Radar Direction-of-Arrival Estimation Using Ego-Velocity Search Grids
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing radar systems for automated driving assistance are computationally expensive and time-consuming in determining the direction of arrival and motion state of objects in the environment, which hinders timely data provision to the automated driving assistance system.
Innovation Solution
A radar system utilizes a simplified two-dimensional angular search grid based on ego velocity and range and radial velocity to reduce the number of search grid points, incorporating a ring for stationary objects and an arc for moving objects, thereby reducing computational cost and processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional grid-based direction of arrival estimation algorithms are used to search every azimuth and elevation in the field of view, then measurement precision is improved, but computational cost and processing time increase significantly
Solution Approach 1:
The patent segments the full 2D angular search space into multiple independent 1D search spaces by dividing the azimuth and elevation fields of view into discrete bins. This allows the system to perform separate 1D FFT operations on azimuth and elevation dimensions, significantly reducing computational complexity while maintaining directional accuracy through the binning approach.
Solution Approach 2:
The patent transforms the 2D angular search problem into two independent 1D search problems by utilizing the separability of azimuth and elevation dimensions. By performing 1D FFT operations on each dimension independently rather than a single 2D operation, the computational burden is dramatically reduced while preserving the ability to accurately estimate direction of arrival.
2Measurement precision
If the radar system searches the entire field of view for objects, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent divides the full field of view into discrete angular bins and performs targeted 1D FFT operations only in the azimuth and elevation dimensions rather than exhaustively searching the entire 2D space. This segmentation approach maintains measurement precision within each bin while dramatically reducing the total processing time required.
Solution Approach 2:
The patent performs preliminary 1D FFT operations on azimuth and elevation data separately, creating transformed data that can be efficiently combined. This preliminary processing enables rapid direction of arrival estimation without requiring a time-consuming full 2D search, thus reducing overall processing time while preserving accuracy.
3Measurement precision
If comprehensive direction of arrival estimation is performed across all search grid points, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent segments the complex 2D direction of arrival estimation problem into two simpler 1D problems by independently processing azimuth and elevation dimensions. Each dimension undergoes a separate 1D FFT operation, which is computationally much simpler than a full 2D operation, thereby reducing device complexity while maintaining estimation accuracy through the binning approach.
Solution Approach 2:
The patent reduces computational complexity by changing from a 2D search approach to two independent 1D searches. By performing 1D FFT operations on azimuth and elevation data separately and then combining the results, the system achieves the same directional information with significantly reduced computational burden and lower device complexity.
Data Source
AI summary
A radar system includes a processor and a non-transitory computer-readable medium storing machine instructions. The processor obtains an ego velocity Vego of a radar system, a range R of an object in an environment of the radar system, and a radial velocity Vr of the object. The processor determines a simplified two-dimensional (2D) angular search grid and performs a grid-based direction-of-arrival algorithm using the simplified 2D angular search grid. In some implementations, the processor determines a ring of possible positions for a stationary object based on the ego velocity Vego, the range R, and the radial velocity Vr, and includes the ring of possible positions in the simplified 2D angular search grid. In some implementations, the processor determines an arc of possible positions for a moving object based on the range R, and includes the arc of possible positions in the simplified 2D angular search grid.