Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DeRO: Dead Reckoning Based on Radar Odometry With Accelerometers Aided for Robot Localization

Hoang Viet Do, Yong Hun Kim, Joo Han Lee, Min Ho Lee, Jin Woo Song

TL;DR

DeRO addresses GNSS-denied robot localization by treating radar velocity as the primary dead reckoning signal, while using accelerometer-derived tilt and radar range for corrections. The method combines a stochastic cloning indirect EKF with a radar Doppler velocity estimate and a scan-matching based distance update, jointly estimating a radar velocity scale factor. Across five open datasets, DeRO achieves about 62% translation and 66% rotation improvements over a state-of-the-art radar-inertial fusion baseline. The results demonstrate robust odometry under challenging conditions, with a public ROS 2 implementation released for reproducibility.

Abstract

In this paper, we propose a radar odometry structure that directly utilizes radar velocity measurements for dead reckoning while maintaining its ability to update estimations within the Kalman filter framework. Specifically, we employ the Doppler velocity obtained by a 4D Frequency Modulated Continuous Wave (FMCW) radar in conjunction with gyroscope data to calculate poses. This approach helps mitigate high drift resulting from accelerometer biases and double integration. Instead, tilt angles measured by gravitational force are utilized alongside relative distance measurements from radar scan matching for the filter's measurement update. Additionally, to further enhance the system's accuracy, we estimate and compensate for the radar velocity scale factor. The performance of the proposed method is verified through five real-world open-source datasets. The results demonstrate that our approach reduces position error by 62% and rotation error by 66% on average compared to the state-of-the-art radar-inertial fusion method in terms of absolute trajectory error.

DeRO: Dead Reckoning Based on Radar Odometry With Accelerometers Aided for Robot Localization

TL;DR

DeRO addresses GNSS-denied robot localization by treating radar velocity as the primary dead reckoning signal, while using accelerometer-derived tilt and radar range for corrections. The method combines a stochastic cloning indirect EKF with a radar Doppler velocity estimate and a scan-matching based distance update, jointly estimating a radar velocity scale factor. Across five open datasets, DeRO achieves about 62% translation and 66% rotation improvements over a state-of-the-art radar-inertial fusion baseline. The results demonstrate robust odometry under challenging conditions, with a public ROS 2 implementation released for reproducibility.

Abstract

In this paper, we propose a radar odometry structure that directly utilizes radar velocity measurements for dead reckoning while maintaining its ability to update estimations within the Kalman filter framework. Specifically, we employ the Doppler velocity obtained by a 4D Frequency Modulated Continuous Wave (FMCW) radar in conjunction with gyroscope data to calculate poses. This approach helps mitigate high drift resulting from accelerometer biases and double integration. Instead, tilt angles measured by gravitational force are utilized alongside relative distance measurements from radar scan matching for the filter's measurement update. Additionally, to further enhance the system's accuracy, we estimate and compensate for the radar velocity scale factor. The performance of the proposed method is verified through five real-world open-source datasets. The results demonstrate that our approach reduces position error by 62% and rotation error by 66% on average compared to the state-of-the-art radar-inertial fusion method in terms of absolute trajectory error.
Paper Structure (22 sections, 26 equations, 5 figures, 1 table, 1 algorithm)

This paper contains 22 sections, 26 equations, 5 figures, 1 table, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORKS
  3. Dead Reckoning
  4. Radar Inertial Odometry
  5. DEAD RECKONING USING RADAR ODOMETRY
  6. Coordinate Frame and Mathematical Notations
  7. System Overview
  8. Radar-based Dead Reckoning Mechanization
  9. STOCHASTIC CLONING INDIRECT EXTENDED KALMAN FILTER
  10. State Augmentation
  11. System Model
  12. Measurement Model
  13. Radar range measurements
  14. Tilt angles
  15. Implementation
  16. ...and 7 more sections

Figures (5)

  • Figure 1: Block diagram illustrates the distinction between conventional RIO and DR-based RO structures.
  • Figure 2: Overview block diagram of the proposed method.
  • Figure 3: Comparison of 2D trajectory estimation between EKF RIO and SC-IEKF DeRO of the Carried 1 dataset.
  • Figure 4: Boxplots of the relative translation and rotation errors over distance traveled between the two investigated method of the Carried 1 dataset.
  • Figure 5: 3D scale factor estimation of the proposed method with the Carried 1 dataset.