Robot Joint Attention via Optical Indication Device
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for establishing joint attention between a user and a robot, such as laser pointer recognition, gesture recognition, touch panel operation, and acceleration sensor use, face challenges in ensuring accuracy and robustness due to environmental factors and natural interface issues, leading to difficulties in specifying the target with certainty.
Innovation Solution
A robot device equipped with an instruction acquisition unit, position/posture estimation unit, and target specifying unit that utilizes an optical indication device to estimate the position and posture of the optical indication device operated by the user, allowing the robot to specify the target direction based on the estimation and confirm interaction establishment with the user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If laser pointer recognition is used with a dedicated filter, then detection robustness is improved, but the camera cannot be used for any other purpose
Solution Approach 1:
The patent divides the camera system into two functional components: a general-purpose camera for various detection tasks and a dedicated filter module specifically for laser pointer recognition. This segmentation allows the general-purpose camera to maintain versatility while the dedicated filter provides reliable laser detection when activated.
Solution Approach 2:
The patent implements a dual-mode camera system where the camera can operate in two distinct modes: general-purpose imaging mode for various detection tasks and laser pointer recognition mode when the dedicated filter is activated. This multi-functionality resolves the contradiction by allowing the same hardware to serve multiple purposes.
2Ease of operation
If only laser pointer recognition is used, then target indication is simplified, but recognition accuracy decreases when other high brightness points are present
Solution Approach 1:
The patent merges two different recognition approaches: laser pointer recognition (based on brightness detection with dedicated filter) and gesture recognition (based on spatial position and motion patterns). By combining these methods, the system can distinguish between actual target indications and spurious high-brightness points, thereby improving recognition accuracy while maintaining operational simplicity.
3Adaptability or versatility
If gesture recognition is used, then the camera can be used for other purposes, but recognition accuracy is insufficient
Solution Approach 1:
The patent introduces an intermediary verification mechanism that uses the dedicated filter's laser detection capability to validate gestures detected by the general-purpose camera. The filter acts as a mediator that confirms whether a detected gesture corresponds to an actual laser pointer indication, thereby improving gesture recognition accuracy without compromising camera versatility.
4Device complexity
If acceleration sensor is not frequently reset, then operation complexity is reduced, but detection error accumulates
Solution Approach 1:
The patent implements a feedback mechanism where the robot's actual position and orientation (obtained from environment map updates and sensor fusion) are continuously compared with the position calculated from acceleration sensor integration. When discrepancies exceed a threshold, the system automatically resets the acceleration sensor, thereby maintaining detection accuracy without requiring frequent manual intervention.
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 solution enables robust and accurate establishment of joint attention between the user and the robot, improving the reliability of target specification and interaction confirmation, even in varying lighting environments and with low display resolutions.
Implementation Method 1
an optical indication device, which is operated by the user to indicate the target by irradiation of a beam
Data Source
AI summary
There is provided a robot device including an instruction acquisition unit that acquires an order for encouraging a robot device to establish joint attention on a target from a user, a position/posture estimation unit that estimates a position and posture of an optical indication device, which is operated by the user to indicate the target by irradiation of a beam, in response to acquisition of the order, and a target specifying unit that specifies a direction of the target indicated by irradiation of the beam based on an estimation result of the position and posture and specifies the target on an environment map representing a surrounding environment based on a specifying result of the direction.