Transmission Guidance
The Terrestrial Network
Hubble's Terrestrial Network is composed of more than 100 million scanning gateways — including smartphones, consumer gateways, and industrial gateways — providing coverage across low- to high-density areas worldwide.
Scanning behavior:
- Scan intervals vary by scanner, with a minimum of approximately 5 minutes and a maximum of 30 minutes.
- Each scan window lasts 25–30 seconds, and reliably detects devices within 10–20 meters.
- There is no hard cap on devices detected per scan, though performance may degrade above 50–100 simultaneous advertisements due to radio saturation on the gateway.
- Offline, powered-off, or low-power smartphones may not execute Bluetooth scans or submit updates.
What to Expect
Detection on a BLE Network is probabilistic. The network provides statistical reliability over time, not guaranteed per-packet delivery. A single detection gap is not evidence of device failure.
Detection Frequency
Detection frequency depends on local gateway density, device configuration, and environment.
- In typical urban deployments, a well-configured device advertising every few seconds may receive anywhere from a handful to dozens of detections per hour.
- In low-density or indoor environments, gaps between detections are normal.
Data latency
Once a scanning gateway captures a BLE advertisement, data is typically delivered to the Hubble backend within seconds to a few minutes, depending on the scanner's connectivity.
Location & Data Quality
How Location is Determined
Position on the Terrestrial Network comes from the scanning gateway, not the tracked device. When a scanner detects a BLE advertisement, it attaches its own GPS coordinates at the time of capture and submits those as the device's reported location.
This means:
- Accuracy depends on the scanner. In open outdoor environments, positions may be accurate to a few meters. In dense urban or indoor environments, error can exceed 50 meters. Each packet includes
location.horizontal_accuracyandlocation.vertical_accuracyfields (in meters), reported directly from the scanning gateway's GPS chipset, which can help you assess the confidence of a given detection's location fix. - Position confidence is not a discrete field in the raw dataset. Customers who need to assess reliability should apply the data cleaning practices below to filter and weight detections based on their use case.
Factors That Affect Data Quality
- Scanner mobility: A moving smartphone may record a location that doesn't accurately reflect where the BLE signal was received.
- Redundant captures: Multiple nearby scanners can detect the same transmission, producing duplicate records with slightly different location and timestamp data.
- Connection backtracking: Scanners that disconnect and reconnect to networks can briefly report stale or conflicting location data.
- Signal interference: Environmental factors or competing wireless signals can affect capture reliability.
Data Cleaning Best Practices
Terrestrial Network data is collected directly from Bluetooth scanners without immediate filtering or aggregation.
- Filter duplicates: Multiple scanners may capture the same transmission. You may choose to filter packets with near-identical timestamps and locations, based on your use case's sensitivity.
- Detect outliers: Flag large, sudden jumps in location data (GPS jumps or teleportation). These often indicate a scanner reconnection event or faulty device reporting.
- Validate timestamps: Missing or inconsistent timestamps can distort transmission pattern analysis.
- Aggregate over time: Smoothing data over suitable intervals helps filter transient inaccuracies and gives a more reliable picture of device activity.
Edge Cases
Colocated devices
Colocated devices, such as multiple devices shipped in the same box, may show different detection rates. BLE operates in shared 2.4 GHz spectrum, and smartphone gateways scan intermittently rather than continuously. Factors like scan window alignment, local RF conditions, and radio collision dynamics mean that two devices placed inches apart can produce different packet counts, even with identical hardware and firmware.
Go Further: For a complete evaluation, we recommend testing network performance over multi-day windows rather than short intervals, looking at average detection rates across the full sample period.
Maximizing Discoverability
Try Network Explorer to test coverage under different beaconing strategies for your devices. For advanced configuration, use the Device Simulator tool in Expert Mode to evaluate probability of detection and time between detection based on transmit power and rate.
1. Set Transmission Power and Advertising Interval
Transmit power and advertising interval work together to determine how reliably your device is detected. Transmit power (Tx power) sets your effective range, and higher power improves detection reliability. The advertising interval (Tx rate) sets how often a passing scanner can catch a transmission, and shorter intervals increase the probability of overlapping with a scan window.
Because higher-power devices reach more scanners per transmission, they can advertise less frequently. The recommended interval therefore scales with device output power class:
| Device Output Power | Tx Power | Recommended Advertising Interval |
|---|---|---|
| Low power | 0–4 dBm | ~1 second |
| Mid power | 4–13 dBm | ~2 seconds |
| High power | 13 dBm and above | ~2–4 seconds |
At lower output powers, advertising frequency can make a large difference in detection performance. If your device is constrained to a low power class, prioritize a shorter advertising interval to compensate.
High-Output / Power-Amplifier Devices
Devices with an external power amplifier (PA) fall into the high-power class and can transmit up to +20 dBm, substantially increasing detection range in open or RF-noisy environments. This is most beneficial in large outdoor deployments, maritime or logistics environments, and areas with significant physical obstructions.
2. Account for mounting and enclosure
Metal enclosures, equipment chassis, and dense packaging materials significantly attenuate BLE signals. If your device is deployed inside or adjacent to metallic structures, factor in signal loss when setting Tx power and advertising interval.
3. Beacon Consistently
Regular transmissions ensure continuous network presence regardless of intermittent scanner coverage. Shorter beacon intervals during periods of movement and longer intervals when stationary can help balance battery life and detectability.
4. Rotate Payloads Frequently
Each payload rotation updates the content of the BLE advertisement. Independently of device ephemeral ID (EID), which rotates based on counter_source, the incremented sequence number for the advertisement packet (used for nonce in encryption) is intended to signal that a fresh payload is available — which anecdotal evidence suggests correlates with more captures from nearby scanners.
Hubble supports up to 1,024 unique sequence numbers per EID rotation period. For applications where discoverability is a priority, rotating at or near this rate is recommended.
Note: For IN100 chips, payload rotation is independent of EID rotation (controlled via
period_exponent). You may rotate payload using a dynamic data source in your configuration file.
5. Rotate MAC Addresses
Regularly changing the MAC address protects user privacy and helps avoid long-term tracking of a single device identifier by the network. We recommend using a Non-Resolvable Private Address (NRPA) in your advertisement.
6. Keep Firmware Current
Keep your device firmware current with the latest improvements in BLE performance and security to support consistent and effective transmissions.
Platform Differences for Mobile App Scanning
Scanning behavior varies by mobile platform and app state:
| Behavior | Android | iOS |
|---|---|---|
| Scan interval | ~15 minutes | ~5–30 minutes |
| Force-quit behavior | Scanning continues in background | Background scanning suspends until app is reopened |
| Foreground | Continuous scanning, real-time detections | Continuous scanning, real-time detections |
| Background | Best-effort; detections sent when location is available | Best-effort; detections sent when location is available |
Learn how Hubble Connect mobile app scans for your devices.