The classic static automation solutions cannot cope with today’s high production variability environment and meet its needs. The defining property of modern AI technologies, on the other hand, is that they are dynamic with high adaptability and continuous improvement. This is exactly how they offer an unprecedented capability to meet automation needs in such a challenging production environment. To realize this promise of AI technologies, the production culture needs to change as well. Lack of an empowering collaborative environment bringing stakeholders together such that AI solutions can be easily designed, deployed, monitored and adapted, is the main obstacle to generate value, eventually slowing down time to value, even blocking it completely.
The philosophy behind the ReliVision platform is to generate value in the shop-floor by exploiting state-of-the-art AI technologies. We enable our customers to overcome the obstacles that block the path to value in several stages, such as, adapting to highly variable production, deploying at and monitoring on the shop-floor, cooperation between data engineers and automation engineers to improve solutions.
The Relimetric’s ReliVision platform is an end-to-end custom solution development environment for designing, training, deploying and perfecting AI powered automated inspection pipelines on any (visual) data. It is specifically tailored for the needs and constraints of the manufacturing industry, gathering data scientists/engineers with industrial engineers/operators around the common goal of generating value on the shop-floor. The pipeline concept, which is an end-to-end inspection solution, is at the heart of the ReliVision platform. In its most general form, a pipeline is composed of AI Blocks (AI models) and Basic Blocks (eg. digital signal/image processing - DSP/DIP - functions).
You can design custom automated inspection pipelines; train, test and compare alternative pipelines; easily deploy, monitor and improve your pipelines on the shop-floor; ReliVision provides a collaborative secure and managed environment for multiple users with different profiles, such as data scientists/engineers, shop-floor operators and automation engineers. It is this unified and efficiently streamlined collaborative environment that facilitates the maximum utilization of AI for automated inspection tasks in real manufacturing environments.
The key differentiation of ReliVision in the AI based automation market comes from
High accuracy in a highly variable production setting:
The unique infrastructure that ReliVision is built upon, enables fast and effective adaptation of AI solutions to production variations due to product customizations which is becoming a standard rather than a specialized service in today’s global markets. In essence, the ability to retrain and deploy your existing AI models with additional data with ease and without any additional investment allows our customers to retain high accuracy levels across a wide range of product portfolio.
Shorter time to value with low-cost maintenance:
All AI based automated inspection systems have a learning curve until they achieve sufficient accuracy and reliability in the field. This earning curve is what determines the time-to-value. ReliVision’s unique infrastructure streamlines the data acquisition, AI model training, deployment, monitoring field operations, getting feedback from the field, retraining and re-deploying AI models optimally which results in a steep learning curve and thus shorter time-to-value. Furthermore, the whole learning process is driven with minimal support from high-cost AI engineers and may even be managed primarily by field operators without any in-depth AI know-how.
Cost effective scale up in multiple dimensions to fit to changing needs and constraints:
The proprietary infrastructure has a microservices based architecture which makes ReliVision scaleable by design and without any extra cost. While the ReliTrainer and ReliAudit engines can handle multiple use-cases simultaneously through a queueing mechanism, they can easily be scaled to multiple production lines and/or plants replicating successful AI solutions. The computation power can be scaled simply by adding new hardware without new software acquisitions or installations.
Specialized for production & manufacturing industries:
ReliVision platform is designed to resolve the procedural and managerial bottlenecks in production and manufacturing industries that hinder value creation with modern AI technologies. The platform brings state-of-the-art AI solutions to shop-floor operations in a flexible, adaptable and self maintainable fashion.
This is achieved by ReliVision platform through providing
Data source (hardware) agnostic integrability.
Guaranteed system up-time.
On-prem and/or cloud installation, as required, with seamless integration.
Scalable-by-design in multiple dimensions: use-case, production line / site, hardware
Secure, collaborative, custom profile multi-user environment with detailed monitoring and logging functions.
Intuitive and simple UIs appealing to collaborating users with different backgrounds, effectively bringing data scientists/engineers together with production engineers and operators
The Relimetric’s ReliVision platform is built upon a modular, flexible and scalable proprietary architecture as depicted below. The main building blocks of the ReliVision architecture are the Relimetrics Training Engines (RMTEs) and the Relimetrics Inference Engines (RMIEs). The whole architecture is based on a distributed micro-services concept. This allows all components to be deployed locally (on-prem) and/or remotely (on cloud), also scaled up easily.
The 3 main modules of ReliVision are ReliUI, ReliTrainer (built on RMTEs) and ReliAudit (built on RMIEs):
ReliUI is the ReliTrainer frontend for users to access and use ReliTrainer’s data curation and AI powered solution designing, (re)training, testing and deploying functions.
ReliTrainer is the data curation and AI powered solution design, (re)train, test and deployment component built on RMTEs. Its functions are accessible via ReliUI. ReliTrainer comes with ReliBoard, a web interface to manage user accounts, to monitor ReliTrainer processes and datasets by means of user friendly dashboards.
ReliAudit is the shop-floor/field component on which multi-modal data is collected, automation systems are controlled, AI solutions are deployed and monitored. It is built on RMIEs. It comes with a web HMI which has access to the local DB to review audit results and provide feedback for AI model improvement.
ReliVision Ecosystem Architecture
The ReliVision architecture is scalable-by-design in multiple dimensions. In the multi-site scaling dimension, the RMTEs and RMIEs can be multiplied as required, both locally and remotely. In multi-usecase scaling dimension, both RMTEs and RMIEs can handle multiple training and inferencing tasks simultaneously by means of an internal queueing system. In hardware scaling dimension, RMTEs and RMIEs can be easily connected to additional GPUs/CPUs, while RMIEs can also receive input from multiple data sources (eg. cameras). This scalability-by-design brings in ease and cost efficiency for multiplying the value generating shop-floor automated inspection operations.
The comprehensive user account management allows defining custom user profiles with variable access rights. The communication between the ReliUI and RMTEs/RMIEs goes through secure VPN connections when all servers are private to the customer, either on-prem or on-cloud. In all other cases where access to an external server, such as one of Relimetrics servers, is required, HTTPS connections are used to ensure security. The shop-floor operations are usually time-sensitive hence it is advised to have RMIEs to be deployed on on-prem servers directly connected to the data sources (sensors, cameras, etc.) with 10 Gbs LAN connections.
Relivision platform modules can be deployed on-prem or on-cloud. Typically, ReliTrainer is deployed on-prem or on-cloud depending on customer requirements, likewise for ReliAudit but it is almost always deployed on-prem to assure swift shop-floor operation and control. The minimum system requirements are as follows:
Client Computer for ReliUI
OS: Windows 10+
HDD: 100+ GB
RAM: 64+ GB
ReliTrainer Server (RMTE) (Cloud + On-prem)
OS: Ubuntu 20+ / Windows 11
GPU: Nvidia GPU with minimum 16GB dedicated VRAM and tensor cores (eg. lowend - T4 GPU; highend - A100 recommended)
HDD: 100+ GB
RAM: 64+ GB
Accessible through 3 ports
ReliAudit Server (RMIE) (On-prem)
OS: Ubuntu 20+
GPU (optional for fast runtime): Nvidia GPU with minimum 16GB dedicated
VRAM and tensor cores (eg. lowend - T4 GPU; highend - A100 recommended)
HDD: 100+ GB
RAM: 64+ GB
Accessible through 3 ports
ReliVision platform’s infrastructure is sensor/camera agnostic. The current ReliVision platform has been configured to work with Basler GigE cameras (https://www.baslerweb.com/en/)