Pixelwise Confidence Depth Maps for Robot-Specific Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robotic systems face challenges in navigating and interacting with environments due to incomplete or inaccurate depth information from various sensors, which can lead to difficulties in performing operations reliably.
Innovation Solution
A method is provided to generate a pixelwise filterable depth map by determining pixelwise confidence levels based on sensor association, allowing for the creation of a robot operation-specific depth map by comparing these confidence levels with operation-specific thresholds, thereby filtering out unreliable depth information for specific robot operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If depth information from various sensors is used for robot navigation and interaction, then the robot can perceive the environment, but the depth information may be incomplete or inaccurate leading to unreliable operations
Solution Approach 1:
The patent segments the depth map into multiple confidence levels at the pixel level. Each pixel is assigned a confidence value indicating the reliability of its depth information, allowing the system to differentiate between high-confidence and low-confidence depth data for more reliable robot operations
Solution Approach 2:
The patent introduces confidence level parameters to characterize the quality of depth information. By changing the parameter from binary (valid/invalid) to multi-level confidence values, the system can make more nuanced decisions about which depth information to trust for navigation and interaction tasks
2Reliability
If all depth information is used for robot operations, then complete environmental coverage is achieved, but unreliable depth information may lead to errors
Solution Approach 1:
The patent performs preliminary filtering of depth information by assigning confidence levels before the robot executes operations. This preliminary assessment allows the system to identify and set aside unreliable depth data while preserving complete environmental coverage for future reference
Solution Approach 2:
The confidence level acts as an intermediary between the raw depth information and the robot's operation decisions. Rather than directly using or discarding depth data, the system uses confidence levels as a mediating layer to make informed decisions about which information to trust
3Measurement precision
If pixelwise confidence filtering is implemented for different robot operations, then accurate navigation is achieved using only reliable depth information, but the processing complexity increases
Solution Approach 1:
The patent implements dynamic confidence thresholding where the filtering criteria can change based on the specific robot operation being performed. Different operations can use different confidence thresholds, allowing the system to adapt the processing complexity to the requirements of each task
Data Source
AI summary
A method includes receiving sensor data from a plurality of robot sensors on a robot. The method includes generating a depth map that includes a plurality of pixel depths. The method includes determining, for each respective pixel depth, based on the at least one robot sensor associated with the respective pixel depth, a pixelwise confidence level indicative of a likelihood that the respective pixel depth accurately represents a distance between the robot and a feature of the environment. The method includes generating a pixelwise filterable depth map for a control system of the robot. The pixelwise filterable depth map is filterable to produce a robot operation specific depth map. The robot operation specific depth map is determined based on a comparison of each respective pixelwise confidence level with a confidence threshold corresponding to at least one operation of the robot controlled by the control system of the robot.