RSSI Fingerprint Indoor Positioning with Dynamic Prior Probability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wireless positioning technologies based on Bayesian algorithms face instability and poor accuracy due to reliance on a constant fingerprint database, leading to significant coordinate jumps in complex wireless environments with fluctuating signal strengths.
Innovation Solution
An indoor positioning method using RSSI fingerprints, where multiple positioning gateways detect signal strengths, generate posterior probabilities, and update prior probabilities continuously, allowing historical data fusion to improve positioning stability and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a constant fingerprint database is used for positioning, then the positioning process is simple to implement, but the positioning stability and accuracy deteriorate due to coordinate jumps in complex wireless environments
Solution Approach 1:
The patent transforms the static fingerprint database into a dynamic system where the prior probability is continuously updated based on historical positioning results. The prior probability adapts to signal fluctuations by incorporating temporal information from previous positioning attempts, making the system responsive to changing wireless conditions while maintaining operational simplicity.
Solution Approach 2:
The patent implements feedback by using historical positioning results to update the prior probability for subsequent positioning operations. The system feeds back the outcomes of previous positioning attempts into the probability calculation, creating a closed-loop system that continuously improves accuracy and reduces coordinate jumps based on accumulated experience.
2Device complexity
If a constant fingerprint database is used for positioning, then the system structure is simple, but the positioning accuracy deteriorates due to isolated positioning events
Solution Approach 1:
The patent introduces dynamics into the system structure by making the prior probability time-varying and adaptive. Instead of a fixed database, the system maintains a dynamic probability distribution that evolves with each positioning attempt, capturing temporal correlations in signal behavior without requiring complex additional hardware or infrastructure.
Solution Approach 2:
The patent ensures continuity by linking successive positioning operations through the updated prior probability. Each positioning result contributes to the next positioning attempt, creating a continuous refinement process that eliminates the isolation between positioning events and progressively improves measurement precision over time.
3Reliability
If historical data is integrated into continuous positioning, then positioning stability and trajectory continuity improve, but the computational complexity increases
Solution Approach 1:
The patent maintains continuity of useful action by continuously updating the prior probability with each positioning result. This continuous refinement process improves positioning stability and trajectory continuity by ensuring that each new positioning attempt builds upon previous information, creating a smooth and reliable positioning trajectory over time.
Solution Approach 2:
The patent manages computational complexity through parameter changes by focusing updates on the prior probability parameters rather than recalculating entire positioning models. The system efficiently updates probability distributions using incremental calculations based on new measurements and historical data, avoiding computationally intensive reprocessing while maintaining improved reliability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to the field of wireless positioning technology, and more particularly, to an indoor positioning method based on RSSI fingerprints. The method comprises: Step S1, upon detection of generating of a positioning signal by a device to be located, obtaining a received signal strength of each positioning gateway; Step S2, generating, according to the received signal strength and a prior probability, a global posterior probability of each fingerprint point; Step S3, outputting, according to the global posterior probability and location coordinates of the fingerprint point, a positioning result, and generating, according to the positioning result and the global posterior probability, a new prior probability corresponding to each fingerprint point, then returning to Step S1. The present invention has the following advantageous effects: positioning results are introduced during a continuous positioning process, as a result, the next prior probability for the same device to be located is generated. In doing so, fusion of historical data is achieved during the continuous positioning process, so that a previous positioning can be associated with the current positioning. The coordinate jump phenomenon occurred during an independent positioning process in the prior art is reduced, and the continuity of a positioning trajectory and the positioning stability are improved.