Inertial Data Segmentation for Mobile Device Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing inertial localization methods using low-cost MEMS IMUs in smartphones suffer from high sensor noise, leading to unbounded system drifts and errors, making accurate indoor localization challenging, especially with dynamic user movements and varying sensor attachments.
Innovation Solution
A computer-implemented method that segments inertial data into pseudo-independent windows, estimates changes in navigation states within each window without initial state propagation, and sums these changes to determine the device's location, using only inertial data and employing a neural network or genetic algorithm to learn the relationship between inertial data and navigation states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional strapdown inertial navigation systems integrate IMU measurements directly, then the system can provide continuous position and velocity estimates, but the low-cost MEMS sensors produce high sensor noise leading to exponential error propagation and unbounded system drifts
Solution Approach 1:
The patent segments the continuous inertial navigation problem into discrete step-based updates. Instead of continuously integrating accelerometer data which causes exponential error growth, the system divides motion into discrete steps, estimating position changes between steps. This segmentation transforms the unbounded error accumulation into bounded step-wise updates, resolving the contradiction between continuous estimation and measurement precision.
Solution Approach 2:
The patent introduces an intermediary step detection and classification mechanism between the raw IMU measurements and the final position estimate. The system uses gyroscope data and acceleration patterns to detect steps, classify them (e.g., walking, running, stationary), and estimate step lengths. This intermediary processing layer filters out high-frequency noise while preserving meaningful motion information, thereby improving position accuracy without sacrificing continuous tracking capability.
2Measurement precision
If step-based pedestrian dead reckoning is used to mitigate drift, then the system reduces error propagation, but it fails when no periodic steps are detected such as when the phone is placed on a trolley or shopping cart
Solution Approach 1:
The patent creates a universal inertial navigation system that handles multiple motion types through a single integrated framework. The system detects motion patterns and automatically adapts its processing: for periodic step-based motion it uses step detection, while for non-periodic motion (trolley pushing, vehicle mounting) it switches to continuous acceleration integration with drift correction. This multi-functional approach maintains position accuracy across diverse应用场景 without requiring separate systems for different motion types.
Solution Approach 2:
The patent implements dynamic adaptation of the navigation algorithm based on detected motion characteristics. The system continuously monitors acceleration and gyroscope data to determine whether the current motion pattern exhibits periodicity characteristic of walking. When periodicity is detected, step-based PDR is activated; when non-periodic motion is detected (as with trolleys or vehicle mounting), the system dynamically switches to a modified continuous integration approach. This dynamic behavior resolves the contradiction between step-based accuracy and versatility.
3Adaptability or versatility
If dynamic step estimation is performed to adapt to user walking habits, then the system can personalize tracking, but the estimation is heavily influenced by sensor noise and phone attachment changes causing unavoidable errors
Solution Approach 1:
The patent implements feedback mechanisms where the system continuously monitors the quality and consistency of step detection results. When sensor noise or attachment changes cause degraded performance (detected through anomalies in step timing, acceleration magnitude, or gyroscope patterns), the system adjusts its processing parameters or switches to alternative estimation methods. This feedback loop allows the system to maintain personalized tracking while compensating for noise and attachment variations, resolving the contradiction between adaptability and precision.
Data Source
AI summary
A computer-implemented method of determining a location of a mobile device is provided. The method can include receiving inertial data generated at the mobile device, the inertial data including a plurality of samples taken at different times, segmenting the inertial data into pseudo-independent windows, wherein each pseudo-independent window can include a plurality of the samples and wherein one or more initial states for each pseudo-independent window are treated as unknown, estimating a change in navigation state over each pseudo-independent window using the samples of inertial data, and summing the changes in the navigation states over the pseudo-independent windows so as to determine the location of the mobile device.