Indoor Camera State Estimation Using Landmark Mapping and TOF
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the state of a camera, such as those used in indoor navigation for robots, suffer from inaccuracies and inefficiencies, particularly in large indoor spaces, due to signal attenuation and multi-path effects, leading to errors in localization and mapping, which are not adequately addressed by current optical localization systems.
Innovation Solution
A method involving an injective mapping estimate and an observation model using a state-space model, combined with a TOF camera or intensity information, to accurately determine the state of a camera by mapping features to known landmarks, utilizing an extended Kalman filter for precise state estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical localization systems use triangulation techniques with multiple cameras to determine object position, then localization accuracy is improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent inverts the traditional outside-in tracking approach by implementing inside-out tracking, where the camera mounted on the moving object tracks the environment instead of external cameras tracking the object. This reduces the number of cameras needed from multiple external units to a single onboard camera, thereby reducing device complexity while maintaining localization capability
Solution Approach 2:
The patent introduces an intermediary coordinate system approach where the camera's local coordinate system is related to the world coordinate system through a transformation matrix. This intermediary mathematical model allows position determination with a single camera instead of requiring multiple physical cameras for direct triangulation, reducing system complexity
2Adaptability or versatility
If inside-out tracking systems generate environment maps to determine position, then scalability to large spaces is improved, but measurement precision deteriorates due to error accumulation
Solution Approach 1:
The patent incorporates feedback mechanisms where the determined camera state (position and orientation) is continuously updated and refined. The system uses the observed features and their known world coordinates to feedback-correct the camera's estimated position, preventing error accumulation that would otherwise occur in pure inside-out tracking systems
Solution Approach 2:
The patent changes the mathematical parameters used in localization by employing a state-space model with state variables representing camera position and orientation. By formulating the problem in terms of state estimation rather than simple map matching, the system can maintain precision across large distances through proper parameter management and error modeling
3Ease of manufacture
If existing methods determine camera state using optical localization, then implementation simplicity is improved, but measurement precision deteriorates in large indoor spaces
Solution Approach 1:
The patent replaces traditional mechanical/optical triangulation systems with a computational state estimation approach. Instead of using multiple physical cameras for geometric triangulation, the system uses a single camera with computational algorithms (state-space models, feature tracking, coordinate transformations) to determine position, maintaining simplicity while improving accuracy in large spaces
Solution Approach 2:
The patent changes the fundamental parameters of the localization system by introducing state variables (position x, y, z and orientation angles) and using statistical estimation methods rather than direct geometric calculation. This parameter transformation allows the system to maintain precision over large distances while keeping the hardware simple
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 enhanced accuracy and scalability in determining the state of a camera, enabling reliable indoor navigation by accurately tracking the camera's position and orientation, even in complex environments with reduced error propagation.
Implementation Method 1
receiving distance data indicative of distance between the M features and the corresponding M landmarks, respectively
Data Source
AI summary
The invention relates to a method for determining a state xk (8) of a camera (11) at a time tk, the state xk (8) being a realization of a state random variable Xk, wherein the state is related to a state-space model of a movement of the camera (11). The method comprises the following steps: a) receiving an image (1) of a scene of interest (15) in an indoor environment (15) captured by the camera (11) at the time tk, wherein the indoor environment (15) comprises N landmarks (9) having known positions in a world coordinate system (12), N being a natural number; b) receiving a state estimate x{circumflex over ( )}k (2) of the camera (11) at the time tk, c) determining (3) positions of M features in the image (1), M being a natural number; d) receiving (4) distance data indicative of distance between the M features and the corresponding M landmarks (9), respectively; e) determining (5) an injective mapping estimate from the M features into the set of the N landmarks (9) using at least (i) the positions of the M features in the image and (ii) the state estimate (2); f) using the determined injective mapping estimate (5) to set up (6) an observation model in the state-space model, wherein the observation model is configured for mapping the state random variable Xk of the camera onto a joint observation random variable Zk, wherein at the time tk, an observation zk is a realization of the joint observation random variable Zk, and wherein the observation zk comprises (i) the position of at least one of the M features in the image, and (ii) the distance data indicative of distance; and g) using (7) (i) the state estimate, (ii) the observation model, and (iii) the observation zk, to determine the state xk (8) of the camera at the time tk. The invention also relates to a computer program product and to an assembly.