Elevation mapping using stereo vision enabled heterogenous multi-agent robotic network

Multi-agent systems and stereo vision navigation is a challenging area in Robotic research. The task can be widely scalable and optimized for a specific application. An efficient centralized heterogeneous swarm robotic system is proposed and implemented for digital elevation mapping of an unknown terrain using stereo vision. The utility of vision sensors compared to other sensor systems is shown in mapping and navigation application on planetary terrain. The cloud computing paradigm is implemented in the communication module via a private cloud over OLSR. The system utilizes real-time odometry, disparity mapping, cloud computing, and network hazard recovery algorithms for efficient terrain mapping. The map is stored in an octree data structure. The system is designed keeping in perspective planetary exploration mission, however it is widely scalable and can be modified for various civilian and military purposes.