Adaptive Trajectory Point Density for Autonomous Vehicle Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicle routing systems are computationally intensive and may not provide safe or comfortable routes, often requiring significant processing power and failing to accurately account for dynamic environments and obstacles.
Innovation Solution
The system adaptively scales the density of trajectory points based on activity levels, using higher densities in high-activity areas and lower densities in low-activity areas, and adjusts weights associated with costs to enhance safety and comfort, with regions around objects dynamically adjusted based on classification and vehicle velocity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If uniform high density of trajectory points is used throughout the route, then route accuracy and safety are improved, but computational burden increases significantly
Solution Approach 1:
The patent applies local quality by varying the density of trajectory points based on the activity level of different route segments. High-activity regions (with obstacles, curvatures, or dynamic objects) use higher point densities for accurate cost evaluation, while low-activity regions use lower densities to reduce computational load. This localized adaptation resolves the contradiction between overall route accuracy and computational efficiency.
Solution Approach 2:
The patent implements dynamics by making the trajectory point density adaptive rather than static. The system dynamically adjusts the density of points along the reference trajectory based on real-time assessment of activity levels in different regions, allowing the computational resources to be allocated flexibly according to the actual environmental conditions and safety requirements.
2Reliability
If uniform high density of trajectory points is used throughout the route, then safety is improved, but processing time increases
Solution Approach 1:
The patent applies local quality by concentrating computational resources (higher point density) only in regions where safety is critical, such as areas with obstacles, sharp curvatures, or dynamic objects. Low-risk regions use lower point density, reducing processing time while maintaining adequate safety monitoring. This localized safety approach resolves the contradiction between overall safety and processing time.
Solution Approach 2:
The patent applies partial action by evaluating costs at fewer points in low-activity regions while maintaining thorough evaluation in high-activity regions. This selective approach ensures safety where needed without unnecessarily processing all points uniformly, thereby reducing overall processing time while maintaining safety standards.
3Device complexity
If fixed region sizes are used around objects, then system complexity is reduced, but adaptability to different objects and velocities decreases
Solution Approach 1:
The patent implements dynamics by making region sizes adaptive based on object classification and vehicle velocity. Different object types (pedestrians, vehicles, cyclists) receive different region sizes, and these sizes further adjust according to the vehicle's velocity. This dynamic adaptation improves safety and accuracy without requiring a completely complex system, as the adjustments are based on simple classification criteria.
Solution Approach 2:
The patent applies parameter changes by modifying region size parameters according to object classification and velocity. Instead of using a fixed region size, the system changes the size parameter dynamically based on the specific object being tracked and the current velocity, allowing the system to adapt to different scenarios while maintaining a relatively simple underlying framework.
Data Source
AI summary
Techniques for generating trajectories and drivable areas for navigating a vehicle in an environment are discussed herein. The techniques can include receiving a reference trajectory representing an initial trajectory for a vehicle, such as an autonomous vehicle, to traverse the environment. Portions of the reference trajectory can be identified as corresponding to actions to navigate around a double-parked vehicle or to change lanes, for example. In some cases, a portion of the reference trajectory can be identified based on a proximity to an object in the environment. A weight can be associated with the portions of the reference trajectory, and the techniques can include evaluating a reference cost function at points of the reference trajectory based on the associated weights to generate a target trajectory. Further, the techniques can include controlling the autonomous vehicle to traverse the environment based at least in part on the target trajectory.