search

PengYu-Team / DCL-SLAM

A ROS package of DCL-SLAM: Distributed Collaborative LiDAR SLAM Framework for a Robotic Swarm.
Apache License 2.0
195 stars 20 forks source link

Replacing other slam algorithms as front end？ #17

Open Joosoo1 opened 10 months ago

Joosoo1 commented 10 months ago

Thanks for open-sourcing this work, what modifications have you made to Lio-sam and fast-lio for DCL-SLAM, and if I want to use other algorithms such as DLIO as a front-end, how should I adapt them, please give me some advice! thanks @zhongshp

zhongshp commented 10 months ago

Hi~ Joosoo1 In general, the process to use other algorithms as a front-end is as follows:

1.define a distributedMapping object, as shown in https://github.com/PengYu-Team/DCL-LIO-SAM/blob/2a68c5f503c04eb1a9ddbac4f9b1fb5ed7bb980a/src/mapOptmization.cpp#L64.

  1. call api function to perform distributed pgo, as shown in https://github.com/PengYu-Team/DCL-LIO-SAM/blob/2a68c5f503c04eb1a9ddbac4f9b1fb5ed7bb980a/src/mapOptmization.cpp#L273.
  2. create a thread for distributed loop closure, as shown in https://github.com/PengYu-Team/DCL-LIO-SAM/blob/2a68c5f503c04eb1a9ddbac4f9b1fb5ed7bb980a/src/mapOptmization.cpp#L1816.

You can refer to https://github.com/PengYu-Team/DCL-LIO-SAM/blob/master/src/mapOptmization.cpp in detail.

Joosoo1 commented 10 months ago

Okay, thanks for the reply.