Demystifying Starlink Network Performance under Vehicular Mobility with Dynamic Beam Switching

In the last few years, considerable research efforts have focused on measuring and improving Starlink network performance, especially for user terminals (UTs) in stationary scenarios. However, the performance of Starlink networks in mobility settings, particularly with frequent changes in the UT’s orientation, and the impact of environmental factors, such as transient obstructions, has not been thoroughly studied, leaving gaps in understanding the causes of performance degradation. Recently, researchers have started identifying the communicating satellites to evaluate satellite selection strategies and the impact on network performance. However, existing Starlink satellite identification methods only work in stationary, obstruction-free scenarios, as they do not account for UT mobility, obstructions or detect dynamic beam switching events. In this paper, we reveal that the UT can perform multiple dynamic beam switching attempts to connect to different satellites when the UT-satellite link is degraded. This degradation can occur either due to the loss of line-of-sight (LoS) from changes in the FOV or obstructions, or due to poor signal quality, extending UT-satellite handovers beyond the well-known 15-second regular handover interval. We propose a mobility-aware Starlink satellite identification method that detects dynamic beam switching events, and plausibly explain network performance using UT’s diagnostic data and connected satellite information. Our findings demystifies the mobile Starlink network performance degradations, which is crucial to enhance the end-to-end performance of transport layer protocols and in diverse application scenarios.

💡 Research Summary

The paper investigates the performance of SpaceX’s Starlink low‑Earth‑orbit (LEO) satellite network when user terminals (UTs) are mounted on moving vehicles. While prior work has largely focused on stationary deployments, this study highlights that vehicular motion introduces frequent changes in UT orientation and transient obstructions (e.g., bridges, trees, signs) that can degrade the signal‑to‑noise ratio (SNR) and trigger “dynamic beam switching” events. Unlike the well‑known regular handover that occurs every 15 seconds at the 12‑27‑42‑57 second marks, dynamic beam switching can happen multiple times within a single handover interval, effectively extending the handover process and causing additional latency and throughput penalties.

To capture these phenomena, the authors built a measurement platform that collects (i) network performance metrics (ICMP ping for round‑trip time, iPerf3 UDP for throughput), (ii) diagnostic data from Starlink’s gRPC interface (UT tilt, boresight azimuth, GPS location, obstruction maps, and outage event logs), and (iii) precise timing information using the UT’s built‑in Stratum‑1 NTP server. Obstruction maps are 123 × 123 pixel 2‑D images that record the accumulated trajectories of connected satellites; they are refreshed roughly once per second. Data were gathered in two phases: a controlled laboratory experiment with a UT mounted on a tripod, and a large‑scale field campaign covering about 500 km of driving across rural, suburban, urban, and highway environments in the Midwestern United States. The mobile UT used a fixed 7.9° tilt and faced north, while a stationary UT was placed in a heavily tree‑obstructed residential setting for baseline comparison.

Analysis of the collected data reveals several key findings. First, when SNR falls below an internal threshold, the UT abandons its current satellite and initiates a dynamic beam‑switching sequence. This sequence can repeat several times within the 15‑second window, especially under obstruction or rapid orientation changes. Second, stationary UTs with a 17.9 % field‑of‑view (FOV) obstruction rate experience multiple failed beam‑switch attempts, leading to prolonged outages that account for roughly 22 % of total connection time. Third, mobile UTs encounter transient obstructions far more often; SKY_SEARCH events—where the UT searches for a new satellite—constitute up to 45.18 % of outage time, roughly double the proportion observed for heavily obstructed stationary terminals. During these events, round‑trip times spike from an average of 30 ms to as high as 120 ms, and UDP throughput drops from the 250 Mbps test rate to below 80 Mbps. Such variability directly harms transport‑layer protocols, especially latency‑sensitive applications such as real‑time video, online gaming, and remote control.

To address these challenges, the authors propose a mobility‑aware satellite identification method. The method first compensates for UT orientation by combining tilt/azimuth data with satellite ephemeris to compute the true beam‑steering angles. It then correlates time‑stamped obstruction‑map updates with outage‑event logs to pinpoint the exact moments of dynamic beam switching. By doing so, the algorithm can infer the currently serving satellite ID in both stationary and mobile scenarios, even though Starlink does not expose satellite IDs directly. The approach enables researchers and network operators to attribute performance degradations to specific satellite transitions and to distinguish between regular handovers and reactive beam switches.

Finally, the paper suggests practical mitigation strategies. Adjusting the UT’s tilt angle dynamically to keep the line‑of‑sight clear can reduce the frequency of beam‑switch attempts. Moreover, in dense constellation regions, a “quality‑first” satellite selection policy—favoring the satellite with the highest measured SNR rather than strictly following the 15‑second schedule—could improve link stability. These recommendations, combined with the proposed identification framework, provide a pathway to enhance end‑to‑end performance for transport‑layer protocols and to support a broader range of mobile applications on Starlink.

In summary, this work is the first to systematically quantify dynamic beam‑switching behavior in vehicular Starlink deployments, to develop a robust method for inferring serving satellites under mobility, and to demonstrate the substantial impact of such events on latency and throughput. The insights and tools presented are valuable for future research on LEO satellite mobility management, handover optimization, and the design of applications that rely on consistent high‑performance satellite connectivity.