Visual-Inertial Mapping With Low-Power Sensor Offloading
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies for providing fast, accurate, and reliable positional awareness to robots and wearable devices, such as those used in virtual reality and augmented reality, face challenges including high computational costs, energy inefficiency, and high costs due to reliance on expensive hardware like stereo RGB systems and active sensing methods, which limit their widespread adoption.
Innovation Solution
The implementation of a visual-inertial sensor system that offloads computational tasks from the main processor to a low-power sensor module, using a combination of low-cost grayscale and colored sensors, and leveraging SIMD architecture to process image and inertial data efficiently, thereby reducing costs and energy consumption while maintaining performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If expensive hardware like stereo RGB systems and active sensing methods are used, then measurement precision and reliability of positional awareness are improved, but device cost increases
Solution Approach 1:
The patent combines visual sensors (cameras) with inertial sensors (IMU) into a unified sensor suite that processes data from both sources through sensor fusion algorithms. This integration allows the system to achieve high measurement precision comparable to expensive stereo RGB systems while using lower-cost individual components working together synergistically
Solution Approach 2:
The visual-inertial sensor system performs multiple functions including pose estimation, mapping, navigation, and localization across different platforms (autonomous vehicles, robots, VR/AR headsets). This multi-functionality reduces the need for specialized expensive hardware for each application, lowering overall device cost while maintaining precision
2Productivity
If high-performance processors are used to process visual and inertial data, then productivity and speed of positional awareness are improved, but use of energy increases
Solution Approach 1:
The patent divides computational tasks into segments: the inertial sensor provides continuous high-frequency pose estimates at low computational cost, while the visual sensor provides periodic correction updates at lower frequency. This segmentation allows the system to maintain high productivity through the inertial component's continuous operation while reducing overall energy consumption by not requiring constant high-power visual processing
Solution Approach 2:
The system uses periodic visual updates to correct and refine the continuous inertial pose estimates. Rather than continuously processing high-power visual data, the system periodically incorporates visual information to correct drift in the inertial solution, maintaining accuracy while significantly reducing energy consumption compared to continuous high-performance visual processing
3Measurement precision
If marker-based approaches are used for location recognition, then measurement precision is improved, but adaptability of the system decreases
Solution Approach 1:
The patent replaces marker-based mechanical/physical systems with a visual-inertial sensing system that uses natural visual features in the environment. Instead of requiring artificially placed markers, the system detects and tracks natural features through camera images and combines them with inertial data, eliminating the need for markers while maintaining location recognition accuracy and enabling operation in any environment with visual features
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The described positional awareness techniques employing visual-inertial 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 provide positional awareness to machines with improved speed and accuracy.