Extended Kalman Filter Trajectory Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the trajectory of a moving object, especially indoors, suffer from significant errors due to reliance on specific infrastructure and interference from electromagnetic signals, and are not applicable to objects other than pedestrians.
Innovation Solution
A method using a Kalman filter to estimate the state of a moving object by acquiring acceleration, angular velocity, local magnetic field, and GNSS signal data, incorporating an extended Kalman filter with a state vector that includes position, speed, and attitude, and averaging instantaneous velocity vectors to reduce errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If WiFi fingerprinting or signal of opportunity fingerprinting methods are used to determine trajectory, then the method can be implemented with existing radio signal infrastructure, but the determination presents significant error when far from this specific infrastructure
Solution Approach 1:
The patent introduces an extended Kalman filter as an intermediary computational system that processes sensor data (accelerometer, gyroscope, magnetometer, barometer) to estimate trajectory. This mediator bridges the gap between raw sensor measurements and accurate position determination, enabling reliable trajectory tracking without dependence on external radio signal infrastructure.
Solution Approach 2:
The patent replaces reliance on electromagnetic signal infrastructure (WiFi, radio signals) with a self-contained inertial measurement system using mechanical sensors (accelerometers, gyroscopes, magnetometers, barometers) combined with mathematical filtering. This substitution eliminates the need for external signal infrastructure while maintaining trajectory determination capability.
2Measurement precision
If satellite receiver and inertial unit measurements are merged with external signals from dedicated systems, then trajectory can be determined with infrastructure support, but the method depends on specific infrastructure deployment and fails far from it
Solution Approach 1:
The patent creates a universal trajectory determination system that functions across diverse environments (indoors, outdoors, urban, rural) using a multi-functional sensor suite. The extended Kalman filter integrates data from multiple sensor types (accelerometer, gyroscope, magnetometer, barometer) to provide consistent performance regardless of location or infrastructure availability, making the system adaptable to any environment.
Solution Approach 2:
The system is self-sufficient, using only onboard sensors and computational algorithms to determine trajectory without requiring external infrastructure services. The extended Kalman filter processes data from the device's own sensors (accelerometer, gyroscope, magnetometer, barometer) to compute position, velocity, and orientation, making the system independent of external signal sources or infrastructure deployment.
3Measurement precision
If ZUPT or ZARU methods are used to reduce error by detecting foot flat phases, then positioning accuracy improves for pedestrians, but the method cannot be generalized to other moving objects
Solution Approach 1:
The patent develops a universal trajectory determination method applicable to any moving object (pedestrians, vehicles, drones, animals) by using a generalized sensor fusion approach. Instead of object-specific gait detection (ZUPT for pedestrians), the extended Kalman filter processes acceleration, angular velocity, magnetic field, and pressure data in a unified framework that adapts to any moving object's motion characteristics without requiring object-specific algorithms.
4Measurement precision
If magnetometer and satellite receiver measurements are used indoors, then trajectory can be determined, but the measurements are greatly degraded due to electromagnetic interference from building infrastructure
Solution Approach 1:
The extended Kalman filter serves as an intermediary that processes and filters sensor data to eliminate the effects of electromagnetic interference. By combining data from multiple sensors (accelerometer, gyroscope, magnetometer, barometer) and using statistical filtering, the system extracts accurate trajectory information while suppressing noise and interference from building infrastructure.
Solution Approach 2:
The patent merges measurements from multiple sensor types (accelerometer, gyroscope, magnetometer, barometer) to compensate for the degradation of individual sensors in electromagnetic interference environments. The combination of these diverse measurement sources through extended Kalman filtering provides robust indoor positioning that overcomes the limitations of any single sensor type under interference conditions.
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 provides a precise trajectory estimation for moving objects indoors without relying on specific infrastructure, significantly reducing positioning errors by utilizing degraded GNSS data and accounting for sensor biases and object orientation.
Implementation Method 1
an acquisition step comprising at least one of the acquisition of an acceleration vector of the moving object
Implementation Method 2
the acquisition of an angular velocity vector of the moving object
Implementation Method 3
the acquisition of a local magnetic field vector at the moving object
Implementation Method 4
the acquisition of a phase of a signal emitted by a global navigation satellite system
Implementation Method 5
an extended Kalman filter with a state vector that includes position, speed, and attitude, and averaging instantaneous velocity vectors to reduce errors
Data Source
Figure 1~2
Figure 3A~3B
Figure 4A~4B
AI summary
A method for determining the trajectory of a moving object comprises acquiring an acceleration vector (ab) and an angular velocity vector (ωb) of the moving object, a local magnetic field vector (Bb) at the moving object, and a phase (φi) of a signal (56) emitted by a global navigation satellite system, and a step of determining a position state of the moving object using an extended Kalman filter. A state (X) of the extended Kalman filter comprises an instantaneous velocity vector (v) and an average velocity vector (vi) of the moving object and a time offset (cδGPSΔdTΔt) between the clock of the satellite receiver and the time of the global navigation satellite system. The average velocity vector (vi) is an average of the instantaneous velocity vectors (vi) of at least N=fIMUfGNSS previous states of the extended Kalman filter.