Robot Navigation for Ordered Multi-Object Search in Real Spaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation systems for robots and vehicles face challenges in efficiently navigating and locating multiple objects in real-world spaces, particularly in environments where objects have different visual attributes, as they often rely on simulated environments and lack effective methods for exploration and semantic mapping in real-world settings.
Innovation Solution
A hybrid approach combining Simultaneous Localization and Mapping (SLAM) with deep neural networks for exploration and goal retrieval, using reinforcement learning and supervised learning to enable robots to detect and navigate to multiple objects with different attributes in a predetermined order, by generating maps from camera and LIDAR inputs and employing exploration and waypoint navigation strategies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If navigation systems rely on simulated environments for training, then the robot can learn navigation tasks in controlled conditions, but the system lacks effectiveness in real-world settings with diverse visual attributes
Solution Approach 1:
The patent creates a virtual copy of the real-world environment through SLAM mapping, generating a digital representation that the robot can navigate in simulation. This virtual map copies the spatial structure and visual attributes of the real environment, allowing the robot to train in simulated conditions while maintaining effectiveness in real-world deployment. The virtual environment serves as a faithful replica that preserves the complexities of real-world navigation.
2Productivity
If the robot uses traditional navigation methods, then the system structure is simple, but the robot cannot efficiently locate and navigate to multiple objects with different attributes in real-world spaces
Solution Approach 1:
The navigation task is segmented into distinct components: SLAM for environmental mapping, object detection for identifying targets with specific attributes, and navigation planning for pathfinding. This segmentation allows each module to specialize in one function, improving overall efficiency while managing complexity through modular architecture. The robot can independently handle mapping, detection, and navigation without requiring a monolithic complex system.
Solution Approach 2:
The system changes parameters by incorporating attribute-based filtering for object identification, using color, shape, and other visual attributes as navigation criteria. This parameter-based approach enables the robot to efficiently locate multiple objects with different attributes by adjusting search parameters rather than using exhaustive search methods, thereby improving productivity without proportionally increasing complexity.
3Loss of information
If the robot explores unknown areas to find objects, then the robot can locate objects not in the map, but the navigation time and energy consumption increase
Solution Approach 1:
The robot performs preliminary exploration and mapping using SLAM before object search begins. By pre-establishing a comprehensive map of the environment during an initial phase, the robot eliminates the need for time-consuming exploration during actual navigation tasks. Objects can then be located efficiently within the pre-mapped areas without requiring real-time environmental scanning, reducing navigation time while maintaining map completeness.
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 allows robots to effectively navigate and locate multiple objects in real-world environments by decomposing the navigation task into exploration and waypoint navigation, improving the efficiency and accuracy of object detection and path planning, bridging the gap between simulated and real-world performance.
Implementation Method 1
The mapping module is configured to generate the map further based on input from a light detection and ranging (LIDAR) sensor of the navigating robot
Data Source
AI summary
A navigating robot includes: a feature module configured to detect objects in images captured by a camera of the navigating robot while the navigating robot is in a real world space; a mapping module configured to generate a map including locations of objects captured in the images and at least one attribute of the objects; and a navigation module trained to find and navigate to N different objects in the real world space in a predetermined order by: when a location of a next one of the N different objects in the predetermined order is stored in the map, navigate toward the next one of the N different objects in the real world space; and when the location of the next one of the N different objects in the predetermined order is not stored in the map, navigate to a portion of the map not yet captured in any images.