Autonomous Path Perception Diversity for Real-Time Reliability Checks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional autonomous vehicle systems rely on a single data source for path perception, which is impractical for real-time accuracy assessment and can lead to system failures, especially in challenging scenarios, due to the lack of redundancy and diversity in path perception inputs.
Innovation Solution
Implementing a system that utilizes a plurality of input signals from deep neural networks, high-definition maps, and object traces to generate a diverse and redundant understanding of the driving surface, allowing for real-time assessment of path perception quality and reliability through ensemble methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single data source is used for path perception, then device complexity is reduced, but reliability deteriorates due to lack of redundancy and inability to perform real-time accuracy assessment
Solution Approach 1:
The patent combines multiple path perception data sources including deep neural network outputs, high-definition map data, and object trace information into a unified ensemble assessment framework. This merging allows the system to evaluate path perception accuracy by comparing multiple independent sources simultaneously, thereby improving reliability without requiring excessive computational complexity.
Solution Approach 2:
The patent creates multiple copies of path perception computations through different data sources and methods (e.g., multiple DNNs, HD map references, object trace analyses). By generating redundant perception results through these copies, the system can perform real-time accuracy assessment and select the most reliable path information, resolving the contradiction between simplicity and reliability.
2Reliability
If multiple input signals from diverse sources are used, then reliability and accuracy are improved, but device complexity increases
Solution Approach 1:
The patent segments the path perception system into distinct functional modules: deep neural network processing, high-definition map integration, object trace analysis, and ensemble aggregation. Each module handles a specific data source independently, allowing the system to leverage diverse inputs for improved reliability while managing complexity through modular architecture and specialized processing for each data type.
3Device complexity
If a single path perception signal is used, then device complexity is reduced, but measurement precision deteriorates due to inability to verify accuracy in real-time
Solution Approach 1:
The patent implements a feedback mechanism where the ensemble of path perception sources continuously evaluates and compares their outputs in real-time. By aggregating results from multiple independent assessments and using this feedback to determine the most accurate path, the system achieves verified measurement precision without requiring impractical computational resources, as the feedback loop efficiently synthesizes information from existing data sources.
4Measurement precision
If diverse path perception inputs are aggregated, then path perception quality is enhanced, but computing requirements increase
Solution Approach 1:
The patent applies partial action by selectively aggregating path perception results from diverse sources based on their individual confidence levels and reliability metrics. Rather than processing all possible data sources equally, the system performs ensemble assessment on a subset of the most reliable sources, achieving high path perception quality while avoiding the computational overhead of processing every available input signal in full detail.
Data Source
AI summary
In various examples, a path perception ensemble is used to produce a more accurate and reliable understanding of a driving surface and/or a path there through. For example, an analysis of a plurality of path perception inputs provides testability and reliability for accurate and redundant lane mapping and/or path planning in real-time or near real-time. By incorporating a plurality of separate path perception computations, a means of metricizing path perception correctness, quality, and reliability is provided by analyzing whether and how much the individual path perception signals agree or disagree. By implementing this approach—where individual path perception inputs fail in almost independent ways—a system failure is less statistically likely. In addition, with diversity and redundancy in path perception, comfortable lane keeping on high curvature roads, under severe road conditions, and/or at complex intersections, as well as autonomous negotiation of turns at intersections, may be enabled.