Floor Plan Repair Using Graph Cuts for Robot Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing household robots face challenges in navigating complex environments safely due to limitations in the number and quality of sensors, which affects their ability to make sharp turns, move at higher speeds, and accurately map and understand their surroundings.
Innovation Solution
The development of a floor plan system that uses data from autonomous mobile devices (AMDs) to generate, update, and repair floor plans, allowing for accurate segmentation of room geometries and integration of user-provided hints to refine the mapping, thereby enhancing navigation and safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If consumer-grade robots are equipped with more high-quality sensors to navigate complex environments safely, then navigation safety and mapping accuracy improve, but device complexity and cost increase
Solution Approach 1:
The patent replaces complex sensor-based detection systems with a computational geometry approach. Instead of relying on multiple high-quality sensors to detect and understand the environment, the system uses floor plan data structures and geometric algorithms to represent, reason about, and navigate environments. This substitution of mechanical/sensor systems with computational methods reduces sensor requirements while maintaining navigation safety.
Solution Approach 2:
The patent introduces floor plan data structures as an intermediary representation between the robot and the physical environment. These structured representations (walls, rooms, corridors, obstacles) serve as a mediator that simplifies complex sensor data into manageable geometric models, enabling safe navigation without requiring direct complex sensor processing.
2Measurement precision
If sensors are enhanced to detect obstacles and hazards more accurately, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent replaces sensor-based obstacle detection with computational geometry methods. Instead of using multiple sensors to detect obstacles in real-time, the system represents obstacles as geometric primitives (polygons, lines, points) in floor plan data structures. This allows for precise obstacle representation and collision detection through mathematical computations rather than complex sensor arrays.
3Speed
If the robot is equipped with advanced sensors for faster movement and sharper turns, then speed and maneuverability improve, but reliability of safe operation decreases due to sensor limitations
Solution Approach 1:
The patent performs preliminary computation of safe paths and collision-free trajectories using floor plan data before the robot executes movements. By pre-calculating navigation paths based on the geometric representation of the environment, the system enables faster speeds and sharper turns with guaranteed safety, as the path validity is determined computationally rather than relying on real-time sensor feedback during high-speed operation.
Data Source
AI summary
Systems and techniques for generation, storage, and updating of common floor plans across shared autonomous mobile devices. The systems and techniques include extraction of semantically meaningful data from one or more occupancy maps to provide a floorplan to a user that aligns with the user's understanding of the space using heuristics and machine learning techniques. The techniques also enable repair of damaged or inconsistent floor plan geometries through the use of polygon triangulation and graph cuts.