Autonomous Navigation Path Generation Using Hierarchical Collision Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous navigation systems for vehicles face instability in path generation due to inaccurate prediction of peripheral vehicle movement, leading to increased collision risk, especially when the predicted path of a moving object is incorrect.
Innovation Solution
A hierarchical collision detection scheme that predicts the movement of a target vehicle based on its size, position, departure angle, and speed, using Vehicle-to-everything (V2X) communication and sensors, and generates a stable navigation path by calculating a collision risk value and determining lane change intentions through probabilistic models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a hierarchical collision detection scheme is used to detect obstacles and generate paths, then the system can operate in real-time with improved processing efficiency, but the accuracy of collision detection deteriorates when the predicted path of a moving object is incorrect
Solution Approach 1:
The system performs preliminary path prediction for target vehicles before collision detection. By estimating future positions of moving objects in advance and incorporating these predictions into the collision detection process, the system maintains both real-time operation and improved accuracy in detecting potential collisions with dynamic objects.
2Device complexity
If the autonomous navigation system generates a traveling path without predicting peripheral vehicle movement, then the path generation process is simpler and faster, but the stability of path generation deteriorates and collision risk increases
Solution Approach 1:
The system incorporates key movement parameters (position, speed, departure angle, size) of peripheral vehicles into the path generation process. By utilizing these parameters to predict target vehicle trajectories and integrate them with the ego-vehicle's path planning, the system achieves stable and reliable path generation that accounts for dynamic environmental factors without excessive complexity.
Data Source
AI summary
A system and method for performing autonomous navigation of a vehicle are disclosed, which recognize a possibility of collision risk through path estimation of a target vehicle, and generate an autonomous navigation path. The autonomous navigation system for a vehicle includes: a target-vehicle information detection unit configured to detect traveling information of a target vehicle; and a path generation unit configured to estimate movement of the target vehicle on the basis of the information received from the target-vehicle information detection unit, calculate a collision risk value by recognizing a possibility of collision risk between an ego-vehicle and the target vehicle, and generate an autonomous navigation path of the ego-vehicle.