LiDAR SLAM with a Wheel Encoder in a Featureless Tunnel Environment

Simultaneous localization and mapping (SLAM) represents a crucial algorithm in the autonomous navigation of ground vehicles. Several studies were conducted to improve SLAM using various sensors robot platforms. However, only few works have focused on applications inside low-illuminated featureless tunnel environments. In this work, we present an improved wheel encoder data from vehicle (AGV) obtain robust performance environment. The system uses FAST-LIO2 LiDAR as baseline algorithm, additional sensor are integrated into structure extended Kalman filter (EKF) algorithm. EKF is used after odometry estimation before process FAST-LIO2. prediction step inertial measurement unit (IMU) data, while correction state estimation. We AGV conduct experiments flat inclined terrain sections results showed that was greatly environment considering both terrains.