Multi-Wavelength LIDAR Object Recognition for Vehicle Safety
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object recognition systems in vehicles face limitations in accurately and rapidly determining obstacles ahead, such as pedestrians or vehicles, using existing camera and range sensor methods.
Innovation Solution
A multi-wavelength LIDAR sensor system that transmits and receives laser light of different wavelengths, utilizing a processing unit with a learning machine and classifier to classify objects based on features like size, speed, reflectance, number plates, and material, increasing weights for accurate recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If camera or range sensor methods are used for object recognition, then the system can detect objects ahead of the vehicle, but the recognition accuracy and speed are insufficient for reliable obstacle determination
Solution Approach 1:
The patent segments the object recognition task into multiple independent analysis dimensions: size classification, speed measurement, reflector detection, number plate recognition, and material identification. Each dimension is processed separately by dedicated units, and the results are combined to achieve high-reliability object classification. This segmentation allows parallel processing and reduces the complexity of each individual analysis task.
Solution Approach 2:
The patent utilizes multi-wavelength laser illumination (different wavelengths) to change the physical parameters of light interaction with objects. By varying the wavelength parameters, the system extracts different reflectance characteristics from the same object, enabling more accurate material identification and object classification. This parameter change approach transforms a single-dimensional detection into multi-dimensional analysis.
2Productivity
If existing sensor methods are used, then object detection can be performed, but rapid and accurate determination of obstacle type cannot be achieved
Solution Approach 1:
The patent performs preliminary classification of objects based on easily measurable parameters such as size and speed before conducting more complex analysis. The learning machine pre-stores feature information for different object types, allowing the system to quickly compare incoming data against known patterns and make rapid preliminary determinations. This preliminary action reduces the time required for full classification while maintaining reliability.
Solution Approach 2:
The system incorporates feedback mechanisms where the learning machine continuously learns from classified objects and updates its feature information database. The classification results feed back into the system to refine future classifications, improving both speed and reliability over time. This feedback loop allows the system to adapt to new object types and improve its determination accuracy progressively.
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 recognition performance by accurately classifying vehicles, two-wheeled vehicles, pedestrians, and unidentified objects, improving safety by providing reliable obstacle detection.
Implementation Method 1
a receiving unit configured to receive light reflected from the object to acquire information on the reflected light
Implementation Method 2
a multi-wavelength LIDAR sensor system that transmits and receives laser light of different wavelengths
Data Source
AI summary
Provided is an object recognition system including: a transmitting unit configured to include one or more light emitting unit to transmit laser light having different wavelengths to an object; a receiving unit configured to receive light reflected from the object to acquire information on the reflected light of a size, a speed, a number plate, a reflector, and material of the object; and a processing unit configured to store the feature information on objects of a vehicle, a two-wheeled vehicle, and a pedestrian and compare the information on the reflected light of the object received from the receiving unit with the feature information on the object to classify and recognize the objects into the vehicle, the two-wheeled vehicle, and the pedestrian.