LiDAR Scan Segment Access for Low-Latency Self-Driving Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-driving vehicles face safety issues due to delays in processing LiDAR data, which can be several cycles old by the time it is used for decision-making, especially in rapidly changing environments, and using faster sensors like cameras introduces errors due to lack of depth information.
Innovation Solution
A control system that allows for immediate access to needed segments of scan data, choosing between using previous or new data based on age and processing delay, ensuring up-to-date information for safer vehicle maneuvering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the LiDAR system uses a default scanning cycle time of 100 ms, then the system can operate at a manageable cost, but the data becomes up to 100 ms old which causes safety issues at high speeds
Solution Approach 1:
The patent divides the scanning cycle into multiple segments or phases, allowing the system to process and provide data from different parts of the scan at different times. This segmentation enables the control system to access the most recent relevant data segments without waiting for the complete 100 ms cycle to finish, thereby reducing the effective data age while maintaining system cost-effectiveness.
Solution Approach 2:
The system performs preliminary processing and availability marking of scan data during the scanning cycle itself. By preparing data in advance and marking it as available before the full cycle completes, the system reduces the delay between data collection and its use in control decisions, improving safety without requiring a faster overall scanning cycle.
2Reliability
If the system waits for new scan data to be collected, then the data is more up-to-date, but the decision-making process is delayed
Solution Approach 1:
The system implements a feedback mechanism where the control system queries the availability of scan data and receives real-time information about which data is ready for use. This feedback loop allows the control system to make informed decisions about whether to wait for newer data or proceed with currently available data, optimizing the balance between data freshness and decision-making speed.
Solution Approach 2:
The system dynamically adjusts its data access strategy based on current conditions. Rather than following a fixed waiting pattern, the system flexibly determines whether to wait for new scans or use previous scans based on the urgency of the situation and the availability of recent data, allowing optimal performance across varying operational contexts.
3Speed
If faster sensors like cameras are used, then the scanning cycle time is reduced, but depth information is lost leading to computational errors
Solution Approach 1:
The patent segments the data requirements into different categories: depth information from LiDAR and 2D visual information from cameras. By segmenting the sensor functions and processing requirements, the system can use camera data for tasks that don't require precise depth measurement while relying on LiDAR for depth-critical decisions, thus leveraging the speed of cameras without sacrificing depth accuracy where needed.
Solution Approach 2:
The system merges data from multiple sensor types (LiDAR and cameras) to achieve both speed and precision. By combining the fast 2D imaging capability of cameras with the accurate depth measurement of LiDAR, the system compensates for the limitations of each individual sensor type, maintaining measurement precision while benefiting from faster overall system response.
Data Source
AI summary
A method for providing scan data for decision making in a self-driving vehicle (200) including a scanning sensor (308) that scans a space around the self-driving vehicle (200) in consecutive scanning cycles. The method includes receiving a request for scan information associated with a requested segment of space, determining whether to wait for a next scan of the requested segment and use scan data from the next scan, or to extract and use scan data from a previous scan of the requested segment. The method also includes processing the scan data to generate the scan information for the requested segment, providing the scan information for the requested segment in response to the request, and operating the self-driving vehicle in accordance with a decision made about a maneuver to be performed by the self-driving vehicle based, at least in part, on the scan information.