Semantic Ground Navigation for Autonomous Vehicles Beyond GPS
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous ground vehicles face challenges in navigating accurately, especially in environments with obstructions and dense urban areas, where GPS signals can be unreliable or insufficient for precise positioning and path determination.
Innovation Solution
The implementation of semantic navigation systems that utilize imaging data to recognize ground features, generate semantic maps, and adjust travel paths accordingly, allowing the vehicle to select courses and speeds based on surface types and obstacles, enhancing navigation precision beyond GPS limitations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS satellites and receivers are used to determine position and navigate, then the autonomous ground vehicle can determine its location and path, but the position information may be inaccurate, irrelevant or unavailable in environments with obstructions or dense urban areas
Solution Approach 1:
The patent introduces semantic maps as an intermediary system that bridges the gap between GPS positioning and actual navigation. When GPS signals are unreliable or unavailable, the vehicle uses semantic maps containing ground feature information (sidewalks, roads, paths) to determine its position and planned path, effectively mediating the navigation function between satellite-based positioning and ground-based movement
Solution Approach 2:
The navigation system is designed to perform multiple functions: it can use GPS positioning when available, switch to semantic map-based navigation when GPS is unreliable, and continuously update its understanding of the environment. This multi-functional approach ensures reliable navigation across diverse conditions, from open areas with clear GPS signals to dense urban environments with obstructions
2Ease of operation
If GPS positioning is used to determine paths and routes, then the vehicle can navigate between locations, but it cannot determine whether paths are clear or obstructed or identify surface types
Solution Approach 1:
The system performs preliminary actions by capturing images and generating semantic maps in advance of actual navigation needs. As the vehicle moves through an environment, it continuously captures images and processes them to create semantic maps that identify ground features (sidewalks, roads, paths) and potential obstacles before the vehicle encounters them, enabling proactive path planning and avoidance
Solution Approach 2:
The patent replaces the mechanical/GPS-based path determination system with an image-processing-based system. Instead of relying solely on satellite positioning and pre-programmed routes, the vehicle uses imaging devices to capture visual information, processes these images to identify ground features, and uses this information to determine and adjust its path in real-time
Data Source
AI summary
Autonomous ground vehicles capture images during operation, and process the images to recognize ground surfaces or features within their vicinity, such as by providing the images to a segmentation network trained to recognize the ground surfaces or features. Semantic maps of the ground surfaces or features are generated from the processed images. A point on a semantic map is selected, and the autonomous ground vehicle is instructed to travel to a location corresponding to the selected point. The point is selected in accordance with one or more goals, such as to maintain the autonomous ground vehicle at a selected distance from a roadway or other hazardous surface, or along a centerline of a sidewalk.