Edge-Cloud Robot Control With Shared Adaptive Learning Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot systems face inefficiencies in controlling heterogeneous robots due to the time-consuming and costly process of parameter exchange and machine learning model training between edge servers and cloud servers, particularly when dealing with diverse robotic tasks and environments.
Innovation Solution
A cloud-based robot system where a cloud server generates a base deep learning model that is tuned and upgraded by edge servers to create adaptive models tailored to specific robots and environments, minimizing data transmission through the cloud and enabling efficient intelligence augmentation by sharing deep learning models among edge servers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If parameters are exchanged among edge nodes for machine learning model updates, then model adaptation capability is improved, but system complexity and difficulty of supporting heterogeneous robots increases
Solution Approach 1:
The patent introduces a cloud server as an intermediary that manages the machine learning model lifecycle. The cloud server receives training data from edge servers, performs centralized model training, and distributes updated models back to edge servers. This intermediary approach simplifies the complexity at edge nodes while maintaining adaptability through centralized model management and coordination.
Solution Approach 2:
The system divides the machine learning workflow into distinct segments: data collection at edge servers, centralized model training at cloud server, and model deployment back to edge servers. This segmentation allows each component to focus on specific tasks, reducing overall system complexity while maintaining adaptability through specialized processing at each stage.
2Measurement precision
If a large amount of data is transmitted between edge server and cloud server for model training, then model accuracy is improved, but transmission time and cost increases
Solution Approach 1:
The patent extracts only the essential training data and model parameters that need to be transmitted between edge and cloud servers, rather than transmitting all raw data. The edge server performs preliminary data processing and selects only the most relevant training samples, reducing transmission volume while maintaining model accuracy.
Solution Approach 2:
The edge server performs preliminary data processing, filtering, and selection before transmitting data to the cloud server. This preliminary action ensures that only high-quality, relevant training data is transmitted, improving model accuracy while minimizing transmission time and bandwidth consumption.
3Adaptability or versatility
If a new machine learning model is trained from scratch at the edge server, then model customization is improved, but processing time and computational resources increase
Solution Approach 1:
The cloud server performs preliminary model training using aggregated data from multiple edge servers, creating a pre-trained base model. This preliminary action provides edge servers with a head start, allowing them to perform only fine-tuning and customization locally, thereby maintaining model adaptability while significantly reducing processing time and computational resources required at the edge.
Solution Approach 2:
The system merges the strengths of centralized and distributed training: the cloud server handles heavy computational lifting by training base models using aggregated data from multiple sources, while edge servers handle customization and fine-tuning. This combination achieves both model customization and processing efficiency by dividing work according to computational capabilities and local needs.
Data Source
AI summary
The present embodiment relates to a cloud-based robot control method for controlling a plurality of robots which are positioned in a plurality of spaces divided arbitrarily, the method comprising the steps of: generating a control base model which can be applied to the plurality of robots in a cloud server; distributing the control base model to edge servers allocated to respective spaces; upgrading the control base model in accordance with the plurality of robots of a space, in the edge server; directly transmitting the upgraded control model from the edge server to another edge server; and controlling the plurality of robots by means of the upgraded control model in the edge server. Therefore, by sharing a deep-learning model among edge servers, supporting heterogeneous robots and heterogeneous services is possible. Further, a base deep-learning model from the cloud server is tuned into a customized deep-learning model to be suitable for respective robots in the edge server, and the deep-learning model is upgraded to an adaptive deep-learning model to be suitable for a service provided by respective robots, and thus an optimized service can be provided.