Navigation Device Path Planning via Depth Map Obstacle Probability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robots used in static environments struggle with navigating unfamiliar spaces and adapting to changes in their environment, as they rely on one-directional radar scans and lack the ability to create two-dimensional or three-dimensional maps for effective obstacle avoidance.
Innovation Solution
A path planning method that involves acquiring a two-dimensional depth map, transforming it into a gray distribution map, computing a space matching map, and determining weighting values for each angle range to assess obstacle probability and distance, allowing for real-time navigation without a database or environment map.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional robots use one-directional radar scans to navigate, then they can determine steering direction, but they cannot create two-dimensional or three-dimensional maps for effective obstacle avoidance
Solution Approach 1:
The patent transforms one-directional radar scan data into two-dimensional depth maps by adding spatial dimensionality. The depth sensing system captures depth information across a field of view and organizes it into a two-dimensional representation, enabling the robot to perceive environmental structure rather than just linear distance measurements.
Solution Approach 2:
The patent introduces a depth sensing system as an intermediary between the robot and its environment. This system acts as a mediator that converts raw optical or acoustic signals into structured depth map data, which then feeds into the path planning algorithm for navigation decisions.
2Adaptability or versatility
If robots rely on pre-established maps for navigation, then they can plan paths in static environments, but they cannot adapt to dynamic changes or unfamiliar spaces
Solution Approach 1:
The patent implements dynamic path planning by continuously updating the two-dimensional depth map with real-time sensor data. The path planning algorithm processes current environmental information rather than relying on static pre-established maps, allowing the robot to adapt its navigation to changing conditions while maintaining reliable obstacle avoidance through computational processing.
3Adaptability or versatility
If robots process real-time environmental data without databases or maps, then they can adapt to dynamic environments, but they require complex real-time computation
Solution Approach 1:
The patent segments the environmental data processing into distinct computational stages: depth map acquisition from sensor data, two-dimensional depth map construction, gray distribution map generation through pixel value statistics, space matching map computation, and weighting value calculation for path planning. This segmentation allows complex real-time processing to be broken into manageable computational tasks.
Solution Approach 2:
The patent performs preliminary computational actions by pre-processing sensor data into structured two-dimensional depth maps and gray distribution maps before path planning. These pre-computed representations organize raw environmental data into formats optimized for navigation decisions, reducing the computational burden during real-time path planning.
4Measurement precision
If robots use detailed environmental mapping for accurate navigation, then they can identify obstacles precisely, but they increase processing time and computational load
Solution Approach 1:
The patent applies partial action by computing weighting values for specific angle ranges in the space matching map rather than processing all possible directions with equal detail. The system focuses computational resources on relevant angular sectors for path planning, achieving sufficient obstacle detection precision without the time cost of exhaustive environmental analysis.
Data Source
AI summary
A path planning method applied to a navigation device includes acquiring a two-dimensional depth map, transforming the two-dimensional depth map into a gray distribution map via statistics of pixel values on the two-dimensional depth map, computing a space matching map by arranging pixel counts on the gray distribution map, and computing a weighting value about each angle range of the space matching map in accordance with a distance from location of a pixel count to a reference point of the space matching map. The weighting value represents existential probability of an obstacle within the said angle range and a probable distance between the navigation device and the obstacle.