New Exercise: HAR(Human Activity Recognition) using LiDAR

[aastha51551]

Hi Everyone!!
I had this idea in my mind for quite some time.
The video-HAR technology approach suffers from occlusion problems, LiDAR technology can tackle these problems more efficiently as it uses 3D point cloud data rather than the 2D RGB images/videos utilized by video cameras.
Although we need to have different kind of models based on the no. of people, we can start at an individual level (one person in a room)

[jmplaza]

Sounds good to me. I suggest you (a) to work in this robotics application (=world in Gazebo Harmonic including lidar sensor and a moving person) out of RoboticsAcademy (RA). And then (b) develop a reference solution (take a look at OpenPose, this webpage and this paper for similar approaches). When you have everything working, then (c) integrate that exercise in RA.

Cheers,

[aastha51551]

Thanks for the directions. I’ll proceed with the tasks as outlined.