360° Depth Camera Navigation for Obstacle and Shadow Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous mobile robots face challenges in navigating indoor environments with unstructured layouts and untrained personnel due to the complexity of obstacle detection and handling, especially when operating at significant heights, as conventional sensors provide inadequate and costly solutions for precise obstacle avoidance.
Innovation Solution
Employing two three-dimensional depth cameras affixed to opposite sides of the robot body, providing a 360-degree field of view, and a processing system to analyze pixel data for missing or erroneous pixels using templates to distinguish obstacles, shadows, and the robot's body, enabling precise obstacle detection and navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional obstacle sensors are used for navigation, then obstacle detection capability is provided, but measurement precision and reliability are insufficient for tall robots in unstructured environments
Solution Approach 1:
The patent combines multiple depth camera sensors (front, rear, and side cameras) into an integrated navigation system. This merging of multiple sensing components provides comprehensive 360-degree depth coverage, eliminating blind spots and improving both measurement precision for obstacle detection and overall navigation reliability in unstructured environments.
Solution Approach 2:
The patent transitions from conventional 2D image capture to 3D depth assessment by employing depth camera sensors that capture depth information in the Z-dimension. This dimensional enhancement enables precise obstacle detection at various altitudes and distances, significantly improving measurement precision for navigation purposes.
2Measurement precision
If multiple sensors are added to improve detection accuracy, then measurement precision improves, but device complexity and cost increase
Solution Approach 1:
The depth camera sensors serve multiple functions: they detect obstacles, assess depth information, identify glass surfaces, and provide navigation data. This multi-functionality allows the system to achieve high measurement precision without proportionally increasing device complexity, as the same sensors perform multiple detection tasks.
Solution Approach 2:
The patent replaces complex mechanical sensor arrays with optical depth camera sensors that use light-based measurement. This substitution reduces mechanical complexity while maintaining or improving measurement precision, as the optical system naturally provides three-dimensional depth information without moving parts.
3Area of stationary object
If depth cameras are positioned higher on the robot body, then field of view increases, but distance to floor increases reducing detection accuracy
Solution Approach 1:
The patent segments the sensing function by placing depth cameras at different locations (front, rear, and sides) rather than relying on a single high-mounted sensor. This segmentation allows each camera to maintain optimal distance to the floor in its respective viewing direction while collectively providing comprehensive 360-degree field of view coverage.
Solution Approach 2:
The patent uses multiple depth cameras positioned at different angular orientations to capture depth information from various perspectives. This angular dimensionality complementation ensures that each camera maintains precise floor detection capability in its specific viewing sector while the collective system achieves omnidirectional coverage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention discloses an autonomous mobile robot (106), comprising a robot body; a first three-dimensional depth camera sensor (102) affixed to a first side of the robot body; a second three-dimensional depth camera sensor (104) affixed to a second side of the robot body; wherein the first and second three-dimensional depth camera sensors (102, 104), in combination, comprise an at least 360 degree field of view (108) of the major floor surface (110) around the robot body; and a processing system (120) communicative with the first and second three dimensional depth camera sensors (102, 104) and comprising non-transitory computing code which, when executed by at least one processor (410) associated with the processing system (120), causes to be executed the steps of: receiving of pixel data (170) within the field of view (108) from the first and second three-dimensional depth camera sensors (102, 104); obtaining missing or erroneous pixels (170a) from the pixel data (170); comparing the missing or erroneous pixels (170a) to at least one template, wherein the at least one template comprises at least an indication of ones of the missing or erroneous pixels (170a) indicative of the robot body and a shadow (106b) of the robot body; and outputting an indication of obstacles (160, 162, 164) in or near the field of view (108) based on the comparing.