Autonomous Vehicle Trajectory Planning Around Roadway Occlusions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous vehicle trajectory planning approaches are computationally demanding and inefficient, especially in complex or chaotic driving environments with occlusions, which increases the need for computational resources.
Innovation Solution
A trajectory planning system that uses a GPS for global scene information and sensors on the autonomous vehicle to collect local scene information, identifying occluding obstacles and scene actors to generate candidate trajectories based on a contracted occluded portion of the roadway.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current planning approaches are used to handle occluded portions, then trajectory planning can be performed, but computational resource requirements increase significantly
Solution Approach 1:
The approach segments the roadway into occluded and non-occluded portions, allowing the planning system to process only relevant segments. By identifying and isolating occluded portions, the system avoids computationally expensive processing of entire trajectories through occluded areas, thereby reducing computational resource consumption while maintaining planning robustness.
Solution Approach 2:
The system performs preliminary identification of occluded portions before generating candidate trajectories. By pre-processing the environment model to mark occluded areas, the system eliminates the need for complex real-time computational analysis during trajectory evaluation, reducing overall computational resource requirements while ensuring reliable handling of occlusions.
2Measurement precision
If comprehensive scene information is collected to handle occlusions, then trajectory accuracy improves, but computational complexity increases
Solution Approach 1:
The approach extracts and isolates only the critical information related to occluded portions from the comprehensive scene data. By separating occlusion-related parameters from other scene information, the system maintains trajectory planning accuracy while simplifying the processing complexity, as only the extracted occlusion data requires specialized handling.
Solution Approach 2:
The system applies different processing qualities to different portions of the scene: comprehensive processing for non-occluded areas and simplified occlusion-aware processing for occluded portions. This local differentiation maintains high trajectory accuracy where needed while reducing overall system complexity by avoiding uniform complex processing throughout the entire scene.
3Productivity
If traditional trajectory planning is used in chaotic driving environments, then basic navigation is achieved, but scalability to complex environments is limited
Solution Approach 1:
The approach dynamically adjusts the planning strategy based on the detected environment characteristics. When occluded portions are identified, the system automatically switches to occlusion-aware trajectory generation; in clear environments, it uses standard planning. This dynamic adaptation maintains high navigation efficiency across diverse environments while improving scalability to complex chaotic driving scenarios.
Solution Approach 2:
The occlusion-aware trajectory planning system serves multiple functions: it handles both occluded and non-occluded environments using a unified framework, and it adapts to various degrees of environmental complexity. This multi-functionality allows the same system to efficiently navigate from simple to chaotic driving environments, improving both productivity and adaptability simultaneously.
Data Source
AI summary
A system and method for trajectory planning selects a trajectory from a plurality of candidate trajectories to direct an autonomous vehicle. The system includes a controller in communication with a global positioning system (GPS) and at least one sensor. The GPS receives a plurality of global scene information. The at least one sensor collects local scene information. The controller is programmed to receive global scene information and local scene information, identify the location of the occluding obstacle. Additionally, the controller is programmed to determine an occluded portion of the roadway, identify a location of the scene actors on the roadway, and calculate a contracted occluded portion within the occluded portion. Furthermore, the system generates candidate trajectories for the autonomous vehicle to travel along, score the candidate trajectories, and select one of the candidate trajectories based on the assigned score.