Compressive Active Range Sampling for Random-Access LiDAR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The implementation of random access scanning LiDAR sensors for object detection in autonomous and driver-assisted transportation vehicles is hindered by the need for continuous algorithm development in live environments and the lack of effective methods to determine optimal scan patterns in dynamic scenes.
Innovation Solution
A system and algorithm that schedules laser pulse emissions based on calculated expected potential range changes, prioritizing rays with larger expected range changes for more frequent sampling, thereby providing compressive active range scanning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If continuous scanning of all rays is performed, then complete scene coverage is achieved, but energy consumption and processing time increase significantly
Solution Approach 1:
The patent segments the scene into multiple rays and further divides them into priority groups based on expected potential range changes. Instead of treating all rays equally, the system segments them into high-priority rays (with larger expected changes) and low-priority rays (with smaller expected changes), allowing selective scanning of only the most informative rays at any given time.
Solution Approach 2:
The patent implements partial scanning by selecting and scanning only a subset of rays that have the highest expected potential range changes, rather than scanning all rays continuously. This partial action approach focuses computational and energetic resources on the most critical portions of the scene that are most likely to contain dynamic changes.
2Stability of the object's composition
If all rays are sampled at equal intervals, then uniform monitoring is achieved, but latency in detecting dynamic changes increases
Solution Approach 1:
The patent transitions from static equal-interval sampling to dynamic adaptive sampling. The sampling interval for each ray is dynamically adjusted based on its expected potential range change, which is calculated using radial velocity information. Rays with higher expected changes are sampled more frequently, while rays with lower expected changes are sampled less frequently, creating a dynamic and adaptive monitoring strategy.
Solution Approach 2:
The system uses feedback from previously measured range values and radial velocities to predict future range changes and adjust sampling frequencies accordingly. The expected potential range change calculation incorporates feedback from historical data, allowing the system to adaptively prioritize rays that are more likely to exhibit dynamic changes, thereby reducing detection latency for moving objects.
3Measurement precision
If sampling frequency is increased for all rays, then detection accuracy improves, but processing load and energy consumption increase
Solution Approach 1:
The patent applies local quality by assigning different sampling frequencies to different rays based on their local characteristics (expected potential range changes). Instead of using a uniform high sampling frequency for all rays, the system selectively applies high sampling frequency only to rays that are more likely to contain dynamic changes, while using lower sampling frequencies for static or slowly changing regions.
Solution Approach 2:
The patent changes the sampling frequency parameter dynamically for each ray based on calculated expected potential range changes. The sampling frequency is adjusted as a variable parameter rather than being fixed, allowing the system to optimize detection accuracy for dynamic regions while reducing processing load in static regions.
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
Maximizes information throughput in dynamic scenes by focusing on areas with the most significant changes, reducing latency and energy consumption, and enabling efficient object detection.
Implementation Method 1
a sensor mounted to a transportation vehicle for sensing and perception of a surrounding environment in which the transportation vehicle is positioned, the sensor comprising at least one LiDAR sensor
Implementation Method 2
emission of range determining laser pulses by the at least one random access scanning LiDAR sensor... range samples returned by the at least one random access scanning LiDAR sensor in response to the emitted range determining laser pulses
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system, methodologies and components utilizing a random access scanning LiDAR sensor for object detection for use in autonomous and driver assisted transportation vehicles, wherein calculated expected potential range changes determined based on emission of range determining laser pulses are used to schedule subsequent emission of range determining laser pulses such that laser pulse generation and ray sample generation for individual rays with a larger expected potential range change is performed more frequently than laser pulse generation and ray sample generation for individual rays with a smaller expected potential change so as to provide compressive active range scanning.