Project: a collaborative project where we work as a team to design for NASA Intelligent Robotics Group.
Problem: design a scalable interface that allows to efficiently remote control large swarms of robots on alien terrains.
Approach: design an interface using new research, and also drawing from already existing video game interfaces, especially real-time strategy games, and integrate first person drone vision (camera or lidar), third-person drone vision, and an isometric overview vision.
The robots will disperse over already mapped terrain and probe soil for a specific mineral, determined by control station. The swarm is mostly autonomous. At the beginning of the mission, the controller will divide the swarm into teams and send them off in several areas to probe. Some robots will encounter problems, and the mission will automatically scale down to adjust for the loss. Some robots will successfully complete the mission, and return to base.
One robot’s arm gets stuck and the robot reports an error to control, and waits. The control tries to troubleshoot with software, then uses a camera from a nearby robot to inspect the error robot. After confirming the situation, control makes software adjustments for robot to move a certain way. The robot is freed, but its arm is broken. Control makes the decision for the robot to either proceed (mission permitting) or return to base as a decommissioned robot.