Self-Driving Car Trajectory Graph Prioritization for Faster Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-Driving Cars face challenges in efficiently generating trajectory data due to the computationally expensive process of creating graph-structures for road segments, which can delay corrective actions in response to detected objects, such as pedestrians, due to the need for extensive computations and resource allocation.
Innovation Solution
The method involves a preliminary phase to initialize a graph-structure and a main iterative phase where edges are ranked by priority, with attributes generated iteratively starting from the highest priority edge until a predetermined limit is met, reducing the computational burden and focusing on high-priority edges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a complete graph-structure is generated for all potential positions and transitions, then the trajectory data is comprehensive and accurate, but the computational time and resources increase significantly
Solution Approach 1:
The patent segments the graph-structure generation process into two distinct phases: a preliminary phase that generates a complete graph-structure with all potential positions and transitions, and a main iterative phase that selectively refines only the necessary portions. This segmentation allows the system to maintain comprehensive trajectory data while reducing computational overhead by avoiding redundant processing of all graph elements in every planning cycle.
Solution Approach 2:
The patent performs preliminary action by generating the complete graph-structure in advance during the preliminary phase, before the main iterative phase begins. This pre-computed graph-structure serves as a foundation that can be quickly refined and reused across multiple planning cycles, eliminating the need to regenerate the entire structure from scratch each time and thus significantly reducing computational time while maintaining accuracy.
2Reliability
If all edges in the graph-structure are processed equally, then the trajectory planning is thorough, but the computational resources are wasted on low-priority edges
Solution Approach 1:
The patent applies local quality by differentiating the processing quality and depth applied to different edges based on their priority. High-priority edges (those more likely to be part of the optimal path) receive thorough processing with detailed attribute generation, while low-priority edges receive minimal or no processing. This selective approach ensures that computational resources are focused on the most critical portions of the graph-structure, maintaining trajectory planning reliability while dramatically improving computational efficiency.
Solution Approach 2:
The patent changes the processing parameters dynamically based on edge priority. The main iterative phase adjusts the depth and intensity of attribute generation for each edge according to its priority ranking, rather than applying a uniform processing level to all edges. This parameter adjustment allows the system to adapt computational effort to the actual needs of each graph element, optimizing the balance between thoroughness and efficiency.
3Measurement precision
If the graph-structure generation is performed completely before each planning cycle, then the trajectory data is up-to-date, but the response time to detected objects is delayed
Solution Approach 1:
The patent performs the computationally intensive portion of graph-structure generation in advance during the preliminary phase, creating a ready-to-use foundation before the main iterative phase begins. This allows the system to have updated trajectory data available much faster than if complete regeneration were required, as the preliminary graph-structure is already computed and can be quickly refined with current sensor data and planning constraints.
Solution Approach 2:
The patent applies partial action by generating only the necessary portions of the graph-structure in the main iterative phase, rather than completely regenerating everything. The system processes edges iteratively until a termination condition is met (such as finding a sufficient number of viable paths or reaching a time threshold), which allows it to obtain sufficiently up-to-date trajectory data without the full computational cost of complete regeneration, thus improving response speed while maintaining data currency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods (900) and devices for operating a Self-Driving Car (SDC) (220) are disclosed. The method (900) includes generating a first graph-structure (450) having nodes and edges, ranking the edges based on a priority logic into a ranked list of edges, and generating a second graph-structure (i) by iteratively generating attributes for respective ones from the ranked list of edges beginning with a highest priority edge in the ranked list of edges and (ii) until a pre-determined limit is met. The method (900) also includes causing operation of the SDC (220) on the road segment using the second graph-structure.