ZJU-FAST-Lab / EGO-Planner-v2

Swarm Playground, the codebase of the paper "Swarm of micro flying robots in the wild"
GNU General Public License v3.0
382 stars 70 forks source link

few questions related to inter drone communication #17

Open shubham-shahh opened 1 year ago

shubham-shahh commented 1 year ago

Hi @bigsuperZZZX, I went through the paper and the supplementary material attached, and the documentation does an excellent job of explaining everything in a lot of detail. I have a few questions related to inter-drone communication 1) in the experiments you performed, were all drones connected to the same ROS master? 2) is there an explicit need for time sync between all the drones? if yes, how did you achieve it? 3) are we handling if there's some offset or delay in trajectories?

thanks

bigsuperZZZX commented 1 year ago
  1. Agents run ROS master individually and communicate via a bridge we wrote.
  2. We use ntpdate for time sync.
  3. Of course because the global time is synchronized.
shubham-shahh commented 1 year ago

Thanks for the info, I'll understand the bridge communication and update this thread

shubham-shahh commented 1 year ago

Hi @bigsuperZZZX I was wondering, initially when the system starts, all the drones will have a odom "0, 0, 0", how did you find the position of each drone in world frame such that, the offsets are added to the odom of each drone. is it made in a way such that they are equally placed away from each other and this offset is added in the odom from the beginning?

thanks

bigsuperZZZX commented 1 year ago

We do it manually

shubham-shahh commented 1 year ago

We do it manually

oh i see, so you transform the odometry of other drones in some central drone's frame right?

bigsuperZZZX commented 1 year ago

yes

shubham-shahh commented 1 year ago

yes

that's interesting, thanks for the info, I am facing difficulties in fetching orientation for accurate transforms, any ideas for that?

shubham-shahh commented 1 year ago

Hi @bigsuperZZZX why is /others_odom being published from swarm_bridge and traj2odom_node I believe the odom published by traj2odom_node node is generated from received traj based on time allocations and the odom published by swarm_bridge is the true odom published by other agents, I am not sure why both of them are published on same topic and for drone_detect, which odom is useful? is traj2odom_node only useful for simulation?

bigsuperZZZX commented 1 year ago

to reduce communication burden

shubham-shahh commented 1 year ago

to reduce communication burden

I understand, but the o/p of traj2odom_node is useful for drone_detect only right?