Indoor Navigation Using Camera-Based Sensor Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In urban and indoor environments, mobile devices struggle to provide accurate position estimates due to the unavailability of satellite and cellular signals, leading to unreliable navigation as inertial sensors accumulate errors over time through conventional dead reckoning.
Innovation Solution
A method using a mobile device's camera to capture images of horizontal features and process them to determine angles relative to the device's orientation, allowing for calibration of inertial sensors by comparing optical and sensor measurements, thereby correcting sensor misalignments and improving navigation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If inertial sensors and conventional dead reckoning are used for position estimation indoors, then navigation capability is maintained without satellite signals, but position estimation accuracy deteriorates over time due to accumulated errors
Solution Approach 1:
The patent applies feedback by using the mobile device's current position estimate to select and process relevant map features, then using the processed feature angles to recalibrate inertial sensors. This closed-loop approach continuously corrects accumulated errors by comparing sensor measurements with reference data from the map, thereby maintaining position estimation accuracy over time.
Solution Approach 2:
The patent performs preliminary action by pre-processing map data to identify geometric features (walls, corridors, intersections) and storing their angular relationships. When the device enters an indoor environment, this pre-processed map information is immediately available for rapid calibration without requiring real-time complex computations, enabling quick error correction.
2Adaptability or versatility
If inertial sensors are used for dead reckoning, then navigation continues without external signals, but sensor calibration deteriorates due to misalignment errors accumulating over time
Solution Approach 1:
The patent replaces mechanical sensor calibration procedures with an automated computational approach. Instead of physically aligning sensors with reference objects, the system uses image processing of map features to compute angular relationships and automatically recalibrates inertial sensor data through algorithmic correction based on the processed feature angles.
Solution Approach 2:
The system performs self-service calibration by using its own camera to capture images of environmental features and its own inertial sensors to measure orientations. The mobile device independently processes this data to detect calibration errors and corrects its own sensor measurements without requiring external calibration equipment or manual intervention.
3Measurement precision
If satellite and cellular signals are used for position estimation, then position accuracy is maintained, but signal availability deteriorates in urban and indoor environments
Solution Approach 1:
The patent introduces map features (walls, corridors, intersections) as intermediaries between the inertial sensors and the final position estimate. The camera captures images of these intermediate features, image processing extracts their angular relationships, and these processed angles serve as intermediary data to calibrate sensors and improve position estimation accuracy in signal-denied environments.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An apparatus and method for providing a direction based on an angle of a reference wall is provided. A mobile device uses an angle of a horizontal feature from an image to calibrate a sensor and future sensor measurements. The angle of the horizontal feature is determined by image processing and this angle is mapped to one of four assumed parallel or perpendicular angles of an interior of a building. A sensor correction value is determined from a difference between the sensor-determined angle and the image- processing determined angle. The image processing determined angle is assumed to be very accurate and without accumulated errors or offsets that the sensor measurements may contain.