Robotic Fleet Maintenance Scheduling Using AI Wear Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The proliferation of data from IoT sensors and other sources in value chain networks overwhelms traditional centralized data management systems, leading to complexity and inefficiencies in data processing and decision-making.
Innovation Solution
A method for processing queries in a distributed database using edge devices, which store and transmit approximate responses based on summary data stored on a dynamic ledger, such as a blockchain, allowing for efficient data retrieval and management across a network of edge devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If centralized data management systems are used to handle IoT sensor data, then data processing can be performed with a single centralized system, but the system becomes overwhelmed by the volume and complexity of data, leading to inefficiencies
Solution Approach 1:
The patent divides the centralized data management system into multiple distributed edge devices that each independently process and manage data locally. This segmentation reduces the burden on any single system while collectively handling the voluminous IoT sensor data through distributed processing nodes throughout the network.
Solution Approach 2:
The patent transitions from a single-dimensional centralized architecture to a multi-dimensional distributed architecture by introducing spatial distribution across multiple edge devices. This dimensional change allows data processing to occur at multiple locations simultaneously, improving overall system capacity and efficiency.
2Loss of information
If all detailed data is stored and transmitted across the network, then complete information is available for analysis, but network bandwidth is consumed and centralized systems become overwhelmed
Solution Approach 1:
The patent extracts only the essential summary data from the complete dataset at edge devices, separating the detailed local data from the transmitted network data. This extraction allows complete information to remain available locally while only condensed summaries consume network bandwidth for transmission to other parts of the system.
Solution Approach 2:
The patent applies partial action by transmitting only a subset (summary) of the complete data rather than all detailed information. This partial transmission suffices for many analytical purposes while dramatically reducing network bandwidth consumption compared to transmitting the full dataset.
3Loss of time
If summary data is stored on a dynamic ledger at edge devices, then data retrieval is faster and more efficient, but the system requires a complex distributed database infrastructure
Solution Approach 1:
The patent performs preliminary actions by pre-computing and storing summary data at edge devices before queries are executed. This advance preparation of aggregated data enables rapid retrieval when needed, as the summary information is already processed and localized rather than requiring real-time computation from raw data.
Solution Approach 2:
The patent introduces a dynamic ledger as an intermediary layer between raw IoT sensor data and query processing systems. This intermediary structure stores pre-computed summaries that mediate between the complexity of distributed data storage and the simplicity of fast data retrieval, reducing the perceived complexity for query operations.
Data Source
AI summary
A robotic fleet management platform includes a resources data store maintaining a fleet resource inventory indicating fleet resources that can be assigned to a robotic fleet. For each respective fleet resource, the fleet resource inventory indicates maintenance status data, a predicted maintenance need, and/or a preventive maintenance schedule. A maintenance management library of fleet resource maintenance requirements facilitates determining maintenance workflows, service actions, and/or service parts for fleet resources. The platform calculates the predicted maintenance need of a fleet resource based anticipated component wear. The anticipated wear/failure is derived from machine learning-based analysis of the maintenance status data. The platform monitors a health state of the fleet resource from sensor data. The platform adapts the preventive maintenance schedule. The platform initiates a service action of the at least one item of maintenance for the fleet resource based on the fleet resource maintenance requirements and/or the new preventive maintenance schedule.