Semantic Map Probability Grids for Sensor-Compatible Autonomous Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles require high-definition maps that are updated in real-time to accommodate changing road conditions, and existing navigation systems face challenges in compatibility due to varying sensor quality among manufacturers.
Innovation Solution
A navigation system that captures a semantic frame from sensor data, corrects pose errors, establishes probability values for grids, generates a semantic grid map, and transmits it for display on a device, using a control circuit and communication circuit for network transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detailed high definition maps are generated to display every aspect of roadway, then map detail and accuracy are improved, but the complexity of map generation and data processing increases
Solution Approach 1:
The patent segments the high definition map into multiple layers including basic map data, semantic information layers, and probability value layers. Each layer represents specific aspects of roadway information (e.g., lane markings, traffic signs, obstacles) allowing complex map data to be organized and processed in manageable segments, thereby improving map detail accuracy while managing generation complexity through structured data organization
Solution Approach 2:
The patent adds a probability value dimension to semantic map data, transforming conventional binary presence/absence information into multi-dimensional probability representations. This allows the system to capture uncertainty and confidence levels in semantic interpretations, enhancing map accuracy by representing real-world ambiguity while maintaining systematic processing through defined probability thresholds
2Reliability
If real-time updates are implemented to capture roadway changes, then map currency and relevance are improved, but the data processing load and system resource requirements increase
Solution Approach 1:
The patent implements preliminary action by pre-processing sensor data into structured semantic frames with defined formats and probability values before integration into the high definition map. This preprocessing step organizes raw sensor information into standardized representations, enabling more efficient real-time updates without proportionally increasing processing energy requirements
Solution Approach 2:
The system employs periodic action through scheduled map updates at defined intervals rather than continuous real-time processing. Semantic frames are captured and integrated at periodic intervals, allowing the system to maintain map currency while managing energy consumption through controlled update frequencies rather than constant processing
3Adaptability or versatility
If sensor data from multiple manufacturers with varying quality is integrated, then system versatility and data coverage are improved, but data compatibility and fusion accuracy deteriorate
Solution Approach 1:
The patent implements universality through a standardized semantic frame format that can accommodate data from multiple sensor types and manufacturers. The semantic frame structure defines universal fields for representing roadway features, allowing diverse sensor inputs to be normalized into a common representation, thereby improving sensor compatibility while maintaining data fusion accuracy through consistent data structures
Solution Approach 2:
The system applies parameter changes by transforming raw sensor data into standardized semantic parameters with defined probability value ranges. This parameter normalization process converts varied sensor qualities into uniform probability representations, enabling accurate data fusion across different manufacturers by focusing on the confidence level of semantic interpretations rather than raw sensor characteristics
Data Source
AI summary
A navigation system includes: a control circuit configured to: a control circuit configured to: capture a semantic frame from a sensor data stream for a region of interest, crop an outer peripheral edge of the semantic frame by removing semantic points beyond the outer peripheral edge, calculate a world coordinate point cloud by calculating a global positioning system (GPS) location of the semantic points includes a pose error corrected, establish a probability value for a base-level grid, generate a semantic grid map by identifying a dominant percentage of the probability value of the base-level grid and an upper-level grid, and generate a semantic map from the semantic grid map to represent the region of interest; and; and a communication circuit, coupled to the control circuit, configured to transmit the semantic map for displaying on a device.