Indoor Pose Estimation Using Reflective Landmarks and SLAM Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the state of a localizing apparatus, such as drones, in indoor environments suffer from inaccuracies and inefficiencies due to signal attenuation and multi-path effects, especially in complex environments with metal obstructions, and existing optical localization systems struggle to provide accurate positioning and scaling in large spaces.
Innovation Solution
A method that utilizes pre-arranged landmarks with known positions and a combination of SLAM features to determine the state of a localizing apparatus by receiving images and using injective mapping estimates to correct and update the state, incorporating a joint observation model with an extended Kalman filter to reduce error accumulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If inside-out tracking SLAM systems are used to determine the location of a localizing apparatus, then the system can operate autonomously without external cameras, but error accumulates as the map is expanded away from the initial field of view
Solution Approach 1:
The system uses pre-arranged landmarks with known positions as feedback references to continuously correct the accumulated errors in SLAM-based location estimates. By comparing the estimated position against the known landmark positions, the system can detect and correct drift, maintaining long-term accuracy without requiring external intervention.
Solution Approach 2:
Pre-arranged landmarks serve as intermediary reference points between the autonomous SLAM system and the external coordinate system. These landmarks provide stable, known positions that mediate the correction of SLAM errors, allowing the system to maintain accurate location estimates relative to external references without requiring direct external observation.
2Measurement precision
If outside-in tracking optical localization systems are used to determine the location of a localizing apparatus, then location accuracy is maintained relative to external coordinates, but the system scales poorly to larger localization systems requiring multiple cameras
Solution Approach 1:
The system uses locally-mounted cameras that autonomously perform both mapping and localization functions without requiring external camera infrastructure. The localizing apparatus carries its own sensing and processing capabilities, enabling it to determine its position relative to pre-arranged landmarks independently, thus eliminating the need for complex external camera networks.
Solution Approach 2:
Instead of using external cameras to observe the localizing apparatus (outside-in tracking), the system inverts the approach by mounting cameras on the localizing apparatus itself to observe the environment (inside-out tracking). This inversion enables autonomous operation while maintaining accuracy through correction using pre-arranged landmarks.
3Adaptability or versatility
If conventional SLAM algorithms are used for mapping and localization, then the system can operate autonomously in unknown environments, but errors accumulate over time and distance from the initial position
Solution Approach 1:
The system performs preliminary action by pre-arranging landmarks with known positions in the environment before the localizing apparatus begins operation. These pre-positioned references are prepared in advance to provide correction points that will counteract the error accumulation inherent in autonomous SLAM operation.
Solution Approach 2:
The system implements feedback by continuously comparing the SLAM-estimated position against the known positions of pre-arranged landmarks. This feedback mechanism enables detection and correction of drift, allowing the system to maintain accurate location estimates over extended periods and distances while preserving autonomous mapping capabilities.
Data Source
AI summary
A method for determining a state of a localizing apparatus in an indoor environment having N pre-arranged landmarks. The method includes receiving the positions of the N pre-arranged landmarks; receiving a first image, and a second image; receiving positions of n currently mapped simultaneous-localization-and-mapping (SLAM) landmarks; determining positions of at least one feature in the first image, and determining an injective mapping estimate from the at least one features into the set of N pre-arranged landmarks; determining positions of at least one SLAM feature in the second image, and determining a first SLAM injective mapping estimate from at least one SLAM features into the set of the n currently mapped SLAM landmarks; using the determined injective mapping estimate and the determined first SLAM injective mapping estimate to determine the state of the localizing apparatus and to update the positions of the n currently mapped SLAM landmarks. A computer program product comprising instructions which when executed by a computer, cause the computer to carry out the method; and an assembly having controller which is configured to carry out the method.