Robot Navigation Using Semantic Cost Maps in Crowded Spaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Navigating robots in crowded and dynamic environments is challenging due to noisy human predictions, leading to unnatural robot movements and difficulty adapting to dynamic situations.
Innovation Solution
A method and system that detect on-site reference features, identify objects of interest, derive semantic data, generate a semantic cost map, and navigate the robot based on this map, enabling adaptive navigation in dynamic environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the robot uses traditional navigation methods in crowded environments, then it can maintain basic movement capability, but the robot freezes or stops due to noisy human predictions, resulting in unnatural movement and inability to adapt to dynamic situations
Solution Approach 1:
The patent transforms the navigation approach by changing from geometric cost parameters to semantic cost parameters. The semantic cost map assigns different cost values based on object categories (e.g., pedestrians, vehicles, buildings) rather than uniform geometric obstacles, allowing the robot to adapt its navigation behavior to the semantic meaning of environment elements while maintaining reliable path planning in crowded settings
Solution Approach 2:
The patent introduces a semantic cost map as an intermediary layer between the robot's navigation system and the real environment. This semantic cost map processes and interprets environmental data, transforming raw sensor inputs into meaningful cost representations that enable the robot to navigate dynamically adapting to crowded environments without freezing or stopping
2Reliability
If the robot navigates through crowded environments with high safety standards, then collision risk is reduced, but the robot's movement efficiency decreases due to frequent stopping and retrying
Solution Approach 1:
The patent changes the cost calculation parameters from binary obstacle detection to multi-category semantic cost assessment. By assigning different cost weights to different object types (e.g., lower cost for stationary objects, higher cost for moving pedestrians), the system achieves both high safety through proper risk assessment and high efficiency through continuous adaptive movement without frequent stops
Solution Approach 2:
The patent implements dynamic navigation by continuously updating the semantic cost map based on real-time environmental changes. The robot's velocity and path are dynamically adjusted according to the current semantic cost landscape, enabling it to navigate through crowded environments with both safety and efficiency by adapting its behavior to changing conditions rather than stopping and retrying
3Productivity
If the robot uses simple geometric cost mapping, then the navigation system remains simple and fast, but it cannot differentiate between different types of objects, leading to suboptimal path planning in complex environments
Solution Approach 1:
The patent enhances the navigation system by changing from simple geometric cost parameters to semantic cost parameters that incorporate object classification and categorization. This allows the system to differentiate between various object types (pedestrians, vehicles, buildings, etc.) and assign appropriate cost values, improving path planning efficiency in complex environments while managing system complexity through modular semantic processing
Data Source
AI summary
A method and system for navigating a robot 100 are provided herein. In an embodiment, the method comprises: detecting on-site reference features of the robot's location when the robot is traversing in an environment; identifying objects of interest 180 from the detected on-site reference features; deriving a semantic data for each object of interest 180 from the on-site reference features; generating a semantic cost map 166 based on the semantic data of the objects of interest 180, the semantic cost map 166 representing cost of traversing in the environment; and navigating the robot 100 based on the semantic cost map 166.