Ecl: V0.9.0 Release

Riseborough, Paul; Bapst, Roman; Olsson, Carl A.; Küng, Beat; Johnson, Walt; Tak, Bharat; Sauder, Mark; pickledgator,; Oes, Julian; georgehines,; Agar, Daniel; Purohit, Siddharth Bharat; nickolasrossi,; Meier, Lorenz; Charlebois, Mark

doi:10.5281/zenodo.55367

Cited by 2 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The community that developed the first Pixhawk is PX4; they maintain a fully functional and open source autopilot software called PX4 autopilot [22]. In addition to the low-level sensor drivers, control loops, and planning algorithm, PX4 autopilot features an INS/GPS integration routine, called EKF2 [23], as a part of its Estimation and Control Library (ECL) [24]. EKF2 fuses measurements from IMU, Barometer, Magnetometer, and GPS to estimate the drone's states, namely, attitude represented in quaternions, velocity, and position in local frame (North-East-Down).…”

Section: Pixhawk and Px4 Drone Flight Stackmentioning

confidence: 99%

GPS-Denied Navigation Using Low-Cost Inertial Sensors and Recurrent Neural Networks

AbdulMajuid,

Mohamady,

Draz

et al. 2021

Preprint

View full text Add to dashboard Cite

Autonomous missions of drones require continuous and reliable estimates for the drone's attitude, velocity, and position. Traditionally, these states are estimated by applying Extended Kalman Filter (EKF) to Accelerometer, Gyroscope, Barometer, Magnetometer, and GPS measurements. When the GPS signal is lost, position and velocity estimates deteriorate quickly, especially when using low-cost inertial sensors. This paper proposes an estimation method that uses a Recurrent Neural Network (RNN) to allow reliable estimation of a drone's position and velocity in the absence of GPS signal. The RNN is trained on a public dataset collected using Pixhawk. This low-cost commercial autopilot logs the raw sensor measurements (network inputs) and corresponding EKF estimates (ground truth outputs). The dataset is comprised of 548 different flight logs with flight durations ranging from 4 to 32 minutes. For training, 465 flights are used, totaling 45 hours. The remaining 83 flights totaling 8 hours are held out for validation. Error in a single flight is taken to be the maximum absolute difference in 3D position (MPE) between the RNN predictions (without GPS) and the ground truth (EKF with GPS). On the validation set, the median MPE is 35 meters. MPE values as low as 2.7 meters in a 5-minutes flight could be achieved using the proposed method. The MPE in 90% of the validation flights is bounded below 166 meters. The network was experimentally tested and worked in real-time.

show abstract

Section: Pixhawk and Px4 Drone Flight Stackmentioning

confidence: 99%

GPS-Denied Navigation Using Low-Cost Inertial Sensors and Recurrent Neural Networks

AbdulMajuid,

Mohamady,

Draz

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

GNSS-INS-dynamic fusion with robustness to outliers based on external force state estimation

Ye,

Song,

Tang

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Multi-source information fusion state estimation algorithms are an important means for drones to perceive ego-state, and accurate and robust estimation of external forces is crucial for precise control of quadrotors. This paper proposes a method that integrates a dynamic model into a multi-rate extended Kalman filter (EKF) framework on manifold. By estimating the magnitude of the external force acting on vehicle, meanwhile, a dynamic constraint on velocity loop is established to reduce the discrepancy between the model-predicted motion and the actual motion. Moreover, the estimated external force is integrated into the zero velocity update criterion for zero speed judgment, effectively reducing false detections while improving the accuracy of zero speed state recognition. However, multi-source measurements significantly increase the probability of data signal errors. To address this issue, we use a robust estimation algorithm to improve EKF’s sensitivity to abnormal measurements, flexibly adjusting measurement weights while rejecting unreasonable measurements. Validation with open-source indoor and outdoor datasets shows that our algorithm improves pose estimation performance while maintaining accurate positioning accuracy compared to non-dynamic fusion under the same filtering parameters, particularly in global navigation satellite system short time denied. It provides accurate external force estimation, offering multi-source data support in areas such as human–machine interaction and carrying variable mass payloads.

show abstract

Ecl: V0.9.0 Release

Cited by 2 publications

References 0 publications

GPS-Denied Navigation Using Low-Cost Inertial Sensors and Recurrent Neural Networks

GPS-Denied Navigation Using Low-Cost Inertial Sensors and Recurrent Neural Networks

GNSS-INS-dynamic fusion with robustness to outliers based on external force state estimation

Contact Info

Product

Resources

About