search

TixiaoShan / LIO-SAM

LIO-SAM: Tightly-coupled Lidar Inertial Odometry via Smoothing and Mapping
BSD 3-Clause "New" or "Revised" License
3.47k stars 1.27k forks source link

"Large velocity, reset IMU-preintegration!" #397

Open wwdlh opened 1 year ago

wwdlh commented 1 year ago

I tried to run this algorithm with livox mid70 lidar. The error "Large velocity, reset IMU preintegration!" appeared, and the trajectory could be seen flying in rviz. The parameters related to the lidar are set according to the Livox Horizon lidar in README.md. The livox_ros_driver used is also the version that can publish the correct message type after the author modifying it. The imu I used is ibc-nav40. This imu will not make mistakes when running the lio-sam with the velodyne vlp16 lidar. Therefore, I guess it is a problem that the mid70 solid-state laser radar is not supported. I want to ask the author and everyone for help, is there a way to solve this problem by modifying the parameter file or part of the code

Zermelo-101 commented 1 year ago

Did you find a fix to this problem ?

JokerJohn commented 1 year ago

I believe the main reason for this issue is that the pre-integration node constructs a factor graph of size 100 and uses optimized lidar pose to estimate IMU bias. This factor graph is reset and re-propagated with new bias every time 100 lidar poses are received, causing the predicted pose to be discontinuous and resulting in system crashes. I suggest trying to reduce the weight of lidar pose. (priorPoseNoise = gtsam::noiseModel::Diagonal::Sigmas((gtsam::Vector(6) << 1e-2, 1e-2, 1e-2, 1e-2, 1e-2, 1e-2).finished()); // rad,rad,rad,m, m, m)

Stream11 commented 1 year ago

@JokerJohn Hello, I encountered this error when I was running a very large package (greater than 1000s). .The front part of the mapping is normal. Do you have any suggestions? Thank you