Autonomous Vehicle Topological Planning for Dense Traffic Lane Changes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicle trajectory planning methods often fail to plan successful lane change maneuvers in dense traffic conditions due to limitations in computational resources and reliance on heuristics, which may not account for complex maneuvers required in such scenarios.
Innovation Solution
The system employs topological planning to identify distinct classes of trajectories, computes constraint sets for each class, optimizes trajectories, and selects the best candidate trajectory for the autonomous vehicle to traverse a local region, incorporating real-time information and biasing for temporal consistency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If prior art methods consider only a finite number of lane change variations based on heuristics, then computational resources are conserved, but the system fails to plan successful lane change maneuvers in dense traffic conditions
Solution Approach 1:
The patent segments the continuous trajectory planning problem into discrete topological classes based on lane change decisions (stay in current lane, change to left lane, change to right lane). This segmentation transforms the infinitely high-dimensional search space into a manageable set of discrete categories, allowing the system to evaluate multiple trajectories within each class without being overwhelmed by computational complexity.
Solution Approach 2:
The patent introduces a topological dimension to the trajectory planning problem by classifying trajectories based on their discrete lane change actions. This adds a categorical layer to the continuous spatial-temporal planning, enabling the system to systematically explore different maneuver types while maintaining computational tractability through structured evaluation of each topological class.
2Reliability
If prior art methods use sampling-based approaches with finite number of candidate trajectories, then computational resources are conserved, but the system may fail to find a feasible trajectory when one exists or finds a suboptimal trajectory
Solution Approach 1:
The patent performs preliminary topological classification of the planning space into distinct trajectory classes based on lane change decisions before conducting detailed trajectory optimization. This preliminary action structures the search space in advance, ensuring that feasible trajectories are systematically explored across all topological categories rather than relying on random sampling, thereby improving the guarantee of finding feasible solutions.
Solution Approach 2:
The patent changes the parameterization of the planning problem from continuous spatial sampling to discrete topological classification combined with continuous optimization within each class. This parameter transformation allows the system to systematically cover the solution space by varying both the discrete lane change decisions and the continuous trajectory parameters, improving both feasibility and optimality.
3Reliability
If the system generates multiple candidate trajectories through topological planning, then the completeness of trajectory coverage is improved, but the computational time and resources increase
Solution Approach 1:
The patent segments the trajectory generation process into two phases: (1) discrete topological classification to identify distinct maneuver categories, and (2) continuous optimization within each topological class. This segmentation allows the system to generate comprehensive trajectory coverage by systematically exploring each class while avoiding redundant computation across equivalent trajectories, thereby managing computational time efficiently.
Solution Approach 2:
The patent creates a universal framework that handles multiple trajectory types (stay-in-lane, left lane change, right lane change) through a single topological planning structure. This multi-functional approach allows the same optimization algorithms to be applied across different maneuver types, improving computational efficiency while maintaining comprehensive trajectory coverage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods of manoeuvring an autonomous vehicle in a local region using topological planning, while traversing a route to a destination location, are disclosed. The system includes an autonomous vehicle including one or more sensors and a processor. The processor is configured to determine the local region on the route and receive real-time information corresponding to the local region. The processor performs topological planning to identify on or more topologically distinct classes of trajectories, compute a constraint set for each of the one or more topologically distinct classes of trajectories, optimize a trajectory to generate a candidate trajectory for each constraint set, and select a trajectory for the autonomous vehicle to traverse the local region from amongst the one or more candidate trajectories. Each of the one or more topologically distinct classes is associated with a plurality of trajectories that take the same combination of discrete actions with respect to objects in the local region.