard sensor combination for navigation (LIDAR, UWB, camera, inertial sensors) and the fusion of corresponding data for state estimation without a GNSS signal, the approach of cooperative state estimation and the linking with the swarm coordination methods. In addition to the fulfillment of the mission specifications the swarm coordination must also allow navigation of the agents in such a way that a cooperative state estimation is possible.
At the end, these components are brought together and subject to functional verification in close-to-real-world laboratory environments on specially developed carrier platforms (UAS and UGV). The result is a proof of concept (PoC) as well as an exploitation strategy at module and system level.