Radar Ego Motion Estimation Using Shifted Velocity Frames

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional ego motion estimation methods in vehicles, such as D-GPS, wheel velocity sensors, and radar sensor systems, face accuracy limitations, particularly in dense urban areas and during slippage or high velocities, impacting autonomous vehicle navigation.

Innovation Solution

A radar sensor system that generates an instantaneous three-dimensional ego motion estimate by processing radar frames with shifted velocity values, using linear least squares - Moore-Penrose inverse estimation to select accurate ego motion estimations from detection points, enabling velocities beyond the unambiguous maximum velocity of the radar sensor system.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If radar sensor systems are used for ego motion estimation, then measurement capability is provided, but the maximum detectable velocity is limited by the unambiguous maximum velocity of the radar sensor system

Engineering Contradiction:
Improvevelocity measurement accuracyVSAvoidmaximum detectable velocity
Core Design Contradiction:
Measurement precisionVSSpeed

Solution Approach 1:

The patent transforms the velocity measurement problem from a one-dimensional limited range measurement into a multi-dimensional analysis by examining detection points across multiple radar frames with different velocity assumptions. By concatenating radar frames with shifted velocity values and analyzing the spatial-temporal patterns of detection points, the system can determine the true velocity even when it exceeds the unambiguous maximum velocity, effectively adding temporal and spatial dimensions to the measurement process.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent changes the velocity parameter interpretation by introducing shifted velocity values (original velocity plus multiples of unambiguous maximum velocity) and analyzing detection point consistency across different velocity hypotheses. This parameter transformation allows the system to resolve the velocity ambiguity and accurately measure velocities beyond the traditional unambiguous maximum velocity limit.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If D-GPS is used for ego motion estimation, then position information is obtained, but accuracy is detrimentally impacted by buildings in dense urban areas

Engineering Contradiction:
Improveposition estimation accuracyVSAvoidsignal blockage by buildings
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent uses static environmental objects (buildings, trees, signs) as intermediary reference points for velocity measurement. Instead of relying on GPS satellite signals that are blocked by buildings, the system measures the apparent motion of these stationary objects through radar detection. By tracking the consistent positions of these intermediary objects across multiple frames, the system can accurately estimate vehicle velocity even in dense urban environments where GPS signals are degraded.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If wheel velocity sensors are used for ego motion estimation, then speed estimates are provided, but accuracy deteriorates in the event of slippage

Engineering Contradiction:
Improvespeed estimation accuracyVSAvoidmeasurement reliability during slippage
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent employs static environmental objects as self-referential measurement references that do not depend on vehicle wheel conditions. By measuring the apparent motion of these fixed external objects through radar, the system obtains velocity information that is independent of wheel slippage, tire traction conditions, or drivetrain mechanics, providing reliable speed estimation even when wheel-based sensors fail.

Inventive Principle:
Principle #25Self-service

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach provides high-accuracy ego motion estimation in both ambiguous and non-ambiguous velocity environments, allowing autonomous vehicles to navigate accurately regardless of their speed relative to the radar sensor system's limitations.

Implementation Method 1

a radar sensor system can include a transmit antenna, a receive antenna, and radar processing circuitry. The transmit antenna can be configured to transmit a radar signal into an environment of the radar sensor system. Further, the receive antenna can be configured to receive a return signal from the environment of the radar sensor system responsive to the radar signal.

Methodology Applied
Scientific EffectRadar: Radar

Implementation Method 2

Each of the detection points can have a radial velocity value and an azimuth angle value. The radar processing circuitry can concatenate the radar frame with at least one shifted radar frame to form a concatenated radar frame. The shifted radar frame can include detection points of the radar frame having respective radial velocity values shifted by a multiple of an unambiguous maximum velocity value

Methodology Applied
Scientific EffectDoppler effect: Doppler Effect

Data Source

PatentEP4455728A1Ego motion estimation on sensor in ambiguous velocity environment
Publication Date: 2024.10.30 GM CRUISE HOLDINGS LLC
  • EP4455728A1 patent drawingFigure 1
  • EP4455728A1 patent drawingFigure 2
  • EP4455728A1 patent drawingFigure 3

AI summary

Various technologies described herein pertain to a radar sensor system that performs ego motion estimation. A radar sensor system can be utilized to generate an instantaneous three-dimensional ego motion estimate of a vehicle. The radar sensor system employs an algorithm that enables generating ego motion estimates for velocities of the vehicle that can be greater than, less then, or equal to the unambiguous maximum velocity of the radar sensor system. Moreover, a single radar sensor system of a vehicle can implement the approaches set forth herein and the techniques can be applicable regardless of modulation of the radar sensor system.