Indoor Radar Mapping with House Plan Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional radar mapping methods for indoor environments, such as those used by sweeping robots, suffer from low accuracy and incompleteness due to the complexity of furniture layouts and obstacles, which limit the robots' ability to traverse all regions and generate noise in the radar maps.
Innovation Solution
A method and apparatus that utilize a deep neural network to generate vectorized house structure data, acquire and filter radar maps, and perform image matching-fusion processing to create an accurate and complete house display plan by aligning radar map data with standard house plans, adjusting for rotation, size, and room identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional radar mapping methods are used for indoor environments, then the robot can obtain a radar map of the house structure, but the map suffers from low accuracy and incompleteness due to furniture and appliance obstructions limiting the robot's travel range
Solution Approach 1:
The patent introduces a standard house plan as an intermediary reference to guide the radar scanning process. The robot uses the standard house plan to identify regions of interest and prioritize scanning areas that are likely to contain important structural features, thereby improving map accuracy without requiring the robot to physically traverse every obstacle-blocked area.
Solution Approach 2:
The patent performs preliminary processing by comparing the obtained radar map with the standard house plan to identify missing or inaccurate regions. This preliminary analysis allows the system to focus subsequent scanning efforts on specific areas that need improvement, rather than requiring complete exhaustive scanning of the entire environment.
2Loss of information
If the robot traverses all regions to improve map completeness, then more areas can be mapped, but the robot's movement is limited by obstacles on the ground
Solution Approach 1:
The standard house plan serves as a mediator that provides prior knowledge about the expected layout and structure of the environment. This allows the system to infer the presence of regions that the robot cannot physically access, using the standard plan as a reference to fill in gaps in the radar map without requiring direct observation of every area.
Solution Approach 2:
The system performs preliminary comparison between the radar map and standard house plan to identify regions that should exist but are missing from the radar data. This preliminary identification allows the system to mark these areas as needing further investigation or to infer their characteristics from the standard plan, compensating for the robot's limited mobility.
3Loss of information
If the robot scans the entire environment to improve map completeness, then all regions can be covered, but noise increases in the radar map due to obstacles and limited traversal
Solution Approach 1:
The standard house plan acts as a filter or intermediary that helps distinguish between valid structural features and noise caused by obstacles. By comparing radar data against the standard plan, the system can identify and suppress noise regions that do not correspond to expected structural elements, thereby improving map completeness without proportionally increasing noise.
Solution Approach 2:
The system performs preliminary noise filtering by comparing the radar map with the standard house plan before final map generation. This preliminary comparison allows the system to identify regions where radar data likely represents noise rather than actual structural features, enabling selective suppression of noise while preserving genuine environmental information.
Data Source
AI summary
A method for mapping indoor includes: generating, based on a standard house plan of a target house, vectorized house structure data for the target house using a deep neural network; acquiring a first radar map by scanning a travelable space of a first region with a radar, the first region containing at least one room of the target house; and performing image matching-fusion processing based on the first radar map and the house structure data to obtain a house display plan of the target house.