Marker-Encoded LiDAR Detection for Traffic Participant Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current sensing systems for autonomous vehicles lack a unified solution for accurate and reliable perception of complex traffic environments, requiring multiple sensing technologies with different advantages and disadvantages, and face challenges in data handling, communication, and energy efficiency.
Innovation Solution
A LIDAR system that emits pulses towards objects with modulated information, using markers to enhance object detection and classification, and an apparatus for communicating with the LIDAR system to output human-readable signals, integrated with a garment or vehicle for improved traffic participant detection and communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensing technologies are used to achieve accurate perception of complex traffic environments, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system segments the sensing function by using markers on traffic participants that contain encoded information. Instead of using complex multi-sensor systems to extract all information, the markers pre-encode object classification data, allowing the LIDAR system to focus on distance measurement while the markers provide classification information, thus reducing overall system complexity while maintaining precision.
Solution Approach 2:
Markers serve as intermediaries between traffic participants and the LIDAR system. These markers contain encoded information about object classification and reflect LIDAR pulses back with modulated signals, enabling the system to obtain both distance and classification data through a single sensing channel rather than requiring multiple specialized sensors.
2Measurement precision
If markers with modulated information are used to enhance object detection, then measurement precision is improved, but loss of information increases due to modulation complexity
Solution Approach 1:
The markers use periodic modulation of reflected LIDAR pulses to encode information. By modulating the reflected signal at specific frequencies or patterns corresponding to different object classes, the system can distinguish between different traffic participants while maintaining reliable detection through the periodic nature of the modulation, which is easier to decode and less prone to information loss.
3Productivity
If LIDAR pulses are modulated with information towards markers, then productivity is improved through faster data acquisition, but use of energy increases due to modulation requirements
Solution Approach 1:
The markers on traffic participants passively modulate the reflected LIDAR pulses without requiring active energy consumption from the traffic participants themselves. The markers encode information by selectively reflecting or absorbing LIDAR pulses in patterns that indicate object classification, allowing the LIDAR system to acquire classified data without the markers consuming additional energy beyond what is already reflected from the incident pulses.
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
Enhances object detection and classification accuracy, improves communication with traffic participants, and optimizes energy usage in autonomous driving systems, addressing the complexity of traffic environments and data handling challenges.
Implementation Method 1
a distance measuring unit configured to emit a plurality of first pulses towards an object located in a field of view (FOV)... and a detector configured to receive at least one second pulse from the one or more markers of the object
Implementation Method 2
each of the at least one second pulse is configured with a particular wavelength which represents an object class of the object... the object information is modulated on the at least one second pulse
Data Source
AI summary
The present disclosure provides a light detection and ranging (LIDAR) system, which comprises: a distance measuring unit configured to emit a plurality of first pulses towards an object located in a field of view (FOV), wherein the object is associated with one or more markers; and a detector configured to receive at least one second pulse from the one or more markers of the object, wherein each of the at least one second pulse indicates object information identifying the object.