Abstract:Anti-collision assistance (as part of the current push towards increasing vehicular autonomy) critically depends on accurate detection/localization of moving targets in vicinity. An effective solution pathway involves removing background or static objects from the scene, so as to enhance the detection/localization of moving targets as a key component for improving overall system performance. In this paper, we present an efficient algorithm for background removal for automotive scenarios, applicable to commodity frequency-modulated continuous wave (FMCW)-based radars. Our proposed algorithm follows a three-step approach: a) preprocessing of back-scattered received radar signal for 4-dimensional (4D) point clouds generation, b) 3-dimensional (3D) radar ego-motion estimation, and c) notch filter-based background removal in the azimuth-elevation-Doppler domain. To begin, we model the received signal corresponding to multiple-input multiple-output (MIMO) FMCW transmissions and develop a signal processing framework for extracting 4D point clouds. Subsequently, we introduce a robust 3D ego-motion estimation algorithm that accurately estimates source radar velocity, accounting for measurement errors and Doppler ambiguity, by processing the point clouds. Additionally, our algorithm leverages the relationship between Doppler velocity, azimuth angle, elevation angle, and radar ego-motion velocity to identify the background clutter spectrum and employ notch filters for its removal. The performance of our algorithm is evaluated using both simulated data and experiments with real-world data. By offering a fast and computationally efficient solution, our approach contributes to a potential pathway for challenges posed by non-homogeneous environments and real-time processing requirements.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to effectively remove static backgrounds or objects in the radar system of self - driving cars to enhance the detection and positioning accuracy of surrounding moving targets. Specifically, the article proposes an efficient algorithm for dealing with the background removal problem based on Frequency - Modulated Continuous - Wave (FMCW) radar, which is especially suitable for complex non - uniform environments and real - time processing requirements.
### Problem Background
In an autonomous driving system, accurate detection and positioning of surrounding moving targets are crucial. However, in actual environments, such as urban roads, there are a large number of static objects (such as buildings, trees, etc.). These static objects can interfere with radar signals and affect the detection effect of moving targets. Therefore, removing these static backgrounds can significantly improve the performance of Moving Target Indication (MTI).
### Main Challenges
1. **Non - uniform environment**: Unlike airborne radars, the environment in which vehicle - mounted radars are located is non - uniform, that is, the types and distributions of background objects may change within the field of view. This makes it difficult for traditional Space - Time Adaptive Processing (STAP) methods to obtain sufficient homogeneous training samples, resulting in performance degradation.
2. **Dynamically changing background statistical characteristics**: Vehicle movement can cause rapid changes in the background statistical characteristics of the ground and stationary targets (such as trees, buildings, etc.), further increasing the difficulty of background removal.
3. **Real - time processing requirements**: Vehicle - mounted radars need to process environmental perception data in real - time, which means that the algorithm must be efficient and have a small amount of computation, and methods with high computational complexity like STAP cannot be directly deployed.
### Solutions
The paper proposes a three - step method to solve the above problems:
1. **Pre - process the received radar echo signals**: Generate 4D point cloud data (distance, speed, azimuth angle, elevation angle).
2. **3D radar self - motion estimation**: By processing point cloud data, accurately estimate the self - motion speed of the radar (speed components along the x, y, and z axes), and consider measurement errors and Doppler ambiguity problems.
3. **Background removal based on notch filters**: Use the relationship between Doppler frequency and azimuth angle and elevation angle to identify and remove background clutter.
Through this method, the paper provides a fast and computationally efficient background removal scheme, which is suitable for complex non - uniform environments and meets the requirements of real - time processing.
### Formula Summary
- The calculation formula for Doppler velocity \( v_r \):
\[
v_r=\frac{(v_{a,y}-v_y)\cos\theta+(v_{a,x}-v_x)\sin\theta}{\cos\phi}+(v_{a,z}-v_z)\sin\phi
\]
When the target is stationary (\( \vec{v}_a = 0 \)):
\[
v_r=-(v_y\cos\theta+v_x\sin\theta)\cos\phi - v_z\sin\phi
\]
- The optimization problem in 3D self - motion estimation:
\[
\arg\min_{\vec{v}_c,\Theta,\Phi,V}\sum_{i = 1}^N\left[\eta_1^2(\Theta_i-\theta_i)^2+\eta_2^2(\Phi_i-\phi_i)^2+(V_i - v_{r,i})^2\right]
\]
Constraints:
\[
V_i=-\sin(\Theta_i)\cos(\Phi_i)v_x-\cos(\Theta_i)\cos(\Phi_i)v_y-\sin(\Phi_i)v_z
\]
Through these steps, the algorithm proposed in the paper can effectively remove static backgrounds in complex environments, thereby improving the detection and positioning accuracy of moving targets.