Monocular Visual-Inertial Localization with Wheel Slip Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Monocular camera-based localization systems face challenges in providing accurate scale information due to scale ambiguity, which affects the robot's ability to determine its precise location in an environment, especially when wheel slip events occur, leading to inaccurate odometry data.
Innovation Solution
A multi-sensory approach combining a monocular camera, MEMS inertial sensors, and optical flow sensors is used to cross-examine poses from different odometry modules, detect and reject wheel slip events, and perform online scale calibration and optimization, thereby improving localization accuracy and stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If monocular camera is used for localization, then device complexity is reduced, but measurement precision of depth and scale is lost
Solution Approach 1:
The patent combines monocular camera visual odometry with inertial measurement unit (IMU) data to create a visual-inertial localization system. The IMU provides depth and scale information that compensates for the monocular camera's inability to measure depth directly, while the camera provides visual context. This merging of sensors resolves the contradiction by maintaining device simplicity while gaining measurement precision through data fusion.
2Stability of the object's composition
If wheel odometry data is used during wheel slip events, then localization continuity is maintained, but measurement precision deteriorates due to accumulated error
Solution Approach 1:
The patent implements a feedback mechanism that continuously monitors wheel slip conditions by comparing wheel odometry data with visual-inertial localization results. When wheel slip is detected (when the discrepancy exceeds a threshold), the system automatically adjusts the fusion weights to reduce reliance on wheel odometry data. This feedback loop maintains localization continuity while preventing accumulated error from degrading measurement precision.
3Measurement precision
If factor-graph-based optimization is used for scale calibration, then measurement precision is improved, but computational burden increases
Solution Approach 1:
The patent applies partial optimization by using factor-graph-based optimization selectively rather than continuously. The system performs full optimization at key moments (such as when scale drift is detected or at regular intervals) and uses simpler interpolation or estimation methods between these optimization points. This approach maintains measurement precision when needed while reducing computational burden during normal operation.
Data Source
AI summary
The method and device disclosed herein presents a method that includes capturing, by an optical sensor disposed on a device moving in an environment, a plurality of optical data at respective locations within a portion of the environment; capturing, by a wheel encoder disposed on the device, a set of encoder data corresponding to the plurality of optical data at the respective locations; determining a first relative motion based on the plurality of optical data; determining a corresponding second relative motion based on the set of encoder data. In accordance with determining that a difference between the first relative motion and the corresponding second relative motion is larger than a first threshold: increasing a counter indicating a slip event of the wheel encoder. The slip event corresponds to a wheel of the device advancing and the corresponding second relative motion being below a second threshold.