Indoor Localization via RSSI Pattern Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Indoor room-localization systems face challenges in accurately determining the location of processor-based client devices within indoor areas due to overlapping RSSI patterns from different access points, leading to misclassification and inefficiencies in localization processes.

Innovation Solution

The system employs a processor-based client device that collects WiFi RSSI data, uses pattern classification algorithms to learn and differentiate RSSI patterns, and implements a sequential decision fusion method with Dirichlet distribution to improve classification accuracy by fusing responses from classifiers and updating probability vectors, thereby reducing misclassification in overlapping regions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional RSSI-based localization methods are used, then the system can determine device location, but misclassification occurs in overlapping regions leading to reduced accuracy

Engineering Contradiction:
Improvelocation determination accuracyVSAvoidclassification reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent combines multiple classifier responses using sequential decision fusion to create a more reliable location determination. Instead of relying on a single classifier that may fail in overlapping RSSI regions, the system merges decisions from multiple classifiers trained on different feature sets, thereby improving both accuracy and reliability simultaneously

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The system changes the parameter representation by using Dirichlet distribution to model probability vectors over time. This statistical approach transforms raw classifier outputs into probabilistic predictions that can be sequentially updated, allowing the system to adapt to changing environmental conditions and resolve ambiguities in overlapping regions

Inventive Principle:
Principle #35Parameter changes

2Reliability

If multiple classifiers are used to improve accuracy, then classification reliability increases, but computational complexity and processing time increase

Engineering Contradiction:
Improveclassification reliabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the classification task into multiple independent classifiers that can be trained and executed separately. Each classifier focuses on specific features or regions, and their results are combined through sequential decision fusion. This segmentation allows the system to achieve high reliability without requiring a single overly complex classifier

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system implements dynamic updating of probability vectors using sequential decision fusion, where classifier results are integrated over time rather than processed statically. This dynamic approach allows the system to adapt to changing conditions and improve reliability while maintaining manageable computational complexity through efficient probability updates

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3397982B1Indoor room-localization system
Publication Date: 2023.10.18 ROBERT BOSCH GMBH
  • EP3397982B1 patent drawingFigure 1
  • EP3397982B1 patent drawingFigure 2
  • EP3397982B1 patent drawingFigure 3

AI summary

A processor-based client device may be localized in an indoor area based on Received Signal Strength Indication (RSSI) values from different access points is provided. A general geographic area in which the processor-based client device is located. A position of the processor-based client device on the identified area is determined. A context-aware information is displayed on the processor-based client device once the identified area is determined.