Service Robot Motion Selection for Natural Conversational Gestures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Service robots performing similar or same motions consecutively can create a perception of malfunctioning or unnatural behavior, leading to user dissatisfaction.
Innovation Solution
A robot equipped with a camera, speaker, driving unit, and processor that detects users, identifies sentences without allocated motions, and selects a different motion from pre-stored options based on similarity and size ratios to ensure varied and natural motion sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the robot extracts a keyword from a sentence and obtains a motion corresponding to the extracted keyword, then the robot can perform motions while uttering sentences, but if a similar keyword is extracted from a following sentence, the robot performs similar motions consecutively which makes the user feel the robot is malfunctioning or the motion is not natural
Solution Approach 1:
The system compares the motion corresponding to the currently extracted keyword with the motion performed in the previous time section. When similarity exceeds a threshold, the system feeds back this information to select an alternative motion from pre-stored motions, preventing repetitive motions and ensuring natural behavior perception.
Solution Approach 2:
The system changes the motion selection parameter by introducing a similarity comparison mechanism. When the similarity between consecutive motions exceeds a threshold, the system switches from direct keyword-motion mapping to alternative motion selection, thereby varying the motion parameters to avoid repetition.
2Ease of operation
If the robot performs motion recognition and selection based on extracted keywords, then the robot can respond to sentences with appropriate motions, but the system becomes complex requiring keyword extraction, motion matching, and similarity comparison mechanisms
Solution Approach 1:
The system performs preliminary action by pre-storing multiple motions for each keyword and pre-calculating similarity relationships between motions. This preparation is done in advance, so that during actual operation, the system only needs to compare and select from pre-prepared options, reducing real-time computational complexity.
Solution Approach 2:
The system introduces an intermediary mechanism - a similarity comparison module that acts as a mediator between keyword extraction and motion selection. This intermediary layer simplifies the control logic by providing a clear decision criterion (similarity threshold) for motion selection.
Data Source
AI summary
A robot and a method for controlling thereof are provided. The robot includes a camera, a speaker, a driving unit, a memory, and a processor, in which the processor is configured to detect a user based on an image obtained through the camera, obtain a plurality of sentences to be uttered by the robot to the user, identify a first sentence to which a motion to be performed while the robot utters is not allocated among the plurality of sentences, identify a second sentence to be uttered in a second time section adjacent to a first time section in which the first sentence is to be uttered, obtain a first motion different from a second motion allocated to the second sentence among a plurality of motions pre-stored in the memory as a motion corresponding to the first sentence, and control the speaker to output a voice corresponding to the first sentence and control the driving unit to perform the first motion while the voice is output.