Robot Hand Sensor Fusion for Occluded Position Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In industrial robot assembly, measuring the position and orientation of complex components is challenging due to partial occlusion by the robot hand, leading to incomplete image capture and failure in calculating the required features, necessitating increased measurement apparatuses or adjusted image capturing positions.
Innovation Solution
An information processing apparatus that combines noncontact sensor data with contact sensor information to calculate the position and orientation of a target object, using error distribution analysis to compensate for measurement inaccuracies and ensure high accuracy even when partial data is obtained.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the robot hand grasps the component to measure position and orientation, then the component can be held for assembly operations, but the component becomes hidden by the robot hand causing incomplete image capture and measurement failure
Solution Approach 1:
The measurement process is segmented into two distinct phases: (1) noncontact measurement phase where the component is measured before grasping using a noncontact sensor to obtain initial positional information, and (2) contact measurement phase where the component is grasped and additional contact positions are measured using contact sensors. This segmentation allows measurement to occur at different stages where different measurement methods are appropriate.
Solution Approach 2:
The noncontact measurement is performed as a preliminary action before the robot hand grasps the component. This preliminary measurement captures the component's position and orientation when it is still visible and accessible, providing initial data that can be used even when subsequent contact measurement is incomplete due to occlusion.
2Device complexity
If only a portion of the object is captured by the camera near the robot hand, then the measurement system remains simple, but the calculation of position-orientation fails due to insufficient features
Solution Approach 1:
The system merges two different measurement approaches: noncontact measurement (using sensors like cameras or laser scanners) and contact measurement (using tactile sensors on the robot hand). By combining the results from both measurement methods, the system compensates for the limitations of each individual method and achieves accurate position-orientation calculation even when only partial object information is available.
Solution Approach 2:
The control device acts as an intermediary that processes and integrates measurement information from both noncontact and contact sensors. It uses the shape information of the component as a reference model to correlate and fuse the partial measurement data from different sources, enabling accurate position-orientation calculation despite incomplete direct observation.
3Measurement precision
If multiple measurement apparatuses are used to ensure complete feature capture, then measurement accuracy improves, but the system complexity and cost increase
Solution Approach 1:
The robot hand serves multiple functions: it is both the actuator for grasping and manipulating the component, and the mounting platform for contact sensors that perform measurement. This multi-functionality eliminates the need for separate dedicated measurement apparatuses, as the same robotic system performs both manipulation and measurement tasks.
Solution Approach 2:
The measurement system uses the robot hand's own contact sensors to measure contact positions on the component, allowing the system to self-measure without requiring external dedicated measurement devices. The robot hand essentially measures the component it is grasping, reducing overall system complexity.
Data Source
AI summary
A position and an orientation of an object are measured with high accuracy. An approximate position-orientation of a target object is obtained, positional information of the target object is obtained by measuring the target object using a noncontact sensor, positional information of contact positions touched by a contact sensor is obtained by bringing the contact sensor into contact with the target object, and a position-orientation of the target object is obtained by associating shape information of the target object with the positional information of the target object and the positional information of the contact positions in accordance with the approximate position-orientation.