Concurrent Static and Dynamic Object Reconstruction in SLAM
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robots face challenges in establishing spatial and temporal relationships between stationary and moving objects in a scene, with existing sensor technologies being noisy and unsuitable for simultaneous localization and mapping (SLAM) and moving object tracking, leading to inefficient performance in dynamic environments.
Innovation Solution
A sparse feature-based SLAM system that concurrently reconstructs static and dynamic objects, using a multi-stage geometric verification approach to distinguish features from static and dynamic regions, allowing for real-time estimation and updating of both static and dynamic object maps, reducing computational and storage burdens.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional sensor devices and separate SLAM/tracking systems are used, then the system can handle static or moving objects independently, but the performance degrades in dynamic environments with both static and moving objects present
Solution Approach 1:
The patent combines SLAM and moving object tracking into a single unified system that simultaneously estimates poses of both static landmarks and moving objects. The system uses a joint state vector that includes both static and dynamic object parameters, and performs concurrent estimation using all available measurements from multiple frames, thereby improving reliability in dynamic environments where both static and moving objects coexist
2Loss of information
If all features from both static and dynamic objects are used in SLAM, then complete scene information is captured, but computational and storage burdens increase significantly
Solution Approach 1:
The patent segments features into two distinct categories: static features that form the SLAM map and dynamic features that belong to moving objects. The system maintains separate data structures for static landmarks and moving object trajectories, allowing efficient storage and processing. Only static features are used for long-term mapping, while dynamic features are tracked temporarily and discarded when objects leave the scene, significantly reducing computational and storage burdens while preserving complete scene information
Solution Approach 2:
The patent extracts and separates dynamic object features from the overall feature set before performing SLAM. By identifying and removing features associated with moving objects from the static map construction process, the system eliminates unnecessary computational overhead and storage requirements while maintaining complete information about both static and dynamic elements through separate tracking mechanisms
3Productivity
If separate estimators are used for static and moving objects, then the estimation problem becomes lower dimensional and feasible for real-time updates, but the system cannot establish spatial relationships between moving objects and static scene
Solution Approach 1:
The patent uses static landmarks as intermediary reference points that mediate between moving objects and the global coordinate system. The joint estimation process establishes spatial relationships by expressing both static landmark positions and moving object poses in the same coordinate frame, with static landmarks serving as fixed reference points. This allows the system to maintain lower dimensional separate estimators for real-time processing while simultaneously establishing accurate spatial relationships between moving objects and the static scene through the shared landmark references
Data Source
AI summary
An imaging system for localization and mapping of a scene including static and dynamic objects. A sensor acquires a sequence of frames in motion or stationary. A memory to store a static map of static objects and an object map of each dynamic object in the scene. The static map includes a set of landmarks, and the object map includes a set of landmarks and a set of segments. A localizer registers keypoints of the frame with landmarks in the static map using frame-based registration and to register some segments in the frame with segments in the object map using a segment-based registration. A mapper to update each object map with keypoints forming each segment and keypoints registered with the corresponding object map according to the segment-based registration, and to update the static map with the remaining keypoints in the frame using the keypoints registered with the static map.