Monocular VSLAM Scale Estimation Using Optical Flow Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous robotic devices face navigation challenges due to temporal and spatial variations in their environment and changes in obstacle positions and motions, which existing techniques struggle to address effectively, particularly when wheel encoder data is erroneous or affected by slippage.
Innovation Solution
The implementation of visual simultaneous localization and mapping (VSLAM) using an optical flow sensor, which processes image frames to generate homography computations, determines scale estimation values, and adjusts robotic device pose based on optical flow and wheel encoder data, ensuring accurate navigation even in conditions of slippage or erroneous sensor readings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If wheel encoder data is used for pose estimation, then the navigation system can operate with existing sensors, but the accuracy deteriorates when slippage or erroneous readings occur
Solution Approach 1:
The patent introduces an optical flow sensor as an intermediary measurement device to bridge the gap between wheel encoder data and actual device motion. The optical flow sensor provides independent motion verification that mediates between the potentially erroneous wheel encoder readings and the true pose changes, enabling accurate pose estimation even when wheel slippage occurs
Solution Approach 2:
The system implements feedback by continuously comparing pose estimates from wheel encoders with measurements from the optical flow sensor and monocular camera. When discrepancies are detected (indicating slippage or erroneous readings), the system adjusts the pose estimation by incorporating corrective information from the optical flow and visual data, creating a closed-loop correction mechanism
2Device complexity
If monocular VSLAM is used for pose estimation, then the system can operate with simple sensors, but the scale estimation accuracy deteriorates due to inherent monocular limitations
Solution Approach 1:
The optical flow sensor serves as an intermediary that provides scale information to complement monocular VSLAM. By combining optical flow measurements (which provide direct scale cues) with monocular visual features, the system achieves accurate scale estimation while maintaining the simplicity of using only a single camera
Solution Approach 2:
The system changes the parameter estimation approach by using optical flow to directly measure motion scale and combining it with monocular VSLAM's structure-from-motion estimates. This multi-parameter fusion (combining optical flow scale with visual structure) resolves the scale ambiguity inherent in monocular systems without adding stereo cameras or other complex sensors
3Measurement precision
If the system uses multiple sensor fusion methods, then navigation accuracy improves in dynamic environments, but the computational complexity increases
Solution Approach 1:
The system applies partial sensor fusion by selectively combining data from wheel encoders, optical flow sensor, and monocular camera only when needed. Rather than continuously fusing all sensors, the system activates specific fusion pathways based on detected conditions (e.g., slippage detection triggers optical flow-based correction), reducing computational overhead while maintaining accuracy when required
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the robotic device's ability to navigate autonomously by accurately determining its pose and scale, improving performance on slippery surfaces and correcting for wheel encoder inaccuracies, thereby ensuring reliable operation in dynamic environments.
Implementation Method 1
an optical flow sensor configured to output optical flow information that characterizes a transition between the two image frames
Implementation Method 2
receiving a first image frame from a monocular image sensor, receiving a second image frame from the monocular image sensor
Data Source
AI summary
Various embodiments include methods for improving navigation by a processor of a robotic device equipped with an image sensor and an optical flow sensor. The robotic device may be configured to capture or receive two image frames from the image sensor, generate a homograph computation based on the image frames, receive optical flow sensor data from an optical flow sensor, and determine a scale estimation value based on the homograph computation and the optical flow sensor data. The robotic device may determine the robotic device pose (or the pose of the image sensor) based on the scale estimation value.