Obstacle Cost Map Expansion for Safer Route Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current obstacle avoidance systems for autonomous vehicles are prone to errors due to inaccuracies in obstacle detection and route planning, leading to potential collisions, especially when encountering dynamic obstacles, as they rely on binary maps that do not effectively expand obstacle costs to ensure safe distances.
Innovation Solution
A method and apparatus for generating an obstacle map by expanding obstacle costs using a binary map model that sets obstacle costs for neighboring positions based on probability, employing functions like sigmoid, linear, or Gaussian to decrease costs away from obstacles, and considering dynamic obstacles' motion vectors to calculate estimated collision areas and set corresponding costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a binary map is used to indicate obstacle positions, then the obstacle detection is simple and fast, but the route planning is unsafe due to insufficient obstacle cost expansion
Solution Approach 1:
The patent applies local quality by assigning different obstacle cost values to different regions around the obstacle. The binary map identifies obstacle positions with value 1, while surrounding areas receive graduated cost values (0.5, 0.25, 0.125, etc.) that decrease with distance. This creates locally differentiated cost structures that guide the mobile platform to maintain safe distances from obstacles while allowing flexible path planning.
Solution Approach 2:
The patent implements preliminary action by pre-calculating and storing obstacle cost values for all positions around detected obstacles before route planning occurs. The obstacle map generation unit computes the expanded cost map in advance, so that when the path planning unit needs to find a safe route, the cost information is already prepared, enabling fast and safe route selection without real-time computation delays.
2Reliability
If the obstacle cost is expanded to ensure safe distance, then the route safety is improved, but the computation time increases
Solution Approach 1:
The patent applies periodic action by using a systematic, repeating pattern for obstacle cost expansion. Instead of complex real-time calculations, the system periodically expands obstacle costs using a fixed algorithm that assigns predetermined values at predetermined intervals around each obstacle. This periodic structure allows for efficient computation while maintaining comprehensive safety coverage.
Solution Approach 2:
The patent implements parameter changes by transforming the binary obstacle detection (0 or 1) into a continuous cost field with multiple graduated values. The obstacle cost parameter is changed from a simple binary indicator to a graduated scale (1, 0.5, 0.25, 0.125, 0.0625, 0.03125), allowing the system to balance safety requirements with computational efficiency by adjusting the granularity of cost values.
3Manufacturing precision
If the obstacle cost is set to uniform values around the obstacle, then the implementation is simple, but the route precision is reduced due to lack of directional information
Solution Approach 1:
The patent applies asymmetry by creating an asymmetric obstacle cost distribution that reflects the directional nature of safe navigation. Rather than uniform circular expansion, the cost values are assigned in a manner that creates asymmetric cost patterns around obstacles, allowing the path planning to naturally avoid obstacles from multiple directions while maintaining precision in route generation. The asymmetric cost field provides directional guidance to the mobile platform.
Solution Approach 2:
The patent implements another dimension by adding a cost magnitude dimension to the traditional binary obstacle map. Instead of only indicating presence/absence of obstacles in two dimensions (x, y), the system adds a third dimension of cost value, creating a 3D cost field that provides richer information for route planning while maintaining computational tractability through systematic value assignment.
Data Source
AI summary
A method for generating an obstacle map is provided. The method includes: receiving an environment map; generating a binary map indicating a position of an obstacle by a value of an obstacle cost corresponding to a probability of presence of the obstacle at the position, based on obstacle information obtained from the environment map; and generating an obstacle map by applying an obstacle expansion model to the binary map, wherein the applying the obstacle expansion model comprises setting a plurality of obstacle costs of a plurality of neighboring positions of the position of the obstacle in the binary map to a plurality of values, respectively. Accordingly, a safer and stable route can be generated to avoid the obstacle.