Cargo tasks arrive at random storage racks and must be shuttled to the launch pad.
Idle robots claim the oldest waiting task, drive to pick it up, then deliver it —
slowing down near other robots to avoid collisions (a simple potential-field
avoidance term).
v_robot = normalize(target - pos) * speed
v_avoid = sum_j (pos - pos_j) / |pos - pos_j|² for |pos-pos_j| < avoidRadius
v_final = v_robot + k_avoid * v_avoid
utilisation = busyRobots / fleetSize
- Robot fleet size — number of autonomous ground vehicles working the floor.
- Task arrival rate — how often new cargo jobs appear at random storage racks.
- Robot speed — top travel speed of each robot.
- Collision-avoid radius — how far away robots start steering around each other; too small causes gridlock, too large wastes travel distance.
This queue-and-fleet model mirrors real autonomous ground support equipment at
spaceports and container terminals, where dozens of robots must be scheduled and
routed without a human directing every trip.