Pedestrian Dead Reckoning via TNB Reference Frame
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pedestrian dead reckoning (PDR) methods face challenges in accurately estimating direction of travel (DoT) without introducing latency, particularly in environments with strong magnetic fields or without magnetometers, due to the need for accurate step detection and data buffering.
Innovation Solution
The use of differential geometric properties of human gait, specifically through the computation of tangent-normal-binormal (TNB) reference frames from acceleration data, allows for real-time and robust estimation of user position, speed, and direction of travel, leveraging kinematic properties like curvature and torsion rotation rates, and employing a low-pass filter for smoothing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If step detection and data buffering are used to estimate direction of travel, then measurement precision is improved, but loss of time increases due to latency
Solution Approach 1:
The patent pre-computes and stores the TNB reference frames from acceleration data as the user moves through the environment. By preparing this geometric information in advance during normal movement, the system eliminates the need for real-time step detection and data buffering when the user stops or needs direction information, thus reducing latency while maintaining precision
Solution Approach 2:
The TNB reference frame acts as an intermediary that bridges acceleration data and direction of travel estimation. Instead of directly computing DoT from raw acceleration data through complex step detection algorithms, the patent uses the TNB frame as an intermediate representation that captures the geometric properties of movement, simplifying the estimation process and reducing computational latency
2Measurement precision
If step detection is used to segment motion data, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent extracts the essential geometric information needed for direction estimation directly from the acceleration data by computing the TNB reference frame. This approach takes out only the necessary geometric properties (tangent, normal, and binormal vectors) from the complex motion data, eliminating the need for complex step detection and segmentation algorithms while maintaining estimation accuracy
Solution Approach 2:
The patent replaces the mechanical step detection system (which requires identifying discrete steps and segmenting data) with a continuous geometric field approach using the TNB reference frame. This substitution eliminates the need for discrete event detection and data segmentation, simplifying the system architecture while preserving measurement precision
Data Source
AI summary
In some embodiments, a method comprises: receiving acceleration data from a motion sensor of a mobile device carried by a user, the acceleration data represented by a space curve in a three-dimensional (3D) acceleration space, the space curve indicative of a cyclical vertical displacement of the user's center of mass accompanied by a lateral left and right sway of the center of mass when the user is stepping; computing a tangent-normal-binormal (TNB) reference frame from the acceleration data, the TNB reference frame describing instantaneous geometric and kinematic properties of the space curve over time; and computing a direction of travel of the user based on an orientation of a unit binormal (B) vector of the TNB reference frame in the 3D acceleration space, and computing a speed of the user based on a linear regression model applied to the kinematic properties of the TNB reference frame.