Monocular-Auxiliary Sensor Fusion for Robot Positional Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies for providing fast, accurate, and reliable positional awareness to robots are inadequate, especially in complex environments, as they fail to incorporate real-time sensory data and are often costly, power-intensive, or prone to interference, limiting their effectiveness in task planning and execution.
Innovation Solution
A monocular-auxiliary sensor system that utilizes visual data from a single camera, combined with inertial measurement units and wheel odometry data, to estimate positional information and adapt plans in real-time, leveraging low-end imaging sensors and offloading computational tasks to reduce energy consumption and enhance processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional SLAM techniques are used for positional awareness, then robots can identify location and plan movement, but the system is slow and inaccurate for fast motion applications
Solution Approach 1:
The system segments the sensing and processing functions by using a camera for visual feature detection and an independent IMU for motion tracking. The IMU processes motion data independently at high frequency while the camera provides periodic visual corrections, allowing fast motion tracking without sacrificing positional accuracy.
Solution Approach 2:
The patent introduces an IMU as an intermediary device between the camera and the SLAM system. The IMU captures high-frequency motion data that bridges the temporal gap between camera frames, providing continuous positional updates during fast motion while the camera provides periodic visual feature matching for accuracy correction.
2Measurement precision
If depth sensors are used for positional awareness, then accurate depth information is obtained, but power consumption increases and interference issues occur
Solution Approach 1:
The system replaces active depth sensing mechanisms (like time-of-flight or structured light sensors) with passive visual odometry using a standard camera combined with IMU data. This substitution eliminates the need for additional power-intensive depth sensing hardware while achieving comparable positional accuracy through computational methods.
Solution Approach 2:
The patent uses low-end, inexpensive cameras instead of expensive depth sensors. While individual camera frames provide limited depth information, the system accumulates positional accuracy over time through visual feature tracking and IMU integration, achieving reliable depth estimation without high-cost hardware.
3Ease of manufacture
If visual approaches are used for positional awareness, then cost is reduced, but processing speed is slow leading to failure in fast motion applications
Solution Approach 1:
The system merges visual odometry from a low-cost camera with inertial navigation from an IMU. The camera provides cost-effective visual feature detection while the IMU provides high-speed motion tracking. By fusing these two data streams, the system achieves both low cost and high processing speed suitable for fast motion applications.
Solution Approach 2:
The system dynamically adjusts the fusion of visual and inertial data based on motion speed and camera frame rate. During fast motion, the IMU data is weighted more heavily to maintain processing speed, while during slower motion, visual corrections are applied more frequently to maintain accuracy, allowing the system to adapt to varying speed requirements.
4Measurement precision
If marker based approaches are used for localization, then accurate positioning is achieved, but the work area is limited by marker placement requirements
Solution Approach 1:
The system extracts and removes the requirement for artificial markers from the environment. Instead of relying on pre-placed markers, the system uses natural visual features from the environment captured by the camera, combined with IMU data, to achieve accurate positioning without restricting the operational area to marked zones.
Solution Approach 2:
The visual-inertial system provides universal positioning capability that works across diverse environments without requiring environment-specific modifications. The system can operate in markerless environments, on various terrains, and in different lighting conditions, making it adaptable to any workspace without the limitation of marker placement.
Data Source
AI summary
The described positional awareness techniques employing sensory data gathering and analysis hardware with reference to specific example implementations implement improvements in the use of sensors, techniques and hardware design that can enable specific embodiments to find new area to cover by a robot performing an area coverage task of an unexplored area. The sensory data are gathered from an operational camera and one or more auxiliary sensors.