MILIOM: Tightly Coupled Multi-Input Lidar-Inertia Odometry and Mapping

In this letter we investigate a tightly coupled Lidar-Inertia Odometry and Mapping (LIOM) scheme, with the capability to incorporate multiple lidars complementary field of view (FOV). essence, devise time-synchronized scheme combine extracted features from separate into single pointcloud, which is then used construct local map compute feature-map matching (FMM) coefficients. These coefficients, along IMU preinteration observations, are factor graph that will be optimized produce an estimate sliding window trajectory. We also propose key frame-based management strategy marginalize certain poses pointclouds in grow global map, assemble later stage. The use FOV ensures our has low drift can sustain good localization situations where lidar gives poor result, or even fails work. Multi-thread computation implementations adopted fractionally cut down time ensure real-time performance. demonstrate efficacy system via series experiments on public datasets collected aerial vehicle.