Autonomous Vehicle Trajectory Search for Reactive Obstacle Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating accurately and efficiently when roadway indicators are obscured, degraded, or invalidated, and in unpredictable environmental and behavioral scenarios, leading to potential stutters or the need for remote operator intervention.
Innovation Solution
An autonomous vehicle guidance system employing a tree search technique that alternately determines candidate actions and predicts future environment states, classifying dynamic objects as reactive or passive using machine-learned models, and adjusting cost estimates to generate paths that avoid obstacles and ensure safe navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the autonomous vehicle uses traditional navigation methods relying on roadway indicators, then navigation is efficient under normal conditions, but navigation accuracy deteriorates when roadway indicators are obscured, degraded, or invalidated
Solution Approach 1:
The system dynamically adapts its navigation approach by switching from relying on static roadway indicators to using a tree search algorithm that explores multiple candidate paths. When abnormal conditions are detected (obscured, degraded, or invalidated indicators), the system transitions to a more flexible, exploratory navigation mode that can handle unpredictable environmental factors and human/animal behavior variations.
Solution Approach 2:
The system changes its operational parameters by adjusting the search depth and cost estimation methods in the tree search algorithm based on the detected condition of roadway indicators. Under normal conditions, the system uses standard navigation parameters for efficiency. When indicators are compromised, the system increases search thoroughness and modifies cost estimates to prioritize safety and adaptability over speed.
2Reliability
If the autonomous vehicle explores more candidate actions and paths using tree search, then navigation safety and scenario coverage increase, but computational complexity and processing time increase
Solution Approach 1:
The system applies partial exploration by selectively expanding only the most promising candidate paths in the tree search, rather than exhaustively exploring all possible paths. The cost estimation function prioritizes paths that are more likely to lead to safe navigation, allowing the system to achieve adequate safety coverage without the full computational burden of complete path exploration.
Solution Approach 2:
The system replaces exhaustive mechanical path verification with a heuristic-based cost estimation approach. Instead of physically testing or simulating every candidate path, the system uses learned cost functions to evaluate and prune unlikely paths, substituting computational heuristics for more resource-intensive verification methods.
3Object-affected harmful factors
If the autonomous vehicle uses conservative navigation to ensure safety in unpredictable scenarios, then collision avoidance improves, but navigation efficiency and speed decrease
Solution Approach 1:
The system uses feedback from the tree search evaluation process to dynamically adjust navigation decisions. As the tree search evaluates candidate paths and their associated costs, the system receives feedback about which paths offer both safety and efficiency. This feedback loop allows the vehicle to select paths that balance collision avoidance with navigation speed, rather than always choosing the safest but slowest option.
Solution Approach 2:
The system performs preliminary evaluation of multiple candidate paths using the tree search algorithm before committing to a navigation decision. By pre-assessing the safety and efficiency characteristics of various paths ahead of time, the system can make faster, more informed decisions during actual navigation, reducing the need for reactive maneuvers that would decrease overall efficiency.
Data Source
AI summary
Trajectory generation for controlling motion or other behavior of an autonomous vehicle may include alternately determining a candidate action and predicting a future state based on that candidate action. The technique may include determining a cost associated with the candidate action that may include an estimation of a transition cost from a current or former state to a next state of the vehicle. This cost estimate may be a lower bound cost or an upper bound cost and the tree search may alternately apply the lower bound cost or upper bound cost exclusively or according to a ratio or changing ratio. The prediction of the future state may be based at least in part on a machine-learned model's classification of a dynamic object as being a reactive object or a passive object, which may change how the dynamic object is modeled for the prediction.