The bus edge in our design includes all the systems deployed on a vehicle. For this solution design testing, we have deployed a physical EPIC iO Universal IOT Gateway (UIG) edge device. The UIG is a rugged device capable of functioning autonomously in harsh environments that combines state-of-the-art, fast wireless connectivity with AIoT solutions for AI computing and video storage at the edge.
The system used in our lab closely resembles the specifications of a device that would typically be used in a public bus transportation environment. The UIG can provide public wireless connectivity that enables riders to stream music/videos, search the web, check email, and more. The mobile wireless connectivity provided by the UIG also enables critical solutions using private wireless networking that operates in and around bus terminals and depots. That helps streamline operations, such as bus route management, alerts on potential delays, asset tracking, and maintenance needs.
The EPIC iO UIG in our lab also hosts a Dell 7090 computer for hosting DeepInsight analytics workloads. The Dell 7090 can process data on the vehicle where it is generated, significantly reducing latency for event detection and network bandwidth usage. DeepInsight's AIoT applications detect abnormal events in near real-time to manage systems more effectively. The 7090 edge device was configured with the latest version of DeepInsights software.
The UIG also includes an R1900 Series 5G Ruggedized Router for internal, external, and wireless communication. The R1900 Series router was designed and built to manage networking traffic for high-performance mobile 5G connectivity and edge computing. The R1900 is an all-in-one mobile networking solution that can withstand extreme environments for installations within a wide range of vehicles, including metro transit buses, school buses, commercial fleets, and first responder vehicles such as police, fire, ambulance, and mobile command trailers. DeepInsights enables enterprise and public organizations to deploy, manage, and update EPIC iO's and third-party vendors' AI models and IoT data flows. The platform was designed to be fully device agnostic and cloud-native with a containerized, microservices design that enables deployment to any compute edge that supports containerization (rugged, MEC, private, public, hybrid, secure, and so forth) EPIC iO's AI Models are optimized for Intel OpenVINO and NVIDIA GPUs (CUDA). IoT components can run on Intel or ARM processors for deployment flexibility.