Autonomous Vehicle Behavioral Planning With Fixed Action Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles require continuous behavioral and trajectory planning to navigate effectively from a starting point to a destination, but existing methods lack efficient filtering mechanisms to maintain a fixed set of actions and trajectories, leading to resource wastage and infeasible trajectory generation.
Innovation Solution
A method and system for behavioral planning in autonomous vehicles that involves generating a set of actions and trajectories, applying filters to remove inapplicable actions and trajectories while maintaining a fixed size and order, and optimizing trajectory planning using a cost function to determine optimal time durations for vehicle control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a fixed size and fixed order set of actions is generated according to a predefined methodology, then the behavioral planning structure is simplified and processing efficiency is improved, but the ability to adapt to diverse driving scenarios is reduced
Solution Approach 1:
The patent applies dynamics by making the action set adaptable through masking mechanisms. While the base action set maintains fixed size and order for efficiency, the masking approach dynamically enables or disables specific actions based on driving context, allowing the system to adapt to diverse scenarios without sacrificing structural efficiency.
Solution Approach 2:
The patent changes parameters by introducing urgency levels and terminal velocities as variable parameters that modify the fixed action set. This allows the same structured framework to generate varied actions by adjusting these parameters, thereby maintaining processing efficiency while improving adaptability to different driving conditions.
2Reliability
If trajectory filtering is applied to assess path and velocity profile, then the reliability of trajectory selection is improved, but the computational time and resource usage increase
Solution Approach 1:
The patent applies preliminary action by pre-defining the structure of action sets and trajectory generation methodologies before actual planning occurs. This preprocessing establishes a framework that reduces the scope of filtering needed during real-time operation, maintaining reliability while reducing computational time through structured pre-planning.
Solution Approach 2:
The patent applies local quality by focusing filtering efforts on specific critical aspects of trajectories (path and velocity profile) rather than进行全面 evaluation. This targeted approach maintains reliability for the most important trajectory characteristics while reducing overall computational burden by not filtering every possible parameter.
3Device complexity
If action filtering is applied to identify inapplicable actions, then the number of trajectories to be generated is reduced, but the complexity of the filtering logic increases
Solution Approach 1:
The patent applies segmentation by dividing the action set into manageable categories (dynamic actions from nine cells, fixed actions from three lanes) that can be independently filtered. This segmentation makes the filtering logic more organized and easier to implement by breaking down the complex filtering task into smaller, more manageable segments based on action type and spatial location.
Data Source
AI summary
Systems and methods to perform behavioral planning in an autonomous vehicle from a reference state involve generating a set of actions of a fixed size and fixed order according to a predefined methodology. Each action is a semantic instruction for a next motion of the vehicle. A set of trajectories is generated from the set of actions as an instruction indicating a path and a velocity profile to generate steering angles and accelerations for implementation by the vehicle. A trajectory filter is applied to filter the set of trajectories such that unfiltered trajectories are candidate trajectories. Applying the trajectory filter includes assessing the path and velocity profile indicated by each of the set of trajectories. A selected trajectory is used to control the vehicle or the action that corresponds to the selected trajectory is used in trajectory planning to generate a final trajectory that is used to control the vehicle.