Journal: Sensors (Basel, Switzerland)

Article Title: Stretch-ICP: A Continuous-Trajectory Registration and Deskewing Algorithm in Scenarios of Aggressive Motions †

doi: 10.3390/s26082567

Figure Lengend Snippet: Composite image showing our localization system tumbling down a steep hill in one of the 32 runs of the TIGS dataset. The inset photo shows our rugged perception rig, with the numbers in the red circles corresponding to (1) XSens MTi-30 IMU, (2) VectorNav VN-100 IMU, (3) RoboSense RS-16 lidar, and (4) Raspberry Pi 4.

Article Snippet: The first IMU is an XSens MTi-30 (Xsens, Enschede, The Netherlands), with a gyroscope saturating at 10.5 rad/s, despite the Xsens specification sheet stating a saturation point of 7.85 rad/s.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: Stretch-ICP: A Continuous-Trajectory Registration and Deskewing Algorithm in Scenarios of Aggressive Motions †

doi: 10.3390/s26082567

Figure Lengend Snippet: Density map of the TIGS dataset. The grayscale represents the number of data points acquired at the specific angular speeds and linear accelerations. The outlines represent the distributions in linear accelerations and angular speeds for similar datasets. The KITTI dataset is shown in green, the Newer College dataset is indicated in orange, and the Hilti-Oxford dataset is illustrated in red. The dashed line represents the manufacturer-specified saturation point of the MTi-30 gyroscope.

Article Snippet: The first IMU is an XSens MTi-30 (Xsens, Enschede, The Netherlands), with a gyroscope saturating at 10.5 rad/s, despite the Xsens specification sheet stating a saturation point of 7.85 rad/s.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: Stretch-ICP: A Continuous-Trajectory Registration and Deskewing Algorithm in Scenarios of Aggressive Motions †

doi: 10.3390/s26082567

Figure Lengend Snippet: Qualitative and quantitative results on angular speed estimates for SAAVE. ( Left ) An example of the angular velocity over time for the saturated gyroscope (MTi-30) axis for one run of the TIGS dataset. The measurements from the MTi-30 gyroscope are shown in dashed blue, the ground truth measurements from a VN-100 gyroscope are indicated in dashed orange, and the angular speeds estimated with SAAVE using MTi-30 measurements are illustrated in purple. The purple-shaded area represents three standard deviations above and below the angular velocities estimated by SAAVE. The light gray zone indicates a gyroscope saturation episode. ( Right ) The error in angular velocities without (in blue) and with SAAVE (in purple) during saturation periods for all the runs of the TIGS dataset.

Article Snippet: The first IMU is an XSens MTi-30 (Xsens, Enschede, The Netherlands), with a gyroscope saturating at 10.5 rad/s, despite the Xsens specification sheet stating a saturation point of 7.85 rad/s.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: Stretch-ICP: A Continuous-Trajectory Registration and Deskewing Algorithm in Scenarios of Aggressive Motions †

doi: 10.3390/s26082567

Figure Lengend Snippet: Composite image showing our localization system tumbling down a steep hill in one of the 32 runs of the TIGS dataset. The inset photo shows our rugged perception rig, with the numbers in the red circles corresponding to (1) XSens MTi-30 IMU, (2) VectorNav VN-100 IMU, (3) RoboSense RS-16 lidar, and (4) Raspberry Pi 4.

Article Snippet: The table shows that the angular velocities reached in TIGS exceed the operating range of several commonly used sensors, including the Xsens MTi-30 used in our setup, whereas those reached in HRMC remain within the range of a broader set of units.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: Stretch-ICP: A Continuous-Trajectory Registration and Deskewing Algorithm in Scenarios of Aggressive Motions †

doi: 10.3390/s26082567

Figure Lengend Snippet: Density map of the TIGS dataset. The grayscale represents the number of data points acquired at the specific angular speeds and linear accelerations. The outlines represent the distributions in linear accelerations and angular speeds for similar datasets. The KITTI dataset is shown in green, the Newer College dataset is indicated in orange, and the Hilti-Oxford dataset is illustrated in red. The dashed line represents the manufacturer-specified saturation point of the MTi-30 gyroscope.

Article Snippet: The table shows that the angular velocities reached in TIGS exceed the operating range of several commonly used sensors, including the Xsens MTi-30 used in our setup, whereas those reached in HRMC remain within the range of a broader set of units.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: Stretch-ICP: A Continuous-Trajectory Registration and Deskewing Algorithm in Scenarios of Aggressive Motions †

doi: 10.3390/s26082567

Figure Lengend Snippet: Qualitative and quantitative results on angular speed estimates for SAAVE. ( Left ) An example of the angular velocity over time for the saturated gyroscope (MTi-30) axis for one run of the TIGS dataset. The measurements from the MTi-30 gyroscope are shown in dashed blue, the ground truth measurements from a VN-100 gyroscope are indicated in dashed orange, and the angular speeds estimated with SAAVE using MTi-30 measurements are illustrated in purple. The purple-shaded area represents three standard deviations above and below the angular velocities estimated by SAAVE. The light gray zone indicates a gyroscope saturation episode. ( Right ) The error in angular velocities without (in blue) and with SAAVE (in purple) during saturation periods for all the runs of the TIGS dataset.

Article Snippet: The table shows that the angular velocities reached in TIGS exceed the operating range of several commonly used sensors, including the Xsens MTi-30 used in our setup, whereas those reached in HRMC remain within the range of a broader set of units.

Techniques: