What is FITS?
Future Intelligent Transportation System - At its core, FITS is AI-based Intelligent Transportation Platform. Most customers have road and transportation sensors from across a spectrum of vendors and suppliers. FITS automatically collects data from all and any such sensors into a uniﬁed data store, fuses it with AI engines and enables business process speciﬁc processing for any data-driven use case - from speed enforcement to traffic optimization. It also monitors the ﬂeet of sensors to ensure maximum data collection uptime and includes 24/7 monitoring centre services.
Benefits of FITS
> Privacy: Single tenant in isolated EUROPEAN data centre, compliance with ISO/IEC 27001/27002:2013, compliance with GDPR
> Optimized Operations: Automatic incident tracking system, 24/7 monitoring and call centre (in English, Latvian, Russian)
> Customized Integration: Case data offload for legal post-processing, cross-checking object registration, whitelist/blacklist and advanced workﬂows Other capabilities on-demand
> Complete Management: Quick deployment, scalability, OpEx over CapEx, risk reduction
Where FITS is used?
FITS runs as a cloud-native solution on Azure, using a mix of PAAS and IAAS services. Architected in scale units separated by Azure Queues (using both Storage and Service Bus for different tasks), it can scale from a deployment of 50 sensors to close to unlimited.
Apart from AI services embedded in Azure core fabric (logging etc.), we are currently not using Azure dedicated AI services. This is due to technical limitations of AI Custom Vision (our proprietary models are mostly in CV domain). Instead, we are training Tensorflow models on-prem and deploying them into Azure VMs for production. We would love to operationalize them at least for inference at PAAS layer of sorts and are closely monitoring the development of Azure capabilities to evaluate the most cost-effective and technically feasible ways of deployment.