Mobile Device Tracking with Low-Load Topological Particle Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing localization methods for mobile devices, such as particle filtering, require a large number of particles to account for measurement uncertainty, leading to excessive computational load that smartphones cannot efficiently handle, and relocating this computation to cloud-based systems is not always feasible due to lack of internet connectivity.
Innovation Solution
A method and system that utilize a topological map with weighted particle tracking, reducing computational effort by generating particles based on motion measurements, displacing them according to uncertainty, and resampling based on path direction and distance, allowing localization to be performed locally on the device with a reduced number of particles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a particle filter based dead reckoning approach is used to track mobile device position, then localization accuracy is improved by accounting for measurement uncertainty, but computational load increases exponentially over time making it infeasible for smartphone processors
Solution Approach 1:
The patent extracts only the essential computational elements needed for particle filtering (particle generation, displacement, and resampling based on topological constraints) while removing the computationally intensive components that cause exponential growth. This allows the core localization functionality to be retained with reduced computational overhead suitable for mobile devices.
Solution Approach 2:
The patent performs preliminary action by pre-computing and storing topological map data (graph structure, path directions, and geometric constraints) before the tracking process begins. This pre-processing allows the particle filter to operate efficiently during actual tracking by utilizing pre-available topological information rather than computing it in real-time, thereby reducing computational load during device movement.
2Productivity
If cloud-based processing is used to handle particle filtering computations, then computational capacity is sufficient for real-time tracking, but internet connectivity is required which is not always available
Solution Approach 1:
The patent implements self-service by enabling the mobile device to perform particle filtering computations locally using its own processor and memory resources. The device utilizes its built-in sensors (accelerometer, gyroscope, compass) and locally stored topological map data to independently execute the tracking algorithm without requiring external cloud processing or internet connectivity, thus achieving operational independence.
3Measurement precision
If a large number of particles are used to represent statistical motion measurement errors, then localization accuracy is improved, but memory requirements and processing time increase
Solution Approach 1:
The patent applies partial action by using a reduced number of particles compared to traditional particle filtering approaches. Instead of employing a large particle set to fully capture measurement uncertainty, the method uses topological constraints and geometric reasoning to guide particle behavior, achieving satisfactory localization accuracy with fewer particles thereby reducing memory and processing requirements.
Solution Approach 2:
The patent introduces topological map data as an intermediary element that mediates between the particle filter and the environment. This topological information (graph structure, path directions, geometric constraints) acts as a mediator that guides particle displacement and resampling, allowing the system to achieve accurate localization with fewer particles by leveraging pre-computed spatial relationships rather than relying solely on large particle ensembles.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer-implemented method for tracking a mobile device based on a series of motion measurements of said device comprises retrieving a topological map providing a graph of topological paths, wherein each topological path is associated with a corresponding path direction; statistically generating, on a virtual map, a plurality of particles around an estimated initial position of the mobile device, representing an initial uncertainty of said estimated initial position of the mobile device; determining, from said series of motion measurements, a measured displacement vector of the mobile device representing an estimated displacement direction and an estimated displacement distance of the mobile device; displacing each of said particles on the virtual map according to a randomized particle displacement vector which is derived from said measured displacement vector based on an uncertainty of the measured displacement vector and recording said particle displacement vector for each displaced particle; determining a weight of each particle based on a comparison of the respective particle displacement vector with the path direction of an adjacent topological path and/or based on a distance of the particle to the adjacent topological path; selecting a subset of said displaced particles for resampling based on a particle-weight dependent selection criterion; and regenerating said selected subset of displaced particles within a particle regeneration area around a current mean location of the plurality of particles according to a probability distribution.