Long-Range LiDAR Speed Estimation Using Segmented Range CNNs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speed estimation systems for vehicles, particularly in autonomous driving and advanced driver assistance systems, face inaccuracies due to dependence on external sensors like IMUs and GPS, and challenges in detecting objects at long distances using cameras, leading to errors in absolute speed measurement.
Innovation Solution
The use of long-range LIDAR sensors to estimate absolute speed by dividing LIDAR frames into short, middle, and long ranges, with separate processing for each range using convolutional neural networks to estimate self and relative speeds, and combining these estimates to achieve accurate absolute speed calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If camera-based object annotation and tracking is used to determine relative speed, then speed estimation can be performed, but detection accuracy deteriorates at long distances causing loss of track
Solution Approach 1:
The patent segments the speed estimation task into two independent components: self-speed estimation using LIDAR data from closer ranges, and relative speed estimation using LIDAR data from longer ranges. This segmentation allows each component to be optimized for its specific range, improving overall reliability at long distances where camera-based tracking fails.
2Measurement precision
If IMU or GPS sensors are used to determine self speed, then speed measurement is available, but accuracy deteriorates due to dependence on external sensors and vehicle operating conditions
Solution Approach 1:
The patent replaces mechanical sensor systems (IMU, GPS, odometer) with an optical-based LIDAR system that uses laser ranging and point cloud analysis. This substitution eliminates dependence on external sensors and vehicle operating conditions, providing more reliable self-speed estimation through direct geometric measurement of motion from sequential LIDAR frames.
3Device complexity
If LIDAR frames are processed as a whole, then processing is simpler, but estimation accuracy deteriorates due to mixed range effects
Solution Approach 1:
The patent divides each LIDAR frame into multiple range segments (e.g., short-range and long-range point clouds) and processes each segment separately through dedicated neural network branches. This segmentation allows the system to apply range-appropriate processing strategies, improving speed estimation accuracy by preventing mixed range effects from degrading the overall estimation.
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 more accurate and reliable speed estimation for objects at longer distances, reducing errors and enhancing safety in autonomous driving environments by leveraging LIDAR's robustness to vehicle operating conditions.
Implementation Method 1
a LIDAR sensor to determine the absolute speed of an object
Implementation Method 2
The absolute speed of an object in the vicinity of a vehicle is typically based on the determined self speed of the ego vehicle and the relative speed of the object
Data Source
AI summary
A LIDAR-based method of determining an absolute speed of an object at a relatively longer distance from an ego vehicle, including: estimating a self speed of the ego vehicle using a first frame t−1 and a second frame t obtained from a LIDAR sensor by estimating an intervening rotation θ about a z axis and translation in orthogonal x and y directions using a deep learning algorithm over a relatively closer distance range; dividing each of the first frame t−1 and the second frame t into multiple adjacent input ranges and estimating a relative speed of the object at the relatively longer distance by subsequently processing each frame using a network, with each input range processed using a corresponding convolutional neural network; and combining the estimation of the estimating the self speed with the estimation of the estimating the relative speed to obtain an estimation of the absolute speed.