IWR6843ISK-ODS: ROS Point Cloud Visualizer demo

user5179531

Part Number: IWR6843ISK-ODS
Other Parts Discussed in Thread: IWR6843ISK

Hi,

I have tried the ROS Point Cloud Visualizer.
"ti\mmwave_industrial_toolbox_3_3_0\labs\lab0006-ros-driver"

Instead of using IWR6843ISK, I use IWR6843ISK-ODS to run this lab because I need the wider FOV to fit the application context

I flash the 68xx demo pre-built binary into the EVM.
"ti\mmwave_sdk_03_01_01_02\packages\ti\demo\xwr68xx\mmw"

And I use the .cfg file in this ROS visualizer lab.

When I mount the sensor on a moving turtlebot and run this lab, I find the point cloud having some curve patterns. But where they are is actually empty space and the obstacle in front is not detected by the sensor. Also there is some points under the ground.

I wonder if the pre-built binary is only suitable for IWR6843ISK. Should I modify the .cfg file or the ROS code to fit the ODS?

Or all I can do now is just change the sensor to IWR6843ISK?

Regards,
Joseph

over 5 years ago

0 Charles Oladimeji over 5 years ago

TI__Mastermind 33805 points

Hi,

The antenna design of the IWR6843ISK and the IWR6843ODS are different, this affects the angle estimation, range etc so it is not direct swap. You will need to account for this in your application.

Regards,
Charles O

0 user5179531 over 5 years ago in reply to Charles Oladimeji

Prodigy 30 points

Hi Charles,

Thanks for your reply.

I am really confused by the output point cloud.

So, the curve pattern and underground points are purely caused by the antenna design of the ODS, right?

Regards,

Joseph

0 Charles Oladimeji over 5 years ago in reply to user5179531

TI__Mastermind 33805 points

Hi,

Most likely, the antenna is different and the points will be interpreted differently.

also the position of the device on the robot also matters, do you have it positioned at the same point as the example?

Regards,

Charles O

0 user5179531 over 5 years ago in reply to Charles Oladimeji

Prodigy 30 points

Hi,

OK, I got it.

I set it up like this. The distance from ground to the top of the sensor is about 35 cm.

I know it is different from the way it is in the lab08.

Due to such patterns, i can't figure out if the output point cloud is upside-down.

Should I reverse the direction of the receiving data in DataHandler.cpp ?

Regards,

Joseph

0 Charles Oladimeji over 5 years ago in reply to user5179531

TI__Mastermind 33805 points

Hi,

The position of your board is inverted compared to the lab, this is why your data is upside-down. Try mounting similar to the lab and see if the data shows the upside-up.

Regards,
Charles O

Sensors

Sensors forum

IWR6843ISK-ODS: ROS Point Cloud Visualizer demo