State Space Representation

Description

State Space Representation (SSR) is a mathematical framework adopted within 3GPP specifications to model the state of a User Equipment (UE). The 'state' typically encompasses variables such as position, velocity, acceleration, and potentially other dynamic parameters. This representation is fundamental for advanced positioning techniques, where the network or the UE estimates the UE's location not just as a single point but as part of a continuous trajectory. The model operates by defining a state vector that contains these key parameters and a state transition model (often based on physics, like constant velocity or acceleration models) that predicts how the state evolves over time. Measurements from network nodes (e.g., gNBs, eNBs) or satellite systems (like GNSS) serve as observations that are fed into estimation algorithms, most commonly a Kalman filter or its variants, to update and refine the predicted state, reducing estimation error.

The architecture for SSR-based positioning involves several key components. The UE or a location server (e.g., Location Management Function - LMF) maintains the state space model. The UE may report its state information (or measurements that allow the server to compute it) to the network. Specifications such as 3GPP TS 37.355 (LTE Positioning Protocol - LPP) and TS 38.305 (NG-RAN; Stage 2 functional specification of User Equipment positioning in NG-RAN) define the protocols and procedures for exchanging this data. The model's parameters, including process noise and measurement noise covariance matrices, are critical for tuning the filter's performance, balancing responsiveness to new measurements against smoothing of noisy data.

SSR's role in the network is pivotal for high-accuracy, low-latency positioning services required in 5G and beyond. Unlike simpler methods that provide snapshot locations, SSR provides a filtered, predictive estimate. This is essential for use cases like vehicle-to-everything (V2X) communication, drone tracking, and industrial IoT, where knowing not just where a device is, but where it will be, is necessary for safety and automation. It integrates with various positioning methods, including Observed Time Difference of Arrival (OTDOA), Uplink Time Difference of Arrival (UTDOA), and multi-round-trip-time (Multi-RTT), by providing a common mathematical framework to fuse these measurements over time, significantly improving accuracy, especially in challenging environments like urban canyons or indoors.

Purpose & Motivation

SSR was introduced to address the limitations of traditional, discrete positioning fixes in mobile networks. Earlier positioning methods often provided independent location estimates at specific request times without leveraging the temporal correlation and motion dynamics of the UE. This resulted in less accurate and 'jumpy' location tracks, especially when measurements were noisy or infrequent. For emerging 5G use cases—such as autonomous driving, augmented reality, and mission-critical communications—these limitations were unacceptable. There was a clear need for a method that could provide smooth, predictive, and highly accurate continuous location tracking.

The creation of SSR was motivated by the need to meet stringent 5G positioning requirements defined by 3GPP, which include sub-meter accuracy and ultra-low latency for certain verticals. By adopting a state-space approach, which is a well-established concept in control theory and signal processing, 3GPP provided a standardized framework for filtering and prediction. This allows the network to maintain a persistent 'understanding' of a UE's kinematic state, rather than treating each positioning event in isolation. It solves the problem of integrating heterogeneous measurement data (from cellular signals, GNSS, sensors) over time in an optimal way, minimizing the impact of individual measurement errors and providing a consistent trajectory.

Historically, similar filtering techniques were used in proprietary or non-standard implementations. SSR's standardization in Rel-15, particularly within the 5G NR positioning architecture, ensured interoperability between network equipment and devices from different vendors. It addressed the challenge of supporting advanced mobility in dense networks and enabled new service level agreements (SLAs) for vertical industries that depend on reliable and precise real-time location information.

Key Features

• Models UE kinematic state (position, velocity, acceleration) as a time-evolving vector
• Utilizes Kalman filtering or similar algorithms for optimal state estimation and prediction
• Integrates measurements from multiple sources (e.g., NR, LTE, GNSS) into a unified framework
• Defines standardized reporting formats and protocols (via LPP) for state information exchange
• Supports both network-based and UE-based positioning architectures
• Enables continuous tracking and predictive location for dynamic use cases

Evolution Across Releases

Defining Specifications