Robot Action Logging for Compact Replayable Operation History
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing logging methods for robot devices generate large amounts of data exponentially with operation time, making it difficult to replay operation history and search for issues, and require separate compiling processes, which are inefficient for development environments.
Innovation Solution
A robot device that generates and stores data in a robot programming language format, allowing for easy replay of operation history and eliminating the need for separate compiling processes, by using instruction data, sensing data, and map data to control hardware elements and store action logs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If text-based logging method is used to record robot operations, then complete operation history can be captured, but the log file size increases exponentially and replay becomes difficult
Solution Approach 1:
The patent creates a simplified copy of operation data in a structured format (JSON) that captures essential information without requiring complete text descriptions. This copy contains only the necessary fields for replay (timestamp, action, parameters), dramatically reducing file size while maintaining operational fidelity.
Solution Approach 2:
The patent transforms unstructured text logs into structured parameter-based records with fixed fields (timestamp, action, parameters, metadata). This parameterization allows for compact storage and efficient processing, changing the data representation from verbose text to concise structured format.
2Loss of information
If detailed text logs are generated for each operation, then complete operation details are preserved, but searching and replaying problems becomes inefficient
Solution Approach 1:
The patent segments operation data into distinct structured fields (timestamp, action, parameters, metadata) that can be independently accessed and searched. This segmentation allows for efficient indexing and retrieval of specific operations without processing entire text logs, significantly reducing search and replay time.
Solution Approach 2:
By converting operation details into parameter-based structured data, the patent enables efficient querying and filtering capabilities. Specific parameters can be searched and filtered programmatically, making problem identification and replay operations much faster than text-based approaches.
3Loss of information
If conventional logging methods are used, then operation data can be recorded, but separate compiling process is required which complicates development
Solution Approach 1:
The patent implements self-service logging where the robot system automatically generates and structures log data without requiring external compilation processes. The logging mechanism is integrated into the robot's operational framework, allowing logs to be created and processed in real-time during operation, eliminating separate compiling steps.
Solution Approach 2:
The patent creates a universal logging format (JSON) that can handle multiple types of robot operations and data sources through a single standardized structure. This multi-functional approach allows the same logging mechanism to record diverse operations without requiring separate compilation processes for different operation types.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A robot device is disclosed. The robot device comprises: a sensor; a communication interface; a memory; and a processor for, when an action of the robot device changes, storing instruction data corresponding to the action, sensing data of a sensor related to the action, and map data related to the action in the memory, transmitting data stored in the memory to a server via the communication interface, and generating an event when it is identified that sensing data corresponding to the action occurred in the robot device does not belong to a threshold range, on the basis of data received from the server.