Crowd-Flow Robot Localization for Socially Friendly Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation algorithms for robots in dense crowds face challenges in mapping and localization due to occlusion by moving obstacles and lack of social-awareness, leading to inefficient and intimidating robot behavior.
Innovation Solution
The use of crowd-flow maps, reconstructed through local crowd observations and clustering algorithms, allows for social-friendly navigation by leveraging pedestrian movement patterns for mapping, localization, and planning, incorporating flow-matching metrics and resistance costs to guide robot movement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If moving obstacles like pedestrians are filtered out from sensory observation, then mapping and localization accuracy is improved, but rich human-robot interaction information is discarded
Solution Approach 1:
The patent converts the harmful effect of moving obstacles (pedestrians) into a beneficial resource by treating them as informative elements rather than noise. The crowd flow map extracts movement patterns, directions, and densities from pedestrians, transforming what was previously filtered out as interference into valuable social context for navigation decisions.
Solution Approach 2:
Instead of filtering out moving obstacles to improve mapping accuracy, the patent inverts the approach by incorporating moving obstacles (pedestrians) as the primary source of environmental information. The crowd flow map is built from pedestrian movement data, reversing the conventional wisdom that static structures are the only reliable mapping features.
2Reliability
If reactive policies or short-horizon planning are used to avoid collisions, then collision avoidance is improved, but social-friendliness and navigation effectiveness are reduced
Solution Approach 1:
The patent performs preliminary analysis of crowd flow patterns and social norms before making navigation decisions. By pre-processing crowd movement data to identify traversable regions, flow directions, and social conventions (e.g., left-hand vs. right-hand traffic), the robot can plan socially-friendly paths in advance rather than reacting impulsively to individual pedestrians.
Solution Approach 2:
The navigation system dynamically adapts to crowd conditions by continuously updating the crowd flow map and adjusting navigation strategies. The robot modifies its behavior based on real-time crowd density, flow patterns, and social norms, transitioning between different navigation modes (e.g., following crowd flow vs. cutting through open spaces) to maintain both safety and social-friendliness.
Data Source
AI summary
Disclosed are methods of navigating a robot in a crowd involving mapping a local crowd and applying a clustering algorithm that reconstructs a crowd-flow map capturing movement patterns of pedestrians in the local crowd; using a flow-matching metric, localizing the robot in the crowd-flow map; and following a movement pattern in the crowd-flow map using a hierarchical crowd-driven planning on a long time horizon and a short time horizon.