Mixed Reality Motion Computing System Drift Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional inertial measurement units in mixed reality systems suffer from drift errors, and head-mounted displays have dead zone issues, leading to inaccuracies in interactive input, while existing calibration methods like US 2021/089162 A1 and raycasting techniques do not fully address these problems.
Innovation Solution
A motion computing system combining a wearable device with an inertial measurement unit and a head-mounted display, utilizing a hand tracking module and device tracking module to generate a hand model and calculate a pointer direction, integrating inertial data with hand tracking data to improve accuracy and eliminate drift and dead zone errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If an inertial measurement unit is used to detect user motion, then interactive input can be obtained, but drift error occurs leading to reduced measurement precision
Solution Approach 1:
The patent introduces a camera as an intermediary device to capture images of the user's hand. This visual data serves as a mediator to correct the drift errors from the inertial measurement unit. The system uses the camera images to determine hand orientation and compares it with IMU data, using the visual reference to recalibrate and eliminate accumulated errors over time.
Solution Approach 2:
The system implements a feedback mechanism where captured images are continuously analyzed to determine hand orientation. This visual feedback is compared with inertial measurement data, and the discrepancy is used to correct and recalibrate the IMU readings. The feedback loop continuously eliminates drift errors by referencing the visual field.
2Ease of operation
If a controller is used to detect interactive input data, then user input can be detected, but the user cannot operate hand-free
Solution Approach 1:
The patent replaces the mechanical controller system with a vision-based detection system. Instead of requiring physical contact with buttons or controls, the system uses a camera to track and detect hand gestures and orientations in the visual field. This substitution eliminates the need for handheld controllers while maintaining interactive input detection capabilities.
3Adaptability or versatility
If a head-mounted display is used to display virtual reality field, then immersive experience is provided, but dead zone error occurs reducing measurement precision
Solution Approach 1:
The system uses a camera mounted on or associated with the head-mounted display as an intermediary to capture external visual references. This camera provides an independent visual reference frame that is not affected by the dead zone limitations of the HMD's internal sensors. The visual data from the camera serves as a mediator to correct positional and orientational errors.
4Productivity
If only inertial measurement unit data is used for positioning, then device position can be tracked, but drift error accumulates over time reducing reliability
Solution Approach 1:
The patent merges inertial measurement unit data with visual data from a camera into a unified positioning system. The IMU provides continuous high-frequency motion data for responsive tracking, while the camera provides periodic visual references for error correction. This combination maintains the speed advantages of IMU while eliminating the drift accumulation problem through visual anchoring.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
A motion computing system for virtual reality is provided, which comprises a wearable device and a head-mounted display. The head-mounted display performs following steps for: setting a wearing position of the wearable device; determining whether to generate a hand model having finger skeleton data for the hand of the user is found in a monitored field by a hand tracking algorithm; in response respond to that determine the hand model is found in the monitored field, identifying a device position according to the wearing position and the finger skeleton data, and identifying a device rotation of the wearable device in the monitored field according to the inertial data; calculating a pointer direction in the monitored field according to the device position and the device rotation; and generating a ray in a virtual reality field according to the pointer direction and the device position.