Zensors AI-Powered Platform Transforms Passenger Experience at Harry Reid International Airport

Maintaining a consistent passenger experience becomes increasingly difficult due to seasonal peaks, event-driven spikes in traffic, and the complexity of balancing staffing to guide passengers to the optimal checkpoints. In the absence of real-time insights, managers relied on phone calls to remain informed on checkpoint statuses, leading to fragmented information, while passengers lacked insight into checkpoint wait times-further exacerbating bottlenecks and information gaps. Without clear, quantitative visibility into the passenger journey through the airport, wait times grew, congestion became prominent, passenger anxiety increased, and the risk of missed flights multiplied.

Although a legacy lidar system was initially implemented to model passenger flow, it was difficult to manage and ultimately was unable to generate real-time actionable insights for end users or airport staff to resolve issues and plan proactively. To enhance the passenger experience for everyone, LAS needed a solution that delivered real-time information, trend analysis, and predictive capabilities-the kind of intelligence its older system could not provide.

Rather than invest in brand new infrastructure, LAS used Zensor's platform to turn its current security surveillance system into a full-scale data collection and analytics pipeline using an established system of over 500 security cameras.

Deployed in two months with zero capital expenditure, this solution shifted operations from manual data collection on an hourly basis to automated, instantaneous updates. Zensors' platform empowers airport operations, personnel, and partners to proactively manage queues, optimize staffing, and significantly enhance passenger experience through data-driven decision-making.

This system can understand and automate the temporal world. Unlike traditional AI models that struggle with 3D spatial awareness and cannot process real-world visual data, Zensors' multimodal transformer-based AI can understand video data in real time while tracking a variety of valuable metrics.

Zensors' physical intelligence platform was selected by LAS based on its comprehensive understanding of airport operational needs and constraints, while having a smooth implementation process with existing infrastructure.

LAS needs to process thousands of hours of video data daily. This massive endeavor is made possible due to the entire processing pipeline being built with Zensors’ application, using an NVIDIA AI technology stack comprised of NVIDIA Dynamo-Triton, TensorRT™, CUDA® Deep Neural Network, NVIDIA CUDA, and NVIDIA Developer Data Loading Library. 

NVIDIA technologies were vital for scaling inference and enabling Zensors' platform to process and produce real-time data analytics. Video workloads are spiky and heterogeneous. The ability to blend streaming and batch data, schedule across whole GPUs, and push pre- and post-processing onto GPUs was needed to avoid CPU bottlenecks.