search

start-jsk / jsk_mbzirc

4 stars 15 forks source link

[Task1]: TF frame tree disconnection in non-ground-truth mode #78

Open tongtybj opened 8 years ago

tongtybj commented 8 years ago

I found that if I use the following non-ground-truth-mode in task1, there is a disconnection in tf frame tree(frames_non_ground_truth.pdf). roslaunch jsk_mbzirc_tasks jsk_mbzirc_task_1.launch use_ground_truth:=false frames_non_ground_truth.pdf. And the "/nav" is fixed at "/world", which "/nav" is the start point of the uav(x:-65m, y: 20m), and "/world" is the the center point of arena.

In comparison, under ground-truth-mode, all frames are connected(frames_ground_truth.pdf) frames_ground_truth.pdf

I will send a PR to fix the tf problem.

tongtybj commented 8 years ago

In hector theory, the "world" means the point where longtitude = 0 and latitude = 0, while the "nav" is the point calculated only by GPS. So in our challenge, the center point of arena does not coincide with the "world" frame.

We have to calculate the offset between the GPS value and relative position of arena(x:-65m, y:20m) in terms of the start point.

k-okada commented 8 years ago

you can define "center_of_the_world"(longitude = 0 and latitude = 0) and "center_of_the_arena", if current tf '/world' is center of arena and GPS publishes the point relative to the center of world, publish the tf between "center of the world" and "center of arena"

◉ Kei Okada

On Thu, Jun 2, 2016 at 7:28 PM, 趙 漠居 notifications@github.com wrote:

In hector theory, the "world" means the point where longtitude = 0 and latitude = 0, while the "nav" is the point calculated only by GPS. So in our challenge, the center point of arena does not coincide with the "world" frame.

We have to calculate the offset between the GPS value and relative position of arena(x:-65m, y:20m) in terms of the start point.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/start-jsk/jsk_mbzirc/issues/78#issuecomment-223254129, or mute the thread https://github.com/notifications/unsubscribe/AAeG3Fc_f0CNmSLm274vTDKvUPOVmU8Yks5qHrAwgaJpZM4IsQXA .

k-okada commented 8 years ago

or consider center of the world is the center of the arena

◉ Kei Okada

On Thu, Jun 2, 2016 at 7:36 PM, Kei Okada k-okada@jsk.t.u-tokyo.ac.jp wrote:

you can define "center_of_the_world"(longitude = 0 and latitude = 0) and "center_of_the_arena", if current tf '/world' is center of arena and GPS publishes the point relative to the center of world, publish the tf between "center of the world" and "center of arena"

◉ Kei Okada

On Thu, Jun 2, 2016 at 7:28 PM, 趙 漠居 notifications@github.com wrote:

In hector theory, the "world" means the point where longtitude = 0 and latitude = 0, while the "nav" is the point calculated only by GPS. So in our challenge, the center point of arena does not coincide with the "world" frame.

We have to calculate the offset between the GPS value and relative position of arena(x:-65m, y:20m) in terms of the start point.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/start-jsk/jsk_mbzirc/issues/78#issuecomment-223254129, or mute the thread https://github.com/notifications/unsubscribe/AAeG3Fc_f0CNmSLm274vTDKvUPOVmU8Yks5qHrAwgaJpZM4IsQXA .