LiDAR Sensor Window Dirt Detection After Cleaner Residue
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor systems face challenges in accurately detecting dirt on transmissive parts of LiDAR due to residual cleaning liquid, which can lead to erroneous determinations, and are limited by reliance on specific detection targets that may not be consistently available, especially when the vehicle is stationary or in changing landscapes.
Innovation Solution
A sensor system that includes a dirt determining unit to detect dirt on transmissive parts by analyzing point group information, using reflection intensity comparisons over time, and considering vehicle speed or landscape stability to determine dirt attachment accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dirt determination is performed immediately after cleaner operation, then dirt detection responsiveness is improved, but erroneous determination occurs due to remaining cleaning liquid
Solution Approach 1:
The system performs preliminary cleaning operation before dirt determination, and introduces a predetermined time delay after cleaning before performing the dirt determination. This preliminary timing control prevents erroneous detection of cleaning liquid as dirt, while maintaining responsive dirt detection capability.
2Device complexity
If a specific detection target (e.g., sky) is used for dirt determination, then detection simplicity is improved, but adaptability to varying environments deteriorates
Solution Approach 1:
The system uses multiple detection targets including sky, road surface, and other objects within the detection range. The dirt determining unit can select appropriate targets based on vehicle state (traveling/stopped) and environmental conditions, making the detection system universally applicable across various environments rather than relying on a single specific target.
Solution Approach 2:
The system dynamically selects detection targets based on vehicle state and environmental conditions. When the vehicle is traveling, moving targets like road surface are used; when stopped, stationary targets are used. This dynamic adaptation ensures reliable dirt detection across varying environments.
3Measurement precision
If landscape changes are used to detect dirt attachment, then detection capability is improved, but applicability when vehicle is stopped deteriorates
Solution Approach 1:
The system performs periodic dirt determination at different vehicle states (traveling and stopped). During traveling, it uses landscape change detection; during stopped states, it uses reflection intensity comparison with predicted values. This periodic operation at different states ensures continuous dirt detection capability regardless of vehicle motion state.
4Speed
If reflection intensity comparison is performed without time delay, then detection speed is improved, but measurement accuracy deteriorates due to cleaning liquid residue
Solution Approach 1:
The system performs preliminary cleaning operation before dirt determination, and introduces a predetermined time delay after cleaning before performing the reflection intensity comparison. This timing control allows cleaning liquid to evaporate or be removed, ensuring accurate reflection intensity measurement without sacrificing overall system responsiveness.
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
The system effectively reduces erroneous dirt detections by accounting for residual cleaning liquid and allows dirt detection regardless of specific targets, ensuring accurate operation in varying environments and vehicle states.
Implementation Method 1
a light receiving unit configured to receive light emitted from the light emitting unit and reflected by hitting at an object
Implementation Method 2
the dirt determining unit is configured to detect dirt attached to a transmissive part, based on the point group information and the target position information
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A cleaner (40) for cleaning a transmissive portion of a sensor including a light receiving unit that receives light from a detection target via the transmissive portion is not operated by a cleaner control unit (25) within a predetermined time after the drive of the cleaner (40) is complete. A dirt determining unit (12) identifies, as a predicted position, position information of a target object after a first time has elapsed, on the basis of the movement history of the target object at a first time point, and determines that dirt is attached if the reflection intensity at the predicted position, acquired after the first time has elapsed after the first time point, differs from the reflection intensity of the target object at the first time point. The dirt determining unit (11) determines that a vehicle is in a specific location on the basis of distance information relating to reference information and distance information relating to point group information when stopped, and determines that dirt is attached if there is a difference between the distance information relating to the point group information when stopped and the distance information relating to the reference information.