LINS: A Lidar-Inertial State Estimator for Robust and Efficient Navigation

Abstract

We present R-LINS, a lightweight robocentric lidarinertial state estimator, which estimates robot ego-motion using a 6-axis IMU and a 3D lidar in a tightly-coupled scheme. To achieve robustness and computational efficiency even in challenging environments, an iterated error-state Kalman filter (ESKF) is designed, which recursively corrects the state via repeatedly generating new corresponding feature pairs. Moreover, a novel robocentric formulation is adopted in which we reformulate the state estimator concerning a moving local frame, rather than a fixed global frame as in the standard world-centric lidarinertial odometry(LIO), in order to prevent filter divergence and lower computational cost. To validate generalizability and long-time practicability, extensive experiments are performed in indoor and outdoor scenarios. The results indicate that R-LINS outperforms lidar-only and loosely-coupled algorithms, and achieve competitive performance as the state-of-the-art LIO with close to an order-of-magnitude improvement in terms of speed.

LIO-SAM: Tightly-coupled Lidar Inertial Odometry via Smoothing and Mapping

“…In [15], preintegrated IMU measurements are exploited for de-skewing point clouds. A robocentric lidar-inertial state estimator, R-LINS, is presented in [16]. R-LINS uses an error-state Kalman filter to correct a robot's state estimate recursively in a tightly-coupled manner.…”

FAST-LIO: A Fast, Robust LiDAR-inertial Odometry Package by Tightly-Coupled Iterated Kalman Filter

“…This method has also been used in the Boston Dynamics Atlas humanoid robot. Since the computation complexity of particle filter grows quickly with the number of LiDAR points and the system dimension, Kalman filter and its variants are usually more preferred, such as extended Kalman filter [20], unscented Kalman filter [21], and iterated Klamn filter [22].…”

“…Our method falls into the tightly-coupled approach. We adopt an iterated extended Kalman filter similar to [22] to mitigate linearization errors. Kalman filter (and its variants) has a time complexity O m 2 where m is the measurement dimension [23], this may lead to remarkably high computation load when dealing with a large number of LiDAR measurements.…”

“…Naive down-sampling would reduce the number of measurements, but at the cost of information loss. [22] reduces the number of measurements by extracting and fitting ground planes similar to [10]. This, however, does not apply to aerial applications where the ground plane may not always present.…”

LION: Lidar-Inertial Observability-Aware Navigator for Vision-Denied Environments

“…In both cases, the odometry estimates can be refined by aligning lidar-scans or lidar-features with a local map [16,18]. Fusing the information from an IMU generally improves the estimation accuracy [17,[19][20][21], since it further constrains the estimate and guarantees a high-output rate, thanks to techniques such as IMU pre-integration [22]. Regardless of the scan-alignment technique used, lidar-based methods produce poor estimates in those scenarios that are not sufficiently geometrically-rich to constrain the motion estimation [23][24][25].…”