NextSense: A Semi-Synthetic Sensing Data generation Platform

TL;DR

NextSense is a flexible semi-synthetic data generation platform for 5G networks, enabling customizable, multi-perspective data outputs to facilitate sensing applications without extensive real-world data collection.

Contribution

It introduces a modular platform that combines a 5G stack, channel, and UE emulators with customizable configurations and multi-perspective outputs for sensing research.

Findings

NextSense produces data closely matching real network measurements.

The platform supports extensive customization of radio and traffic profiles.

Validation shows it effectively proxies real sensing data.

Abstract

Emerging integrated sensing and communication (ISAC) applications require large volumes of data, but collecting such datasets in real networks is costly, time consuming, and often infeasible due to limited access to low level measurements. In this paper we present NextSense, an open and modular semi-synthetic data generation platform that consists of a 5G stack, a channel emulator, and an UE emulator. The platform allows users full customization on radio configuration, channel and mobility, and traffic profiles through an API and GUI, and produces multi-perspective outputs that combine symbol-level IQ samples, protocol traces, and key performance indicators across UE, RAN, and CN. This paper describes the NextSense's architecture, and validates its ability to act as a faithful proxy for real measurements in sensing use cases.