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, as they often require high point densities for navigating through environments with obstacles, leading to increased processing burdens and potential safety risks.
Innovation Solution
Adaptive scaling of trajectory point density based on activity levels, with lower densities in low-activity areas and higher densities in high-activity areas, combined with dynamic adjustment of region sizes for drivable areas based on object classification and vehicle velocity, to optimize route planning and safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If high point density is used for trajectory generation, then route safety and accuracy are improved, but computational burden increases
Solution Approach 1:
The patent applies local quality by varying the point density along different portions of the reference trajectory based on local activity levels. High point density is applied to high-activity portions (where objects are present or curvature is high) and low point density to low-activity portions, thereby maintaining safety where needed while reducing overall computational burden.
Solution Approach 2:
The patent implements dynamics by adaptively adjusting the point density in real-time based on the detected activity level along the reference trajectory. The system dynamically determines which portions require higher density and which can use lower density, making the trajectory generation process adaptive rather than static.
2Measurement precision
If high point density is used for trajectory generation, then route accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies local quality by varying the point density along different portions of the reference trajectory based on local activity levels. High point density is applied to high-activity portions (where objects are present or curvature is high) and low point density to low-activity portions, thereby maintaining safety where needed while reducing overall computational burden.
Solution Approach 2:
The patent applies partial action by using high point density only where necessary (in high-activity portions) rather than uniformly across the entire trajectory. This partial application of high density maintains accuracy where needed while avoiding the excessive processing time that would result from applying high density everywhere.
3Reliability
If large region sizes are used for drivable areas, then safety margins are improved, but computational complexity increases
Solution Approach 1:
The patent applies local quality by adjusting the size of drivable area regions based on local conditions such as object classification and vehicle velocity. Larger regions are created where safety margins are more critical, while smaller regions are used in less critical areas, thereby maintaining safety without uniformly increasing computational complexity across the entire environment.
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. A point density can be determined for various portions of the reference trajectory. In some cases, the point density can be based at least in part on a cost associated with a curvature value associated the reference trajectory or a cost associated with a distance between the reference trajectory and an obstacle in the environment. Further, the techniques can include evaluating a cost function at points on the reference trajectory to generate a target trajectory with respect to the reference trajectory, and controlling the autonomous vehicle to traverse the environment based at least in part on the target trajectory.