/PETcat_Benchmark

Repository for benchmarking various SLAM algorithms

PETcat_Benchmark

Repository for benchmarking various SLAM algorithms

The two main repos being used are :

https://github.com/chvmp/champ

https://github.com/xdspacelab/openvslam

The localization methods used:

  1. GMapping
  2. OpenVSLAM

CPU usage

OpenVSLAM : running with gazebo, pangolin viewer

The rqt_top data:

GMapping : running with gazebo and rviz

The rqt_top data:


GPU info

OpenVSLAM : data without gazebo running, input is a video file

GMapping : running with gazebo and rviz


Odometry Drift

Openvslam : Running run_slam on the aist dataset gave three .txt files. They contain frame coordinates, keyframe coordinates and the track timings for each frame.