Adaptive Pedestrian Inertial Navigation Using Floor Type Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pedestrian inertial navigation systems face challenges in accurately navigating on varying floor types and gait frequencies due to fixed parameters and lack of adaptive detection methods.
Innovation Solution
The system employs an inertial measurement unit (IMU) to receive and partition data based on toe-off and heel strike events, using principle component analysis to reduce dimensionality and an artificial neural network to identify floor types, thereby adapting navigation parameters for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If fixed parameters are used in ZUPT-aided navigation, then the system is simple to implement, but navigation accuracy deteriorates for different users and floor types
Solution Approach 1:
The patent applies dynamics by transitioning from fixed parameters to adaptive parameters that automatically adjust based on detected floor types and gait characteristics. The system dynamically modifies navigation parameters including zero-velocity update thresholds, stance phase detection criteria, and gait frequency estimates according to real-time sensor data, enabling accurate navigation across diverse conditions without manual reconfiguration.
Solution Approach 2:
The patent implements parameter changes by modifying navigation parameters based on detected floor types and gait patterns. The system changes parameters such as zero-velocity update thresholds, stance phase duration estimates, and acceleration thresholds according to the detected environment (e.g., hard floor vs. soft floor) and user gait characteristics, thereby maintaining high navigation accuracy across different scenarios.
2Measurement precision
If additional sensors are added for floor type detection, then floor type detection accuracy improves, but device complexity increases
Solution Approach 1:
The patent applies universality by making the IMU serve multiple functions: it not only performs inertial navigation but also detects floor types and gait characteristics using the same sensor data. The system extracts floor type information from IMU acceleration and orientation patterns during gait cycles, eliminating the need for separate floor detection sensors while achieving accurate multi-functional operation.
Solution Approach 2:
The patent implements self-service by having the navigation system use its own IMU data to detect floor types and adapt parameters automatically. The system processes its intrinsic sensor measurements to identify environmental conditions and adjusts its navigation behavior accordingly, without requiring external sensors or manual input, thereby reducing overall system complexity.
3Measurement precision
If conventional floor type detection systems are used, then floor type identification is achieved, but they are not suitable for pedestrian navigation conditions
Solution Approach 1:
The patent applies local quality by tailoring the floor detection method specifically to pedestrian gait conditions rather than using generic floor detection. The system analyzes local characteristics of IMU data during stance phases, heel strikes, and toe-offs to identify floor types, creating a specialized detection approach that adapts to the specific motion patterns and environmental conditions of pedestrian navigation.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances navigation accuracy by dynamically adapting parameters to different floor types and gait frequencies, reducing navigation errors and improving position estimation.
Implementation Method 1
pedestrian inertial navigation systems have been developed using microelectromechanical systems (MEMS) based inertial measurement units (IMUs)
Implementation Method 2
partitioning the inertial data into a plurality of partitions. The method also includes reducing, by the processor, dimensionality of the plurality of partitions using principle component analysis
Data Source
AI summary
Processes and systems for adaptive pedestrian inertial navigation are provided. Configurations can adjust to various navigation scenarios, including different floor types and different gait paces. A combination of IMU data partition, principal component analysis (PCA), and artificial neural network may be used to perform the floor type detection. Floor type results may be used in the multiple-model extended Kalman filter. In each extended Kalman filter, an adaptive threshold is used for the stance phase detection to enable the detector to adjust to gait frequency without tuning design parameters during navigation. A floor type classification of high accuracy is demonstrated, and the position error in a velocity-changing navigation system using adaptive threshold is reduced.